Public Administration Select Committee
Oral evidence: Crime Statistics, HC 760
Tuesday 21 January 2014]
Ordered by the House of Commons to be published on 21 January 2014.
Written evidence from witnesses:
Members present: Bernard Jenkin (Chair); Paul Flynn; Priti Patel; Lindsay Roy; Mr Andrew Turner.
Questions 463-603
Witnesses: Sir Andrew Dilnot, Chair, UK Statistics Authority, Ed Humpherson, Head of Assessment, UK Statistics Authority, Jil Matheson, National Statistician, Office for National Statistics, and John Flatley, Head of Crime Statistics, gave evidence.
Q463 Chair: May I welcome our first panel to this last session on crime statistics? Could I ask each of you to identify yourselves for the record?
Sir Andrew Dilnot: I am Andrew Dilnot. I am Chairman of the UK Statistics Authority.
Ed Humpherson: Ed Humpherson. I am the Head of Assessment at the Statistics Authority.
Jil Matheson: Jil Matheson, National Statistician.
John Flatley: John Flatley, Head of Crime Statistics at ONS.
Q464 Chair: Thank you all for coming. A very warm welcome to Ed Humpherson, who was appointed only a few days ago and will be bringing a fresh perspective to these matters, but also a particularly warm welcome to Jil, a veteran of this Committee, who has just announced her retirement. Can I take this opportunity to thank you for all the work you have done and you are continuing to do until you leave your post, I believe, this summer?
Jil Matheson: Thank you, Chairman.
Chair: I am afraid that is the nice bit of the cross-examination. All our questions will not be that easy to respond to, but can I enjoin you to give as short and crisp answers as possible? I will pull you up if I think you are going on for too long. I also enjoin my colleagues to ask crisp and to-the-point questions.
First of all, as Chairman of this Committee, I would express some disappointment that, from a state of reasonable and fairly commonplace ignorance about the very technical matter of police-recorded crime statistics, we have stumbled into what appears to be a very well-kept and widely-shared secret: that these figures were never very reliable, to the extent that the Home Editor of the BBC even told me that he did not think the de-designation of the statistics was much of a story and the BBC did not even run it on their website. What have you to say to this situation?
Sir Andrew Dilnot: This issue of police-recorded crime statistics is something that has been in the eye of this Authority and the ONS for some years. Going back to the early part of this decade, there was a monitoring report on this that raised concerns about the lack of regular auditing since 2007-2008. The National Statistician’s report, now nearly three years ago, raised those concerns. There was concern raised in the initial assessment report some three years ago.
Concerns in particular about the gap between the Crime Survey for England and Wales and the police-recorded crime statistics have increased in recent years. A paper published by the Office for National Statistics in January of last year drew that—and the way in which that has increased—very firmly to our attention. That was part of the reason that led to the Statistics Authority’s monitoring and assessment team starting off a conversation with the ONS last January—a year ago—to move towards a reassessment. This is one of the first reassessments of statistics that we have conducted. The assessment we hoped would start in May 2013; it actually started in June 2013, and is now completed. However, if you were to ask me whether I wished we had been even swifter than that, the answer is, honestly, yes. Given the concerns that we had repeatedly expressed about the lack of adequate audit and the lack of progress in periodic and regular audit of these figures in the last three years, it would, in an ideal world, have been good to have got to where we got to more quickly than we did.
Jil Matheson: When I did my review, following the Authority’s monitoring review, and spoke to a lot of people—
Chair: This was the 2011 review.
Jil Matheson: That is right. It was clear that it was one of the issues that people were expressing concern about, hence my recommendations, which included recommendations for regular periodic audit. When you have statistics being produced from the kind of data that police‑recorded crime is, it is important to have in place a lot of things: quality assurance, trained staff, recognition that this is really important, plus strong independent audit. At that time, I was not convinced that that strong independent audit was in place, hence the comments in the report.
Q465 Chair: How surprised were either of you that you finished up de‑designating police-recorded crime?
Sir Andrew Dilnot: The Authority has rarely de-designated a series of statistics.
Chair: So how surprised were you?
Sir Andrew Dilnot: I was not surprised by the time we did it, because we had been working on it for some considerable number of months.
Q466 Chair: How surprised would you have been if we had not had our inquiry?
Sir Andrew Dilnot: I do not think we would have been surprised. There are four reasons—
Q467 Chair: It is a Catch‑22 question, is it not? If you were not surprised, why were you not doing something more about it at an earlier stage?
Sir Andrew Dilnot: There are four reasons that we have de-designated.
Chair: We know that. I am asking why more was not done at an earlier stage. You have expressed regret for that, and I am very grateful for that frankness, but why do you think more was not done earlier?
Sir Andrew Dilnot: It is very hard to answer such hypothetical questions, but I think the answer is, in the first round of assessments of statistics, the Authority quite properly sought to assess all of the statistics that sought to be national statistics—more than 1,000 series; 240 reports. When we got to the end of that sweep through all of the statistics for the first time, the right thing to do—Mr Humpherson might want to say more about plans for this in the future—was to move to a more risk-related programme of assessments, which we have moved on to. We got pretty soon to the national statistics on crime, so a year ago we alerted ourselves and got on with discussions with the ONS. In the best of all possible worlds, we would have seen even greater urgency at that point. We did not. That is something that we can think about in the future, but these processes also have, as it were, dignified time to them. Just as your own report, which was announced last autumn, will not be concluded until early this year, our assessment report, which began in June, before your Committee had started its own work, has taken six months, as most of our assessment reports have done.
Q468 Chair: Jil Matheson, you wrote in the ONS Statistical Bulletin as recently as June 2013: “The system for recording crime in England and Wales by the police is widely recognised by international standards to be one of the best in the world”. This was, in your defence, quite a widely-held view. Do you still hold this view?
Jil Matheson: It was a widely-held view, and there are still elements where I do hold that view; for example, the fact that there is a standard—the crime-recording standard—and that there are procedures whereby the Home Office sets those standards and requires police forces to record. The implementation of it I think is the question, but lots of countries do not even have that basic system of setting a standard of having, as we do here, Force Crime Registrars, who are charged with a role in the operation. They do not have the kind of quality assurance, and they do not have some of the reporting at a national level that we do here. Those elements are really important, and they are important to a world-class system. I do hold to that view still in that respect.
Q469 Chair: How much do you think it is fair to say ONS has been a little bit too complacent and a little bit too comfortable with the way things were?
Jil Matheson: Let us just go back. ONS took responsibility for this in April 2012. Within nine months, John Flatley here had analysed, written and published the report that Sir Andrew has already referred to.
Chair: That is the 2011 report.
Jil Matheson: No, January 2013.
Q470 Chair: The January 2013 report, yes. That is a year ago today. In that report you said: “We have already highlighted our concerns about the quality of crime recording by the police, and the variations in trends between recorded crime and our own Crime Survey for England and Wales. I am pleased this has been recognised by the Authority.” Then what?
John Flatley: That is a quote from Glen Watson, which was only made last week in response to the Authority designation report.
Chair: I beg your pardon?
John Flatley: Going back to the point that Jil was making earlier—just to add to the points about the crime-recording system in the UK—we are quite unique in having this victim‑focused approach. Clearly there are issues and concerns about whether that is implemented in the way it should be in every force for every offence, but in many other jurisdictions, crimes only get recorded if a prosecution takes place. We have got a much fuller picture of crime in this country than in many other jurisdictions.
Q471 Chair: But the point remains: if ONS was so pleased that UKSA had recognised these deficiencies, the implication is that you have been hammering at the door for a long time to try to get UKSA to do something about this.
Sir Andrew Dilnot: My interpretation of this is that nine months after the ONS took responsibility for this, Mr Flatley and a colleague of his published a report in January of last year, exactly a year ago. It was in that very month that Mr Humpherson’s predecessors first discussed with the ONS the fact that we wanted to reassess crime statistics. We published something saying we hoped to start that reassessment in May; we did not start it until June. It then took us six months. I would that each of those steps had been a few months sooner than it was, but the process has got us to the right conclusion.
Chair: How fair is it for us to claim the credit for highlighting these problems and promoting your attention on to this, so that you have taken this decision? You credit us in your letter, do you not?
Sir Andrew Dilnot: Yes. The evidence that has been produced for the Committee is one of four elements of our concern, and they all fit together; they are not inconsistent with one another. It is certainly the case that the assessment process took seriously this public evidence. One of the things that we note from our own assessment process is that although more than three times as many users responded to our call in the consultation for the assessment report this time than last time, we did not in that consultation receive the kind of evidence that was made available on the record to this Committee, and certainly that evidence helped us be sure that de-designation was the right thing. The hypothetical question—“Would we have done it anyway?”—I do not know the answer to.
Q472 Chair: The next question is, therefore: how overstretched is your assessment function?
Sir Andrew Dilnot: I will turn in a moment to Mr Humpherson to say more on this. It is the case that we need to change—and we are changing—the way in which we do our assessment work from, in the first instance, doing everything to now doing work that is much more risk-centred. That means that we need to think quite hard about how we work out where the high-priority risks are. Perhaps Mr Humpherson would like to say something on that.
Ed Humpherson: That is absolutely right. The assessment team is well equipped to undertake the kinds of assessment of the high-materiality, high-risk official statistics and the extent to which they comply with the code of practice. That is the job now. Now that the first 240 assessments have been completed and we are into the phase of reassessment, we need to concentrate our resources and focus the team, and we have the resources to do that.
Q473 Chair: You said yourself earlier you had this huge programme of assessment, and at that point police-recorded crime got through the assessment and maintained its designation. Is this a problem of resources for the assessment function or of the independence of the assessment function?
Sir Andrew Dilnot: If the implication in your question is that they should not have been given national-statistics designation in 2011, I am not sure that I agree with that. The world has changed between 2011 and now in ways that are material. We said in our letter to you that we noted from the accumulation of evidence in the last three or four years that there were problems.
Q474 Chair: The answer is that we have been looking for the evidence. Maybe you were not looking for the evidence sufficiently before. If anything, I would suspect there was a greater awareness amongst the police before we started our inquiry than there was two years ago that they had a persistent problem.
Sir Andrew Dilnot: That is almost certainly true. It is also the case that if you look at the work demonstrating the gap between the data from the Crime Survey for England and Wales and the police-recorded data, that has grown very significantly in recent years. At the time that the last report was done, it was only relatively recently that the Audit Commission’s regular audits had taken place, and the work on that gap shows that while that work was going on, we got over 90% of comparable crimes being recorded by the police-recorded crime statistics, compared to the 50% that it was before and the 70% to which it has now fallen. At the time of the initial assessment report, there simply was not the evidence that there is now. The auditing process had only recently stopped from the Audit Commission and there had been a suggestion from the Home Office that the Home Office and HMIC would be doing auditing, which simply has not happened over these three years.
Q475 Chair: What implications do you think this has for other assessments of administrative data that you make that are national statistics? Have you got the resource to look at the sheer volume of material that you need to look at and to make that kind of judgment?
Sir Andrew Dilnot: I will turn to Ed in a moment for more detail, and also to Jil, who has a responsibility as National Statistician. It certainly is the case that there are particular risks with administrative data, and especially administrative data where the performance of the organisation is assessed on the basis of the data. The thing that matters crucially in all of this is the integrity of the data. It is very important for us to realise that the data does not just matter for purposes of accountability and transparency—although it matters a great deal for that—but it also matters for the functioning of the underlying service. If you are running a police force, or any other large public-sector body, you need the data to know what is working and how to allocate your resources. Do we have the resource to do this? We certainly need to focus our resources and work out where we should be putting effort. That is something that Mr Humpherson is thinking about.
Ed Humpherson: I have just a couple of observations to add to that. The first is on your question of administrative data and assessments thereof. We need to be clear in the way we prioritise the reassessments of official statistics that we place those statistics that are drawn from administrative data high up our list so we get to them first. That is the first thing.
Secondly, I would say that we need to have a clear idea of what good looks like. We would be looking, in any situation where administrative data is called on, for clear standards at the operational level for how data is recorded, done by accredited people, equivalent to Force Crime Registrars, and that being the primary source of assurance that the underlying data is right. On top of that, we would always want to see some assurance from a regulatory or an audit function—the analogue to the role that we hope HMIC take on in this instance. We would always be looking for that. We would be looking for the producer of the statistics—in this case the ONS—to be suitably sceptical and to be asking the kinds of questions that, indeed, John Flatley and his team did ask in January 2013, so when we conduct these reassessments, we go in with a clear idea of what we are expecting to see to ensure compliance with the Code of Practice for Official Statistics.
The third thing to say is we are very keen to ensure that we think about this systemically and not just statistic-by-statistic. We have announced that we are going to start undertaking two monitoring reviews: one of situations where statistics are used to support compliance with performance targets; and the other on the general question of what are the standards of statistical audit of administrative data. Those are the things that we will be taking forward in the assessment function.
Q476 Chair: Thank you. I must just interject an apology to Mark Easton, who has tweeted that he did blog about the de-designation of the crime statistics. I said that it did not appear on the main news bulletins of the BBC, but he did say in his blog: “Recorded crime figures have always been a lousy way to identify crime trends”. If that was his view, why was it not UKSA’s view?
Sir Andrew Dilnot: Our strong view is that we need both forms of data. It is very important that we have something like the Crime Survey for England and Wales. It is also very important that the police have data that is reliable. Our whole purpose is to be here to promote the collection and communication of good data for the public good, and it seems to me inconceivable that it could not be for the public good for the police to be collecting and presenting data that is extremely clear, overwhelming in its integrity, and available to all so that the police themselves in the management of their own affairs, and the public, can see what is going on.
Q477 Chair: I do not feel that I have yet got a sense of what lessons the assessment function of UKSA learns from this episode.
Sir Andrew Dilnot: What lessons did we learn? A number of lessons. One is to be even more risk-based in our decisions about what to do assessments on. We are pleased that we chose a year ago to reassess crime statistics, one of the very first series that we reassessed. We might have done that more quickly. We might, given the prima facie evidence that we had from the January 2013 paper of the seriousness of this, have accelerated the process. It is quite conceivable that we could have aimed to produce a report six months ago. Had we produced that report six months ago, we would not have had the benefit of the evidence that has been given to this Committee, which certainly has been helpful in our report. We need to be a bit more nimble, we need to be more aware of possible big risks, and we need to be very aware of the risks around the use of administrative data.
Q478 Chair: But that proactivity requires an independent function in assessment, and it requires resource.
Sir Andrew Dilnot: Yes.
Q479 Chair: How independent and how well-resourced is the assessment function?
Sir Andrew Dilnot: The assessment function is completely independent. No pressure has ever been put on me, with my assessment hat on. Ed has only been in post for a couple of weeks, but I do not think Richard Alldritt ever felt that pressure was put on him by Government. Indeed, were Government ever to put pressure on, one would, to quote a famous film star—I am not quite sure which—say, “Go on, make my day.” If any pressure on our independence were ever to be put, I would come straight to you, Chairman; I would also go straight to the press. I have not seen it and I do not imagine that will be the case.
With regard to resource, I said repeatedly to Mr Humpherson’s predecessor, and I say to him too, that, given that the assessment function is a small part of the overall budget of the Statistics Authority, he should not feel, as Mr Alldritt should not ever have felt, budget constrained. If the Head of Assessment needs more resource and they come to the board of the Authority, the board of the Authority would, I hope, always say yes.
Chair: What a wonderful start to your new job, Mr Humpherson.
Ed Humpherson: Yes, indeed. He has said that to me. I would just like to say something else on independence. My background is working in an organisation where independence is deeply prized, the National Audit Office. I come to this role with an understanding of what it is to form an independent judgment, independent of pressure, and to form your judgment on the basis of the criteria laid down for the work. I can say it is exactly the same in the assessment function; I am seeing exactly the same disciplines, rigour and independence. It feels like a process of forming independent judgments, just to confirm what Andrew said about the lack of any sort of pressure.
Q480 Chair: Lastly, Jil Matheson, in a way you are caught in a rather complex set of relationships, because you do not initiate the data—you do not create the data—the Home Office remains responsible for collecting the data and passing it to you for publication; and ultimately UKSA’s assessment authority is responsible for verification. How much confusion is there about accountability and who is actually responsible for ensuring that police-recorded crime statistics are accurate?
Jil Matheson: If I go back to the report and the recommendations in the report that I made nearly three years ago now, the reasons—and these were accepted—for recommending that the collection of police-recorded crime statistics remained the responsibility of the Home Office to mandate and of the police themselves to comply with that mandate were several. One of the reasons is that we know from other data collections, from statistics more generally, that you get a better job done if the people doing it are the people who are ultimately going to use that data. It is the point that Sir Andrew was making: that this is not just an add-on for statistical purposes; this is an integral part of police operation and Home Office responsibility.
Q481 Chair: But I recall you yourself have said: “We are not responsible for collecting the data”. To what extent does that absolve you for producing, albeit data that is accurately referenced to itself—
Jil Matheson: Let me go on. The first line of responsibility is with the people who collect the data.
Q482 Chair: It is not, is it?
Jil Matheson: ONS has a responsibility to seek assurance from those who collect the data about the quality of that data. That is one of the issues that the assessment report found, quite rightly: that ONS could not get that assurance from an independent audit function.
Q483 Chair: When did you first realise you were not getting that assurance?
Jil Matheson: This goes back to John Flatley’s article. He can talk more about this, but in being able to plot the decline in the proportion of crime reported to the police that appears in the recorded crime figures, the question gets raised, “How is this data quality‑assured, and how is it audited?” We know that there was not a full, clear, clean, independent audit function.
Sir Andrew Dilnot: In a way this is a further answer to one of your earlier questions, “What lessons have we learnt?” It is very important for the assessment function to be very plugged in to concerns that are being raised within the production function, both within the ONS and in the wider GSS. We need to think quite carefully about making sure that those channels of communication are very open, so that if there are concerns being felt by the producers about some of the administrative data that is coming to them, there is a very clear and open channel. In principle it is clear at the moment, but we might want to think about whether we should put a mechanism or a structure in place that makes it easier and more obvious for somebody—
Q484 Chair: Is it not more about attitudes than mechanisms, and being receptive?
Sir Andrew Dilnot: It is probably both about being receptive, which on the assessment side we always would be, and also the knowledge on the production side that there will be receptiveness on the assessment side.
The second point that I want to note is that it is striking that in our open consultations, either for this assessment report or the previous one, we have not received the kind of evidence and input that this Committee has. We should reflect more on whether there are ways in which we could change our process such that that might encourage more of the whistle-blowing type of activity—I am not sure about that—but also we should reflect on whether there are ways in which we can work even more closely with this Committee. In a way, it is a shame that we did not think in, if not January of last year, when we first thought of doing an assessment report, certainly in June of last year, when we actually started it, of alerting this Committee to the fact that we had concerns about this area and we were starting an assessment report, and asking whether there was any way in which we could collaborate. That is something that we might think more about in the future.
Q485 Lindsay Roy: Good morning, panel. Can you just summarise for me what were the key indicators that led you to believe that all was not well with the recording of crime statistics?
Sir Andrew Dilnot: There were four issues. The first was that we have a survey of crime in England and Wales that does not involve the police-recording of crime, and the proportion of crimes for a comparable set of crimes reported in that survey that were also reported police-recorded statistics fell from 92% to 70% over a period of five years. That was prima facie evidence that the gap was growing between the experience of crime in individuals’ and households’ lives and the data that was being reported by police.
The second was that in a whole series of reports going back to 2006, the importance of regular, periodic, independent audit of crime data had been pointed out. That largely stopped in 2007-2008, since when the gap has grown and, although HMIC now has plans to do significant audit, it has not been doing large-scale regular audit of these figures. The third is that, to the extent that there has been audit—for example, the HMIC audit of Kent, which was commissioned by Kent—that has raised concerns that seem to us serious concerns. 90% does not seem to me obviously all right, and if there is a dispersion across the country around 90%, which is also clear from the HMIC work that has been done, that is cause for concern.
To cap all of those suggestions that there was a problem, your Committee has heard evidence from a range of senior and less senior officials—current officials and past members of the police force—all of whom suggest that they have evidence of serious misreporting, under-recording and mis-recording of crime. Taken together, that seemed to us to be an argument for immediate de-designation of the statistics.
Q486 Lindsay Roy: The Chairman asked why you had not acted earlier. I am not completely satisfied. Can you add to what you said before?
Sir Andrew Dilnot: It is always hard to work out why you did not do something; it is easy to work out why you did do something. The absolute truth is that until about a year ago we were completing the assessment of all national statistics. We assessed this set of statistics first in 2011; at that stage we did not have the accumulation of evidence that I have just described to you and, although none of us was around at that time, the decision was made—it seems to me properly—to allow designation. Once we got to the end of that process of looking at everything, we thought about what should be done next, and this was on the list of about the first set of things to be done next. I wish that we had taken it even more seriously. I think we could probably have accelerated that process, perhaps by six months. I think it would have been possible to have produced a report in the summer of last year rather than now. Of course, that report would not have had the benefit of your evidence, although if we had also when we thought about that report got in touch with the Committee, then perhaps the Committee itself could have accelerated its inquiries.
Q487 Lindsay Roy: To what extent was there too much focus on quantitative monitoring rather than a qualitative analysis?
Sir Andrew Dilnot: The quantitative modelling helped us in that it established this gap. In a sense, the qualitative process is precisely what auditing would do for us. Auditing, going in both deep and broad, would have given qualitative assurance. Probably all of us would recognise that the fundamental problem here has been the lack of auditing. Actually, the fundamental problem, if there is a problem, is about the integrity of the underlying data. That is the crucial thing, but the way of finding that out must be independent, regular, periodic and known auditing.
Q488 Priti Patel: Sir Andrew, you have just mentioned the issue associated with the lack of auditing. There are two issues here. One is: what would you do differently with regards to the lack of auditing? Second—because it has to be associated with the degree of public confidence around crime statistics—in your view, in light of everything that has happened, subsequent reports and investigations, what do you now think is considered to be an appropriate level of public confidence around crime statistics?
Sir Andrew Dilnot: The Crime Survey for England and Wales, which is a large part of it, the public can be really rather confident in. It is worth noting that the broad picture you get from the Crime Survey for England and Wales is roughly similar to the broad picture you get from the police-recorded crime statistics. Both have been showing reductions, although the police-recorded crime statistics in the last four or five years have been showing almost twice as rapid a reduction. The public can be confident in the broad picture given by the Crime Survey for England and Wales. We have suspended our designation of the police‑recorded crime statistics. They are still official statistics, but we have said, effectively, that we cannot be confident on the basis of the information that is available at the moment that these statistics give a full, accurate and fair reflection of the underlying truths of crime.
Q489 Priti Patel: Specific to auditing, what would or could be done differently?
Sir Andrew Dilnot: Regular, periodic, rigorous audits. The precise judgment about exactly how frequently, how large the samples should be and how the sampling frame should be done is for auditing experts on our assessment side, for the producers, for HMIC and for the police, but what I am convinced of—and what the UKSA board is convinced of—is the need for regular auditing. When you look back at the records, it is almost as if there was a sense in 2007-2008, at the end of a period when the Audit Commission had been doing annual audits, that things had improved—you certainly see that in the data—and, “Since it’s improved, we don’t need to do it anymore”. Well no, the whole point about audit is you have to know that it is coming at you every single year.
Q490 Chair: Any administrative data that is subject to any kind of target culture must be vulnerable to this kind of manipulation. Perhaps that would be a good generic test to apply across designated data.
Sir Andrew Dilnot: Absolutely. One of the monitoring reports that Mr Humpherson just described, which was commissioned in the second half of last year and will be done during this year, tackles precisely that set of issues. I think that is right.
Ed Humpherson: That is right.
Q491 Lindsay Roy: What are the possible reasons for the mis‑recording of crime that you have discovered and which is the most significant?
Sir Andrew Dilnot: I will pass to Mr Flatley in a moment, who describes in his January 2013 paper a range of possibilities. Those are still the range of possibilities, and precisely until we have audit that is rigorous and reliable, it is very hard for us to know what the balance of importance of those is.
Chair: Just to mention, if anybody is following this on Twitter, “#crimestats” is a good hashtag.
John Flatley: In our paper, we highlighted a number of reasons. A target-driven culture was one of the key ones that we identified from the anecdotal evidence that had been fed back to us from the front line, so to speak. The issue of training and awareness of the counting rules is a very important one. A lot of officers are confused about the distinction between charging standards, or the standards the CPS require officers to adhere to when they take a case through to prosecution, and the counting rules, which require them to record in accordance with the law.
HMIC did an interesting review back in 2012, and they identified three key factors that they thought were very influential in getting good crime‑recording decisions. One of them was leadership and good governance; the second was having proper checks and quality assurance proportionate to the risk; and the third was having skilled people, and in particular having objective and strong Force Crime Registrars who could stand up to performance pressures and ensure that the standards were followed correctly. All those factors play into the mix.
Q492 Lindsay Roy: To what extent have you been able to evaluate the importance or the significance of each?
Sir Andrew Dilnot: In our assessment report, we were not able to evaluate the importance of each. In our assessment report, we had the evidence on the gap; then we had our concern about auditing, which is our overriding concern; and then we had the evidence that has come to you and a little bit of evidence that came to us from users in our assessment consultation, although not very much. None of that is statistical evidence, so we are simply not in a position to be able to say what we think, in quantitative terms, the various possible causes have contributed.
Q493 Lindsay Roy: But are you pursuing that or is somebody else pursuing that on your behalf?
Sir Andrew Dilnot: In evidence to your Committee earlier this month, HMIC committed to and said that they had been given resources to engage in detailed crime‑data integrity auditing over the next period. That is what can provide that kind of statistical information, although their current proposals will only give statistically significant answers to some of those questions, and to get statistically significant answers across all forces would probably require larger sample sizes and more money than thus far they appear to have been suggesting, but I do not think that is yet finalised.
Q494 Lindsay Roy: How do you explain the divergence between police-recorded crime data and the Crime Survey for England and Wales? What action have you taken in relation to that?
Sir Andrew Dilnot: The action we have taken in relation to that is primarily that we have de‑designated the police-recorded crime data, because that is one of the bases on which it seems to us unsatisfactory. The possible reasons for that are as described by Mr Flatley earlier on: performance-related target regimes; the possibility that the complexity causes problems; the possibility that the training needs to be better; and the possibility that Force Crime Registrars do not have as much authority and domain as they need to have. In the end, though, the core point must be that if there is a problem, it is in the interests not just of the whole country but also of the police force to get this right, because until we have accurate data, not only do we not know how well the police force is doing from the outside, but it will be very hard for the police force to run itself effectively.
Q495 Lindsay Roy: To what extent, then, do you think there need to be cultural change within the police?
Sir Andrew Dilnot: I do not think I can really comment on that. What we see are the outputs of the statistics, and we see outputs that demonstrate to us that there is a reasonable case to be made that there is a gap that we cannot explain. I do not know enough about what is going on and I do not think that can be our job as a statistics authority. Our job is to say, “There is evidence here that something needs to be done to demonstrate that things are done well.” We should not become another regulator of the way that the police itself works.
Q496 Chair: But if you had understood the cultural deficiency that has led to this underperformance in the production of statistics, maybe you might have acted earlier. Is your job not to understand that sort of aspect of policing?
Sir Andrew Dilnot: It is our job to understand the potential interactions between things like performance-related-target measures and the potential performance of organisations in terms of the creation of data. That is something that we are seeking to understand more clearly.
Q497 Lindsay Roy: Your colleague, John Flatley, suggested professional development was important. I take it you agree with that.
Sir Andrew Dilnot: Yes, absolutely.
Lindsay Roy: That is one response.
Sir Andrew Dilnot: It is certainly one response. It goes back to something that the National Statistician mentioned earlier. One of the attractive characteristics of the British system is that it is very detailed. A contrast is often drawn with the American system, where there are only eight principal heads of crime for which there is recorded data, whereas in this country at the moment it is close to 150 in many cases. That means the system is rather complicated, and so one can see that training is needed. That is a tricky judgment.
Q498 Paul Flynn: As the two main sources of crime statistics—the police and the Home Office—have political imperatives, do you think the holy grail of accurate crime statistics is unattainable?
Sir Andrew Dilnot: No, I do not think it is unattainable. In the Crime Survey for England and Wales, we have a good survey that gives us a good and accurate idea. On police-recorded crime statistics, I imagine the police themselves would want to say that an organisation such as that should be quite capable, subject to proper administration and auditing, of producing good data.
Q499 Paul Flynn: 18 months ago, the London’s Mayor’s Office for Policing and Crime determined three targets: a reduction in crime of 20%—not 17%, not 24%, but 20%— an increase in public confidence of 20% and a reduction of costs for the police of 20%. Do you think that is likely to be a statistically accurate way of approaching using targets like that, or just a simple way of chasing headlines?
Sir Andrew Dilnot: The job of all statisticians is to produce the data that all of their users want.
Paul Flynn: Could you repeat that?
Sir Andrew Dilnot: The job of statisticians, particularly official statisticians, is to produce statistics that the users want and need.
Q500 Paul Flynn: Their business is not to produce an objective truth; it is to serve their masters with agreeable figures.
Sir Andrew Dilnot: No, statistics should all be objectively true—I mean the range of statistics. If there is a desire for statistics on the number of crimes that are committed, then it is very important that they should be produced. I have not thought carefully about the three targets you describe. The definition of those three targets would all be relatively complicated. Let us take it away from London and the police. Once you have got targets for any organisation, there is a risk that the people who are responsible for that organisation are subject to a temptation to manipulate the data. That is why independent, rigorous audit is an absolute essential if such targets are to be taken seriously and if the achievement of such targets is to be taken to be an achievement. It is no good to achieve a target that you are able to manipulate; there is only credit in achieving a target if it is independently measured.
Q501 Paul Flynn: Is it a retrograde step that the situation we will have when the present Police and Crime Commissioners come for re-election that they will be under great pressure to present a picture of reducing crime, otherwise they are unlikely to be re-elected? Is the danger of political manipulation not increased greatly with the institution of Police and Crime Commissioners?
Sir Andrew Dilnot: I feel myself insufficiently expert on the matter of police to answer that fully, but there is one thing I will say, which is that if, through the introduction of a rigorous auditing process, we see anything like what we saw when that first happened some decade or so ago, it is quite conceivable that the police-recorded crime data could show an increase in the number of crimes recorded over the next little while, while one might see, if the trends we have seen in the Crime Survey for England and Wales continue, a decline there. It is quite conceivable—indeed, I think it might be quite likely—that if there is proper audit introduced and if the recording of crime is improved, we will see recorded crime increase, and I imagine that might lead to some interesting discussions.
Q502 Paul Flynn: We have had a wealth of evidence, from police commissioners down to police constables, and other commentators, that crime statistics in the recent past have been fictitious and we should put no credibility in them, but huge decisions have been made on these. Can you think of some way of handling crime statistics in the future that would remove the contaminated sources? Should we increase the use of the Crime Survey for England and Wales? Should we develop that in some way? Or are we still going to be taking the primary sources from groups who have vested interests in manipulating the figures?
Sir Andrew Dilnot: Now that we have removed the national-statistics designation from police-recorded crime, I imagine there will be greater attention given to the Crime Survey for England and Wales.
Q503 Paul Flynn: So they lose your imprimatur. What gets the imprimatur in future?
Sir Andrew Dilnot: The Crime Survey for England and Wales still has that imprimatur. There is a legitimate question about the Crime Survey for England and Wales, which is: should the sample size be increased such that the CSEW can give data at the police force level? The difficulty with that is that there are 43 police forces, and so the increase in the sample size that would be necessary would be very large indeed. I will turn to Mr Flatley in a minute for numbers; I think we think that would cost £13.7 million. Mr Flatley, could you say something about the costs of increasing the CSEW sample size?
John Flatley: Yes. As Mr Dilnot said, the sample size at the moment in each police force area is restricted to 650 individuals, which is nowhere near big enough to get reliable estimates at a police force area level. We reckon you would need a sample of at least 6,000 in each police force area, and given, as he said, there are 43 police forces, that is a massive expansion of the sample, at large cost.
Q504 Paul Flynn: The nightmare you paint, if I can describe it, is of a police force inside on desks somewhere trying to work their alchemy on the figures, whereas their job is to reduce crime. No one wants to go down that course. Should we just disregard the crime statistics and treat them as being fictitious, or should we allow the police to get on with their job of catching criminals and not spend their time fiddling the figures?
Sir Andrew Dilnot: My guess is that the police find—and they ought to find—that the data is a very important part of their job of catching criminals. Let me tell you a nice story about all of this.
Paul Flynn: Please do.
Sir Andrew Dilnot: I spent a happy afternoon at the Home Office six months or perhaps nine months ago looking at some very exciting real-time capturing of data from police forces into the central Home Office that was helping in working out what kinds of policing really worked—at what times of the day, week and month you needed the largest amount of available police resource. The data is not useful just for accountability and for public discussion, although it is very useful for that; it is crucial for all public organisations to be running themselves effectively. I do not think we can simply wash our hands and say, “Since we are not convinced that the police-recorded crime data is good enough, we will not bother with it”. We have to say, “We are not convinced the police-recorded crime data is good enough; therefore, it needs to be made better”.
Q505 Paul Flynn: But you cannot spin gold thread out of dross. Huge decisions—decisions on promotion and where resources are put—have been taken on the basis of untrue crime figures. This is the situation we have before us. We had Lord Stevens saying figures have always been fiddled; they have always been nonsense. What do we do about it?
Sir Andrew Dilnot: We have had Lord Stevens and others say that. My own take on what the data shows—and in the end, we are statisticians, so we look at the data—is that there was a period during the period of annual Audit Commission careful auditing of four or five years when certainly the relationship between the Crime Survey for England and Wales and the police-recorded crime data looked very different to now and looked plausible. I do not think it is hopelessly idealistic to imagine that we can not only insist but also expect that this situation can be put right. In doing so, it is quite likely that there will be a period when it looks as though the recorded crime data is rising, as we capture data.
Q506 Paul Flynn: The process for achieving the crime figures is from police forces, through the Home Office, through the Office of National Statistics, with UKSA also having an input into this. At what stages on the way are those crime figures verified? What credibility could we give to each stage?
Jil Matheson: Let me just give the headline answer and John can add more detail. The basic answer is: at each stage. The first stage is within the police, with the basic quality assurance and the role of the Force Crime Registrars—I agree with the comments that have been made about the need for a strengthened role for those Force Crime Registrars and for training and accreditation. Getting it right at the beginning is the basic step. The second stage is when the data go to the Home Office, who have the advantage of being able to see across forces. The first stage is what is happening within the force, but you can also then detect what is happening in different forces and whether there are different patterns. That is the second stage. The third stage is the work that ONS does, which is more analytical. It is not based on individual records of what individual police officers have done, but it is exactly the kind of analysis that we have been talking about: looking at the patterns, looking at the relationship, and pointing out where discrepancies appear to be and where they are changing.
John Flatley: We are dealing with aggregate data, so we can look, as Jil said, at patterns and we can look for things that look out of kilter with what we expect and we can then go back and query that, but ultimately we cannot see whether a crime has been recorded in the right category or whether it has not been recorded at all without this independent regular audit shining a light on these matters.
Q507 Priti Patel: I would like to ask the representatives from both ONS and UKSA: if you were starting with a blank sheet of paper all over again, what guidance would you give on how to resolve the issues that have been highlighted in the last hour?
Sir Andrew Dilnot: My central point would be that audit is essential. These are crucial figures—crucial to the public good, crucial to the public debate and crucial to the running of the police force. In the sorts of systems that pervade all public administrations now, where data is used to assess performance and there are often targets, the role of this kind of auditing to make sure that data is accurate and comparable is just inescapable. Had we had that, we might not have got into the muddle that we are in at the moment.
Ed Humpherson: I mentioned earlier defining—in situations where official statistics are drawn from administrative data—what good looks like. As an authority, we have a responsibility to advocate that to the people who produce statistics and the people who provide them—in this example, the police forces—so that this central importance of discipline and rigour at the police force level and of regular periodic audit is well understood, not just in the context of the police but in the context of any environment where admin data is used to support a statistic. That is the first thing. With a blank sheet of paper, I would want to be absolutely sure that we are being sufficiently risk-based in our assessments—in other words, choosing for reassessment those things that look like they might be the greatest cause for concern. The Chairman has also already mentioned being agile and being able to respond to concerns quickly as they emerge.
Jil Matheson: I have three points. I agree with the importance of independent and regular audit. I agree with the point that Mr Humpherson made that a recognition that this is important is essential. It is not an add-on, it is not a burden that can be cut and it is essential for the operation of whichever organisation is concerned. The third point—it is linked to that—is the importance of regular training and governance of this whole process.
Q508 Chair: Just finally, I have one point that presumably ONS and UKSA will be concerned about: the complaint that the procedures for ensuring accurate recorded crime figures is quite a burdensome red-tape problem for the police. How much do you think that has to be taken into consideration?
Sir Andrew Dilnot: Our code of practice makes it clear that things should be done in a way that is proportionate.
Q509 Chair: How do you take that into account?
Sir Andrew Dilnot: My own view is that the imposition of rigorous external independent audit should not be a red-tape exercise. That is not what it is. Internal audit can be a red-tape exercise, but external audit is something coming in that should not take very large and significant resources from the police. My view is that it is inescapable. There are other red-tape-related issues that we have already touched on, such as that there are very many different types of crime being recorded. That is an issue that it is legitimate to consider.
Ed Humpherson: It is really important to be clear that the systems of assurance that we have been describing—such as the regular auditing—are not add-ons simply for the purpose of satisfying the Authority for designation as an official statistic; they are operational needs to support decision making in every police force.
Jil Matheson: I was going to make the same point. I do not see recording or auditing as red tape. I do not very much buy the burden argument, particularly if it is integrated within the operation of the police force. I do accept that you get a much better job done if you make the job as simple as possible, so having systems and rules that are easy to implement is part of it.
Q510 Chair: It is not so much about simplicity; it is about where officers feel it is appropriate to use their discretion—for example, giving a young person a criminal record at a very early stage in their otherwise rather unblemished life. Do you recognise that that discretion is a necessary part the data-gathering process? How do you recognise that?
Jil Matheson: There will be discretion in judgments; of course there will. However, the whole basis of statistics is consistency and high standards—that discretion is used in a consistent way and that the way those judgments are being made is transparent.
Q511 Chair: Do you think the code of practice needs to be reviewed or rewritten in any way to allow UKSA improved access to designated data?
Sir Andrew Dilnot: We said at the outset of the Authority’s existence when the code was first published that we would keep it under review, and we certainly will keep it under review. Off the top of my head, I cannot think of any need precisely like that you have described, but we are certainly happy to take that away and think about it.
Ed Humpherson: My instinctive reaction is to say we would not want to revise the code in that direction, because we need to be assured that there are systems in place that confirm the quality of the data; we do not need to directly delve down to that level of granularity ourselves.
Q512 Chair: Do you have no concern about your access to the source data?
Ed Humpherson: I do not, no.
Q513 Chair: Could you please update us regularly on your programme of assessment and update it on your website? I think that was last published in March 2013. It might be quite helpful if it was published on, say, a three-monthly basis.
Ed Humpherson: Yes.
Q514 Chair: Lastly, should we also be having a look at the Crime Survey for England and Wales? Am I right in saying it is generally done by telephone surveys?
Jil Matheson: No.
Q515 Chair: Could you just tell us how it is done, very briefly?
John Flatley: The Crime Survey is considered to be a gold-standard survey. We have a random probability sample of the population drawn from Post Office address lists. A letter goes to that household informing them they have been selected for the survey, and then an interviewer calls at that address and carries out a face-to-face interview in the respondent’s home, capturing their experience of crime over the last 12 months. Outside the interview, we have a quality-assurance process, checking that the information collected has been correctly collected.
Q516 Chair: They are face-to-face interviews.
John Flatley: Yes.
Sir Andrew Dilnot: The response rate is 75%, which is unusually high, and it has been pretty stable.
John Flatley: It has been consistently at that level for about a decade.
Q517 Chair: How many homeless people get interviewed? Homeless people are classically regular victims of crime. How do you pick them up?
John Flatley: The survey is restricted to the population living in households. It does not cover all groups that fall victim—you mention homeless people.
Q518 Chair: What about under‑16‑year‑olds who are victims of crime?
John Flatley: Since January 2009, we have extended the survey to cover 10 to 15‑year‑olds. They have been included in the survey for about three or four years now. We do not cover foreign visitors, tourists or people living in institutions of one sort or another.
Sir Andrew Dilnot: Or companies.
Jil Matheson: There is a separate survey of businesses.
Q519 Chair: What about domestic violence, and particularly violence against women, which we know is a hidden iceberg in terms of crime? That is not picked up in the survey.
John Flatley: It is not very well measured in the main survey, but we do have an additional module. At the end of the main survey, the interviewer will hand the tablet computer over to the respondent, who will answer questions in private. That section does capture experience of domestic violence and sexual assault, which we do report on annually alongside our main statistics.
Q520 Chair: I am told that five out of seven people in accident and emergency who have clearly been involved in a crime and injured through some criminal act do not report this to the police, so it is not recorded. Would the Crime Survey for England and Wales pick that up?
John Flatley: The Crime Survey would pick that up if the respondents tell us about that and we also make use of that data. We make use of NHS data and a survey by Cardiff University of accident and emergency admissions, which help to triangulate the data from the other sources to give us assurance about the broad trends.
Q521 Chair: Do you think your household survey picks up kinds of crime like cyber and ID fraud, internet crime and industrial espionage?
John Flatley: Certainly not industrial espionage, but cybercrime is something we are actively looking at. This is a challenging area to ask about, but we are developing a programme of work and on Thursday, when we publish our next statistics, we will be issuing a discussion paper seeking views on some of the issues that we are seeking to tackle. We hope by the end of this year to have some questions included in the Crime Survey to get us a better handle on that emerging crime type.
Q522 Chair: How sufficiently do you think all the data we collect about recorded crime and the Crime Survey for England and Wales are focused on harm reduction, which is, after all, why we want to prevent crime?
John Flatley: We can be quite proud of the general system of statistics we have in this country in relation to crime. Clearly there are areas for improvement and there are gaps in statistics that we are seeking to address, but in general terms I think other countries would rival our system of statistics.
Q523 Chair: Is there anything you want to add on harm reduction as the overall objective of recording and measuring crime?
Sir Andrew Dilnot: I do not think so.
Chair: Okay. We will leave it at that. Can I thank you all very much? It has been a very informative session—perhaps not a very easy one, but I feel you have been very frank and open with us and we have learnt a lot from this. Thank you very much indeed.
Examination of Witnesses
Witnesses: Norman Baker MP, Minister of State for Crime Prevention, and Professor Bernard Silverman, Chief Scientific Adviser, Home Office, gave evidence.
Q524 Chair: Can I ask each of you to identify yourselves for the record?
Norman Baker: I am Norman Baker, the Home Office Minister with responsibility for crime prevention.
Professor Silverman: I am Bernard Silverman, the Home Office’s Chief Scientific Adviser.
Q525 Chair: Professor Silverman, could you first of all just describe to us what your role is in this matter? As a scientist, we do not think you are a statistician—or maybe you are.
Professor Silverman: I am a statistician. I used to be the President of the Royal Statistical Society before I took up my present post.
Chair: Oh, my goodness.
Professor Silverman: I am the head of Home Office Science, which is a consolidated organisation that includes Home Office Statistics. The Chief Statistician, for example, is managed by me, though he has a line of professional accountability to Jil Matheson. In general, the role of the Chief Scientific Adviser is to be an independent voice in the Department. Having the Chief Statistician under my purview stresses the point that statistics is an independent part of the Home Office’s workings, and so it is not part of one of the policy units or anything like that.
Q526 Chair: That is very informative and reassuring; thank you very much. Minister, how surprised were you when you heard that UKSA was going to downgrade police-recorded crime statistics?
Norman Baker: I was slightly surprised—not because there are not questions that have been raised about the validity of the statistics, because there have been. In fact, we have had concerns ourselves, as I will go into shortly if you wish. I was slightly surprised in the sense that we have got the HMIC analysis underway and we are due to have an interim report from them in April. I was surprised that they, in a sense, jumped the gun before that interim report was ready. Nevertheless, there are clearly questions about the validity of the statistics, so, to that extent I understand why they have taken the action they have. I would also stress that they are entirely independent, so it is not for me to make a judgment as to whether they are correct or not, really.
Q527 Chair: How much have you learnt about this question since—and I am afraid I have to quote you from BBC Radio 5 Live in December—you said, “I do not believe there is massaging of crime figures. There are strong safeguards in place to ensure that police‑recorded crime is accurate. There is no incentive for police officers to do anything other than record crime accurately. There may have been in the past when there were centrally-driven targets, but there is no incentive to do that now”?
Norman Baker: I still hold to the view that there is not an endemic problem in the sense that there is a deliberate attempt across forces in a coherent way to massage figures. If you want to ask me whether I stick entirely to those views from December, I would say that, having gone into this in more detail, I am slightly more concerned than I was in December when I made those comments.
Q528 Chair: That is very refreshingly honest. Thank you very much. Why do you think the Home Office did not really recognise the problem that you recognise now?
Norman Baker: We recognised there was a potential problem, but we could not quantify it. We recognised it because we asked HMIC to undertake this audit last July, and the Home Secretary, in her letter to Tom Winsor in June last year, if I may just quote from that one sentence, said, “It is vital that the public have access to transparent and trustworthy statistics on recorded crime. It is clearly critical a report on this is published within the new inspection period,” so the Home Office was taking action last year on that. The Home Secretary also provided an extra £9.4 million to HMIC to facilitate this exercise, and indeed their other work. Going further back, in 2011 we took action as a Department to transfer responsibility from the Home Office to the ONS deliberately to make sure that independence and transparency were enhanced. The Home Secretary only earlier this month wrote to chief constables to stress the importance of crime recording being as accurate as possible and that “any suspicion that officers are mis-recording crime figures is investigated”.
We have had concerns as a Department based, frankly, on anecdotes and we wanted to get to the bottom of this. That is why the Home Secretary commissioned HMIC to undertake this detailed audit of all 43 forces. It is also worth pointing out that we did abolish the national targets when the Government came to power. It was not the only reason for that, but part of the reason was a suspicion—and I put it no higher than that—that national targets were in some sense driving behaviour that was not necessarily very helpful.
Q529 Chair: A lot of the actions you refer to have occurred since we started our inquiry and we had the dramatic evidence from the Metropolitan Police whistle‑blower. How much would you credit this Committee for highlighting these concerns, both to UKSA and to the Home Office, and bringing about a change of attitude?
Norman Baker: I welcome the Committee’s inquiry. It is very helpful. We have in the Home Office no interest in crime figures being inaccurate; it does not help us any way to have that. We want to cut crime; we do not necessarily want to cut crime figures if the crime figures are wrong. Insofar as crime figures are wrong, that is unhelpful and inaccurate information on which to base future policy decisions, let alone in terms of chief constables deciding on the priorities.
Q530 Chair: What specific actions did the Home Office take to address the requirements of the UKSA 2011 assessment of crime data?
Norman Baker: Do you mean when they concluded that there was a green light?
Chair: They also highlighted some concerns. What action did you take about those concerns?
Norman Baker: I have mentioned some of the actions already that the Home Secretary has taken as a consequence of that. HMIC clearly went away to do further work and revisited 13 forces to dig further into that. There was an understanding that although a green light had been given, further work was necessary. The Home Office was certainly supportive of that further work. But to be fair—and I think you would recognise this—the concerns about the accuracy of recorded-crime figures have not quite mushroomed, but I would certainly say they have grown since that time.
Q531 Chair: The UKSA report says: “The Authority is also concerned that its earlier requirements, notably around the need for a rigorous audit of the quality of the data, have not been fully addressed to date”. Why is that?
Norman Baker: You will have to ask them that question, but as I understand it, HMIC undertook further work; the follow-up work was not published, for example, and the conclusion was that it was not as detailed as it might have been. That has now been corrected by the Home Secretary’s request that HMIC do a full audit. It is also worth pointing out that while that was going on, the ONS in October 2013 said in their report—and I have the exact quote here—“the system for recording crime in England and Wales by the police is widely recognised by international standards to be one of the best in the world”. That was just last October, so quite clearly we are all on a learning curve with this.
Q532 Chair: What particular lesson have you learnt from this episode?
Norman Baker: It is important to draw conclusions fully when HMIC have undertaken their audit and we are able to look at that dispassionately in the light of day when we have got full evidence. What I have concluded is that if you have targets that are there for no doubt good purposes, they can drive behaviour that is not necessarily in the interests of the police or the public at large.
Q533 Chair: I think I have to correct you on one point. I am advised that the Home Secretary commissioned HMIC’s inspection in June 2013 on crime data—
Norman Baker: Yes.
Chair—but it had already been announced in HMIC’s programme of work published in May 2013.
Norman Baker: Yes, but the programme of work follows discussions between Ministers, Home Office officials and HMIC. The letter from the Home Secretary I quoted was a formal approval of that.
Chair: We will have to make a judgment as to how accurate that comment is. It may be in your brief, but it may not be accurate. We will move on at that point.
Q534 Paul Flynn: Did the Home Secretary write to Boris Johnson in the autumn of 2012 when he declared a target of 20% for reduction in crime?
Norman Baker: I do not know the answer to that. I do not know whether Professor Silverman knows.
Professor Silverman: I do not know either, but we can find that out and let the Committee know.
Q535 Paul Flynn: You said the Government is on a learning curve. Should the Government not be convinced in what they are doing, rather than seeing the process of government as an educational opportunity?
Norman Baker: No, to be fair, whenever circumstances change or new information comes forward, it is incumbent upon the Government to welcome that new information, to absorb it, and to use it to decide whether or not existing policy, practices or processes have to change. That is perfectly reasonable. It is also worth pointing out that some of the problems we are dealing with now are a backwash from what happened in earlier times. For example—
Q536 Paul Flynn: Not the last lot—blame the last Government.
Norman Baker: I am going to—
Paul Flynn: We can go through the thing. We presume it is all the fault of the last Government, the European Union or civil servants. We take that as read, because this is the answer from Government to almost every question. We will take it that you blame the last Government.
Norman Baker: Mr Flynn, first of all, that is a grotesque description of what I—
Paul Flynn: No, it is an entirely accurate description of what happens here every day.
Chair: Mr Flynn, allow the Minister to answer.
Norman Baker: It is a grotesque description of what I have not actually even said yet.
Paul Flynn: But we know what you are going to say.
Norman Baker: I think if I am allowed say it before you categorise it as that, that might be helpful. It is the case that the Audit Commission used to analyse how crime statistics were recorded by the police. That was abandoned in 2006. That, coupled with national targets—which were introduced, I am sure, for very good reasons and with good intentions—did mean that the police took their eye off the ball in the latter part of the last decade. HMIC, when they started analysing recorded statistics again in 2010, did so in a way that was not as thorough as we think now should occur. That is one of the reasons why they are now carrying out full audits. In a sense, there has been a period when we have not seen the messages from above that having accurate recorded crime data is very important. We are now sending out that message very clearly.
Q537 Paul Flynn: Why do the Met Police and 13 other police forces have numerical targets now?
Norman Baker: We have abolished national targets.
Q538 Paul Flynn: Yes, but if you can deal with the question, you were talking about things not being endemic, not being complete and so on, yet that is the situation. The Met and 13 others still have targets.
Norman Baker: We have abolished national targets. It is a matter for individual PCCs if they wish to have localised targets. I think they will want to reflect on whether those targets drive better police performance or distort police activity when they see the reports from HMIC, but, for example, some of the targets that our PCCs have adopted relate to anti‑social behaviour. They do not necessarily, depending on what the offence is, show up in the official police-recorded crime statistics in any case, because they are summary offences.
Q539 Paul Flynn: So the Government are pulling on the rubber levers and nothing is happening in 14 police forces.
Norman Baker: No. PCCs are entitled to reach their own conclusions as to what their priorities are in their areas. They are doing that. They may wish to incentivise particular behaviour; that is a matter for them. As I say, the PCCs, as well as the Home Office, Members of Parliament and everybody else, will draw their conclusions both from this inquiry, which is very helpful, and from the detailed HMIC audit that is now underway.
Q540 Paul Flynn: Lord Stevens, who is greatly experienced in these matters, told the Home Affairs Committee that the police crime figures are fiddled and always have been. Would you agree with that?
Norman Baker: I note that Tom Winsor, for example, when he gave evidence here also used the word “fiddled”, but he also said he did not think there was—I have forgotten what he said exactly—an endemic problem across the 43 forces. There has clearly been some suggestion from a number of sources—including, Chair, evidence to your Committee—of practices that are questionable. They are being looked into by HMIC, as they should be, and any practices that prevent us from having accurate information and accurate recorded figures are unacceptable. The message we want to send from the Home Office is: we want accurate figures from the police, not figures that are convenient either for them or, indeed, for us.
Q541 Paul Flynn: On the basis of the evidence that we have, why do you think the police appear to be recording a lower proportion of crimes reported to them than in previous years?
Norman Baker: Do you mean why is there divergence from the Crime Survey for England and Wales? Is that what you are asking?
Q542 Paul Flynn: No. I am saying crimes are being reported to police and they have records of them coming in, but a lower proportion are being recorded.
Norman Baker: That is a matter that you have probably raised with previous witnesses.
Paul Flynn: We did.
Norman Baker: My answer, I hope, is the same. There are a number of factors. One of the factors is the one I gave you a moment ago, which is that when the Audit Commission discontinued its activities and they were not replaced by another body with the same rigour, or indeed at all for a number of years, that sent a message that recording crime was not as important as previously it had been suggested it was. Secondly, when you had national targets imposed, that drove behaviour in an unfortunate way. Both those matters have now been corrected. Thirdly, it is human nature that people want to present their activities in the best possible light, and that may have led to skewing of figures. You will not change human nature, but what we can do, perhaps, is put in place processes to discourage that from occurring, by taking away incentives and, secondly, by making sure that the scrutiny of recorded data is of a quality that means that is identified and stopped.
We have to identify also the difference between misreporting, which may be entirely innocent, albeit unhelpful, and deliberate under-reporting for particular reasons. If there has been deliberate under‑reporting or there have been individual officers who have undertaken their activities in a way so as to distort the truth deliberately, that is a very serious matter that we expect chief constables to deal with.
Q543 Paul Flynn: What effect do you think performance-related pay and promotion has in discouraging officers from presenting a true picture of the levels of crime?
Norman Baker: The chief-officer bonus scheme was suspended in April 2012; it was abolished in April 2013 and replaced by annual performance appraisal for ACCs. I do not think it is helpful to directly link—in a way that perhaps was done to an unhealthy degree—the official crime figures from an individual police force with the pay of an individual officer. It is not helpful to do that.
Q544 Paul Flynn: How do you see the future? You have used words like “there are problems, but they are not endemic”, which might mean that a third of the police forces are still fiddling the figures. How do you see the future? If there is a collapse of trust in police figures—and it is pretty well near that now—do you see any reforms that could be taken that would restore belief without encouraging further our policemen from engaging in mathematical gymnastics and manipulating the figures?
Norman Baker: I do not accept that there has been a collapse, to use your word, in confidence in the police figures.
Q545 Paul Flynn: Can I quote the Chairman, who said at the last meeting that every policeman he has spoken to recently has said to him, “Of course we fiddle the figures”? I think it is the experience of all of us that there is virtually no faith in the accuracy of police figures now.
Norman Baker: If you will let me finish, I will explain why I think that is not right. I do not believe there has been a collapse; there has been a dent in the confidence with which people view police figures.
Paul Flynn: It is a very big dent.
Norman Baker: But it has to be seen in the context of the comparison between the Crime Survey for England and Wales, which people have not, by and large, queried and which retains its official statistics status, and the police-recorded crime. They have diverged slightly since 2007. I have tried to explain to you why I believe that may have occurred, but essentially they are both pointing in the same direction. They both point to crime being reduced. In the last year, for example, to June 2013, the Crime Survey for England and Wales showed a decrease in crime over the year of 7%. The police-recorded crime was, I think, down 3% or 5%—one of those two. That is different, but it is not significantly different, and they are both going in the same direction. Yes, there are issues about why there has been a divergence; yes, there are issues about particular practices that have been exposed; but that does not mean people should conclude there is a collapse. Having said that, Chairman, that is just my best analysis at the moment. We will see what HMIC says when it comes out with their report. We should not pre-judge that deep analysis that is now underway.
Q546 Paul Flynn: We have heard from ONS and UKSA that they are dependent entirely on getting their raw material from the Home Office and the police, both of whom have a vested interest in manipulating those figures. What chance is there of achieving faith?
Norman Baker: I do not accept the Home Office has any interest in manipulating figures.
Paul Flynn: No, you are entirely non‑political. You have no interest in winning the next election or presenting the most favourable case. You are entirely dispassionate about these things.
Norman Baker: You must let me finish my answer, Mr Flynn, if you ask me a question. We have no interest in manipulating the figures. We have an interest in cutting crime, not in cutting crime figures. We have an interest in making sure we have the best possible data on which to base our policy. We have no interest in having figures that do not tell us what we need to know. The reason statistics are collected, ultimately, is to inform both the Home Office and the Government, and the police themselves as to what is happening on the ground and how the police should deploy their resources, what changes in the law we have to make, or whatever it may be. We have no interest in having figures that do not tell us the truth.
Paul Flynn: There is a gap between the aspiration and the reality, in my opinion.
Professor Silverman: Can I come back to what I said at the beginning? The teams that deal with these figures are completely separate from any part of the Home Office that does policy work, and so even if there is no interest in manipulating figures, it is my role to make sure the Home Office has the best possible information to do its work, and my teams do not manipulate the figures.
Q547 Paul Flynn: Why do you think the ONS have taken their imprimatur from you and humiliated you by—
Professor Silverman: They have not taken their imprimatur from us.
Paul Flynn: They have. They have taken you off the gold standard.
Professor Silverman: They have come to the view that this statistic has fallen below a line. They have drawn a very high bar, and they have said that this statistic has not quite reached that bar. I was quite encouraged by the tone of their report, because what it does is list a series of actions that can be taken to get back there again. Many of those are around documentation and explanation, not at all around the levels or the quality of the statistics.
Q548 Paul Flynn: You clearly both have a remarkable capacity for self‑delusion, I believe. Most departments, if they had been handed that rebuff from the Office for National Statistics, would have felt a little repentant and determined to do better in future.
Professor Silverman: It is not a matter of that. It is a matter of going carefully through the recommendations and making sure that they are all done. Most of the recommendations are on the ONS themselves. There is not, as far as I see, a particular rebuff of the Home Office in it, but we will be looking very closely to see where our systems can contribute to the things that the assessment report requires, and we will be working with ONS to do that. We do not see ONS as opponents; they are our partners in producing these statistics.
Q549 Chair: May I just interject at this point? I confess to sensing that a degree of complacency is still persisting in your evidence to this Committee. By saying that there has merely been “dented” confidence in these statistics that is surely underestimating the significance of UKSA’s decision to de-designate. Can I just remind you of what Professor Hough said? He said: “If they were downgraded, the system would collapse…I would be sorry to see them lose their status.” But in the Home Office written evidence, Minister, you were insistent: “We have seen no evidence to call into question the classification of these data as national statistics”. There does seem to be a divergence of views between UKSA and the Home Office about the seriousness of the state of these statistics. Do you accept that?
Norman Baker: I accept there may be a difference of view. I would not put it as strongly as you have just put it. They remain official statistics, even national statistics, if that is the difference between them. It is not one that says, “These have got no validity whatsoever”. For example, the retail price index was de‑designated.
Q550 Chair: We will not go into that. That is a different matter. There is an inconsistency. On the one hand, as soon as we held our first evidence session with the whistle-blower, you said, “We have abandoned all targets for policing”, implying, I presume, that targets are a bad thing. Have I understood your view correctly?
Norman Baker: The national targets have indeed been abandoned.
Q551 Chair: Why were national targets damaging?
Norman Baker: Because I believe they can lead to police responding by trying to meet the targets irrespective of whether that is the most sensible operational decision of the day.
Q552 Chair: Why is it okay for police authorities or police commissioners to set targets locally if they have that effect?
Norman Baker: It may not be okay. They will have to reflect on the conclusions of both your Committee’s work and, indeed, the work that HMIC are doing in looking at individual audits.
Q553 Chair: What responsibility do you have to ensure that values, and not targets, are embedded in the culture of our police in this country?
Norman Baker: We do that through our discussions with the PCCs; we do it through the College of Policing, which has a major role in co-ordinating—
Q554 Chair: Do you enjoin PCCs not to set targets?
Norman Baker: They need to look at their targets in the light of the evidence that your Committee is collecting and in the light of the HMIC report that will come out, and decide whether or not those targets are helpful in meeting the aspirations of their local populations and cutting crime, or whether they lead to decisions being taken that are not the most sensible operational decisions. They will have to look at that.
Q555 Chair: How much have you altered your view as a result of all this exposure of these problems to one that says that, in fact, the Government should take a stronger line to encourage PCCs to adopt at least intelligent aspirations rather than targets that override the values of policing?
Norman Baker: What I do not want to do is to pre-judge what is a very detailed exercise by HMIC, looking at all 43 forces. It is right that they are allowed to carry out that work, and they should do it diligently, competently and thoroughly. By the way, I think they should do it regularly as well, to make sure that these doubts do not arise in future.
Q556 Chair: But do you accept that setting targets distorts behaviour and distorts the statistics?
Norman Baker: I certainly accept that the national targets that were set previously may well have had that consequence. Local targets may or may not have that consequence.
Q557 Chair: Why are you avoiding the obvious fact that the same kinds of targets set locally are going to have the same effect?
Norman Baker: It depends what targets are set locally. There is a wide variation in the targets that are set.
Q558 Chair: What sorts of targets should they be setting and what sorts of targets should they not be setting?
Norman Baker: They may have a target that is general. They may want a target whereby they have—
Q559 Chair: What have you got to say to the 13 Police and Crime Commissioners, including the Mayor of London, who are still setting very specific targets?
Norman Baker: If they are setting numeric targets, which are different from general targets, they have to be very careful that those targets are well-founded, that they do not lead to a distortion in operational priorities and that they do not lead to police officers chasing targets rather than cutting crime.
Q560 Chair: But you think if you set such targets nationally, you could not avoid that.
Norman Baker: The conclusion we reached was that the national targets were not helpful and had too many dangers in them, apart from the fact that national targets took away responsibility from local areas, where, indeed, the priorities may be different across the country. What is important in Devon and Cornwall may not be the same as what is important in the north-east.
Q561 Chair: I appreciate the decentralisation point, but there is a very simple policy question here: how unreasonable would it be for this Committee to recommend that the Government should consider taking a generally tougher line against Police and Crime Commissioners setting numerical targets for police-recorded crime?
Norman Baker: You are right to question the setting of numeric targets, both historically nationally, before this Government took office, and, indeed, locally now by the PCCs. In terms of that particular issue in individual forces, I would want to see what HMIC says.
Q562 Chair: Professor Silverman, what would be your objective advice on this question?
Professor Silverman: The important thing is to stress that targets should not distort behaviour.
Chair: Well, they do.
Professor Silverman: The reason you set targets is to direct behaviour. Targets, in any field, are there in order to direct behaviour. As the Minister will say, the objective is to cut crime, not to cut crime figures. If targets are set and if you are looking at behaviour, you need to ask the managers to make quite sure that people are not distorting the reporting and they are genuinely directing their behaviour in the right direction. Good targets are a good idea. One does not want to say as a blanket thing, “You must never have targets”. If there is evidence that targets have distorted police behaviour inappropriately, then that is quite wrong. What I would come to in the end is that HMIC should look specifically at places where targets have been set to see whether those targets have distorted the reporting, and that ought to be part of their reporting process.
Q563 Chair: How unfair is it for me to say that that was a tellingly long answer?
Professor Silverman: You would like me to say that the Government should intervene in this way. That, honestly, would be a policy view and not a scientific view. The scientific view I would say is: get the audit people to look closely at the targets and their effects.
Q564 Lindsay Roy: Professor Silverman, you rightly said that our expectation is the highest of standards. How far have we fallen below that, and how do you know?
Professor Silverman: The entire crime statistics thing is a system, and it is a system that we are all part of. You are part of it; the Home Office is part of it; ONS, the Statistics Authority and HMIC are part of it, and so on. The system as a whole has its internal regulation. Even this process is part of aspiring towards the highest possible standards. We will never be 100% perfect. The evidence that has been shown demonstrates not just that there may have been some falling away from perfect, but also that there is a real desire to move back towards it.
Q565 Lindsay Roy: But how far are we below the threshold of acceptability?
Professor Silverman: We were told in 2013 that we had a police recording system that is the best in the world; now we are told by the same people that we have lost the National Statistics status for police-recorded crime. These two together demonstrate that we are not that far from where we should be, but we are not quite there.
Q566 Chair: Twenty-five per cent mis-recording of sexual offences in the Metropolitan Police—
Professor Silverman: No, that is not acceptable.
Q567 Chair: That is a bit worse than “not quite there”.
Professor Silverman: That is a bit worse; that is right.
Chair: Do you want to revise what you have just said, Professor?
Professor Silverman: Yes, I will revise it in that sense. The question is: is the system working towards where it should be? What I would say is that with serious offences, no mis‑recording is acceptable. In the end, we need a culture where recording and performance are seen as mutually reinforcing.
Q568 Chair: How can you be sure that you are getting accurate data if targets are being set for that data at the same time?
Professor Silverman: If the people responsible for the audit put their hands on their hearts and say the targets have not had a deleterious effect.
Q569 Chair: Good practice and, indeed, the values of the people recording that crime have to come above the targets.
Professor Silverman: Yes. Thank you. Very well put—absolutely.
Chair: So, how can you set targets as the measurement of performance of any police service? The values and the procedures are going to most reduce harm.
Professor Silverman: I am moving away again from a technical matter, but if the values and the procedures are right, then the targets will not be a negative thing. In other words, if you set a target when the values and procedures are right, then it is not a problem. That is why I cannot say categorically that targets are the problem. The important thing is exactly as you say: the values and procedures. I took great comfort from what Tom Winsor said when he said that nobody joins the police with the intention of fiddling figures, that HMIC were going to really get on top of all this, and so on. I can see there an example of continuous improvement of values and procedures.
Q570 Lindsay Roy: I am surprised you cannot give us a more quantitative response.
Professor Silverman: You wanted me to give a quantitative answer about values?
Q571 Lindsay Roy: How far below have we to go to reach an acceptable threshold?
Professor Silverman: That is not a numerical question, and I cannot give a quantitative answer to a question like, “How far do we fall short of what is perfect?”
Q572 Lindsay Roy: Qualitatively, what do we have to do?
Professor Silverman: We have to continue along the path we are on. In other words, we have to wait for HMIC to do their work, and we have to pay close attention to what HMIC says. By “we” I do not mean the Home Office; I mean everyone in the system. Attention has to be paid to the assessment report, which is itself part of the process. For example, if we look in the assessment report, there were issues such as the fact that although users were happy, the documentation of their happiness was not quite sufficient. That is not a question for the Home Office, but that is the sort of thing that will be done to get us to a position of better assurance.
Q573 Lindsay Roy: So it is monitoring the rigour of objectives and the evaluation.
Professor Silverman: It is all the different parts of the system together.
Q574 Lindsay Roy: Can I ask to what extent the status and role of Force Crime Registrars is an issue?
Norman Baker: I spoke to the Force Crime Registrar in Sussex last week—in my own force—just to go through what the Sussex procedure was. It was a very interesting meeting indeed. I was quite surprised that there is no standard job description for Force Crime Registrars across the patch, and I have asked officials to look into whether there should be a common job description. The College of Policing is going some way to having a standardised arrangement for that, but you have Force Incident Registrars who are sometimes different people and sometimes the same person, and that is a bit unhelpfully varied across the piece. Some more standardisation there might be helpful.
It is quite clear as well that Force Crime Registrars approach their jobs in different ways. The Sussex registrar I spoke to told me, for example, when you take serious sexual crimes—which I know is a particular issue in terms of recording and in terms of how many are deemed to be “no crime”—the default position, in Sussex, certainly, is that it is a crime unless evidence is provided to the contrary and it is recorded as a crime as soon as the victim says something. It can then be un-recorded if evidence comes forward that no crime has been committed, but the default position is that it is a crime. I think that is the right way to go forward; it is a victim-centred approach, which is what the crime statistics should be. In other forces, I am told by the Sussex registrar that an officer may take 72 hours, or thereabouts, to investigate whether or not there is a crime and then take a view on whether it should be recorded. That is a quite different approach. That may lead to some variations, unnecessarily, between forces. My view is that they should record it to start with and then un-record it if that is the conclusion subsequently, rather than waiting to see whether they should record it.
Q575 Chair: Because it creates an audit trail.
Norman Baker: Yes.
Q576 Lindsay Roy: Variation seems to be a real issue. Chief Constable Jeff Farrar said, “In some forces, the Force Crime Registrar regularly meets with the Deputy Chief Constable or Chief Constable. In some forces, the Chief Constable or the Deputy Chief Constable might not even know their name” or, if they walk past them in the corridor, recognise them.
Norman Baker: Yes. My view is it is very important that the Force Crime Registrar is independent, is seen to be independent, and should not answer to somebody who may have an interest in an operational day‑to‑day activity in coming out with particular crime figures. Therefore, I think it is appropriate myself that these registrars should report to the Deputy Chief Constable.
Q577 Lindsay Roy: Who has the prime responsibility for accuracy and quality of recorded crime statistics?
Norman Baker: It is shared. We have responsibility in the Home Office to collect the data from police forces; police forces have responsibility to record it accurately; the ONS has responsibility, which we have passed on to them, for the collation and dissemination of that; and the UK Statistics Authority has responsibility for ensuring the whole system works.
Q578 Lindsay Roy: Let us take the Home Office. How does it ensure the statistical data is accurate and of high quality?
Norman Baker: If there are queries that are spotted by our Home Office statisticians, then they will query those with individual forces. For example, there was a spike in the number of serious sexual offences, and that was queried with individual forces by our statisticians. The conclusion was that the spike was a genuine spike, probably generated by Operation Yewtree and all the coverage that that generated, because there were a lot of historic allegations or suggestions of crime that came in on the back of all the publicity for all these so-called celebrities who have been subjected to extensive newspaper coverage. That was an example of where the Home Office statisticians would become involved with the forces, to just check at that point, before then passing the information on to the ONS.
Q579 Lindsay Roy: Would you accept that the Home Office has prime responsibility? If so, what communication do you have with other partners?
Norman Baker: I do not accept we have got prime responsibility. As I explained a moment ago, responsibility for crime statistics is shared. We have been deliberately sharing it and passing over responsibility in some degree to the ONS, because we think it is right that there should be as much independence, transparency and accountability on these statistics as possible. Mr Flynn earlier on was alleging that the Home Office had a reason to manipulate figures, which I reject, but nevertheless, passing some responsibility to an independent body, which is not subject to any pressure at all from the Government, is one way of guaranteeing, I hope, the independence and validity of the figures.
Q580 Chair: Can I just press on this point for a moment? Where a responsibility is shared, there is nobody with prime responsibility, unless, Minister, you identify some other party as primarily responsible for the quality of police-recorded crime.
Norman Baker: We are all responsible for different elements of it.
Q581 Chair: In a way, you are a bit of a piggy-in-the-middle here. You do not collect the data; you collate the data. You then pass it on to ONS and they publish it, and then somebody else assesses it. Is this divided responsibility not part of why this has happened?
Norman Baker: No. The division of responsibility between different bodies is actually quite helpful.
Chair: It is helpful for avoiding accountability.
Norman Baker: No, that is not the purpose of it.
Q582 Chair: In a positive sense, who should we hold accountable for the failure of this data?
Norman Baker: It depends on which part of the chain the faults have arisen in. You could hold a Government accountable—and, with due respect to Mr Flynn, the last Government—for setting national targets, which have driven unhelpful behaviour. You could hold individual police forces accountable, if you wished, for the way in which they have interpreted their duties in terms of recording crime. You could hold HMIC accountable for being a bit slow on the uptake, if you wanted to. You could hold the ONS accountable, if you liked, because they said in 2013 we had one of the best systems in the world and now they appear to have a different view. It depends on which bit of the chain you want to look at at any one time.
Q583 Chair: Can we propose a novel concept here: that the Minister responsible for policing should be responsible for the integrity of the data to Parliament, so when we have the Minister in front of the Select Committee, he accepts responsibility? I appreciate that might not be the position at the moment, but if we were to make a recommendation that suggests that the Minister should take responsibility for ensuring that the police-recorded crime statistics are reliable data, how unreasonable would that be?
Norman Baker: It would be reasonable to suggest that the Home Office or the Minister of the Government—whichever way you want to put it—is responsible for creating the framework within which this whole business operates. If the framework is wrong or the Minister or the Government sets targets or does other interventions that produce unhelpful results, then—
Q584 Chair: If you are responsible for the framework, why are you not ultimately holding prime responsibility for the quality of the data?
Norman Baker: I do not think I am responsible, or the Government is responsible, for the actions of individual police forces, independent chief constables and independent PCCs. We are in a country where we have got devolution and we have, quite rightly, independent police forces, where the Home Office does not interfere on a day‑to‑day basis, so we should not be held responsible for operational matters. That would be quite wrong.
Q585 Chair: I understand that. There is a difference between operational matters and the overall integrity of a body of statistics that has failed the test, has it not?
Norman Baker: We are having the same conversation again, in a way, but there are certainly questions that have been raised and there are certainly doubts that have been generated, quite rightly. It is right that your Committee should look into those. I will want to wait and see what HMIC produce at the end of their rigorous activity, and the Home Office will look very carefully at what the recommendations are and whether those relate to us or to different bodies.
Q586 Mr Turner: What improvements do you think can be made to assessments of the validity of crime data?
Norman Baker: Let me ask Professor Silverman if he wants to comment on that.
Professor Silverman: We see the assessment reports, so we have a very good assessment of the system of crime data before us today. It is the job of the UK Statistics Authority to assess national and official statistics, and that is what they have done. Are you asking what improvements we could make to that process?
Mr Turner: Yes.
Professor Silverman: What I would say is that, having looked at this set of administrative data, as you heard in the last evidence session, they will be clearly looking at administrative data sources across Government. It is important to understand that they are shining a bright light not just on the quality of the data but on all the audit processes that go around it. That is very good, and that is part of our excellent system. I would say that our assessment process is world-leading, and I do not see any immediate improvements that are needed in it.
Q587 Mr Turner: What will you do differently in the future to ensure good crime statistics?
Norman Baker: That partly depends on what HMIC produce by way of their report. If they identify systemic failures across all forces or a large number of forces as a consequence of particular processes, for example, then those processes will have to be changed, if they are generating those unwelcome results. It really depends on what they come up with. That is why we have been very keen to engage them in this way; that is why we have provided extra money; and that is why we are determined that Tom Winsor and his colleagues should have the freedom and support from us to go through rigorously the 43 forces we have.
Q588 Mr Turner: Just walking across from Waterloo today, I saw an advertisement describing what rape is. It was meant to attract one’s attention and prompt one to say, “Gosh, I did not know I committed this crime last night”. If members of the public do not know—
Norman Baker: That is a very important point. It is not always clear whether a crime has been committed, and it is not always clear to the victim that they have been subject to a crime. This does not simply apply to sexual offences. It certainly does apply to that, and some of the work that I do in the Department on child sexual exploitation, for example, involving gangs and others demonstrates that some people do not regard themselves as having been a victim of a crime, when an objective view would suggest that is the case. You could have a situation where somebody has their driveway re‑laid by a gang of people going round exploiting vulnerable people. They may charge an excessive amount for that work to be done but the person whose drive is laid may think that it is good value and they may think the drive is all right. Is that a crime or is it not a crime?
It is not always entirely clear whether a crime has been committed. It is also not clear that we always know about crimes. For example, fraud may occur that nobody identifies has occurred. Theft from a shop may occur and the shopkeeper may not be aware that something has been stolen. There are crimes that we are simply not aware of. To that extent, we will never, ever have 100% accurate crime figures, because we cannot record everything that is in those sorts of categories.
Q589 Mr Turner: In the case of the rape, who is responsible for making sure youngsters in particular know what is and is not legal?
Norman Baker: I do not know if the Home Office is responsible, but we take it upon ourselves to try to engage with young people. We are giving a lot of time to child sex exploitation at the moment, not least through the Modern Slavery Bill that is coming forward in terms of trafficking. I have been engaging with our work on gangs to make sure that we find a way of getting to young people. Presently, we have a “This is Abuse” campaign running, which is using Hollyoaks actors and MTV and other means of communication, which are probably more accessible to young people, dare I say, Chairman, than your Committee or me, in order to get those messages across. We do take it upon ourselves to have a public‑education role, to try to help young people in particular.
There is an issue about sexual crime which makes it particularly difficult to record. We have a situation where sometimes someone who has been subject to an attack—a criminal act—does not want to report that, and sometimes when they do report it, they then want to withdraw their statement and say that no crime has occurred. The police then have to take a view on whether a crime has occurred, notwithstanding the withdrawing of the statement by the person who originally advanced it. My view is that the police should not assume, as has happened sometimes—this is what happened in Kent; we saw some of this in the Kent HMIC report—that if somebody withdraws a statement that means no crime has been committed. That is a factor they should take into account in their analysis, but it may well be the case that they should conclude a crime has been committed and pursue it as they would otherwise do.
Q590 Mr Turner: When do you expect to reach not 25% but, say, 5% unreported crimes?
Norman Baker: I do not know if I want to trade percentages or dates. What I have as a general objective is to make sure that crime reporting is as accurate as possible, and I particularly want to make sure that when serious crimes occur, they are pursued by the police with due rigour. I am not convinced that is always the case with sexual crime.
Q591 Mr Turner: How can the integrity of crime statistics be more thoroughly investigated without over-burdening forces?
Norman Baker: As I suspect your previous witnesses said, this is a matter of proportionality. You can have a system that is so busy analysing everything that nothing ever gets done. That would not be a sensible outcome. The purpose of statistics, in my view, is to inform the individual police forces as to what is happening in their areas and the PCCs as to how they should approach their operational priorities; to inform the Government, so that we can decide whether there is a particular social problem that needs further intervention—not necessarily in police terms, but perhaps through other departments, such as Education or Health—and also to give a rudimentary check on how police forces are performing. That is why the Crime Survey for England and Wales, which has not been de-designated, is so valuable: because that gives us a picture from the victim’s point of view as to what is happening. It is important to say, Chairman, that that also has gaps. It is a victim-led survey; it does not, for example, pick up possession of drugs, where there is, in a sense, no victim, and it will not pick up homicides, because clearly the person who has been murdered is not in a position to report to the Crime Survey that they have been murdered.
Q592 Chair: But more importantly, as I pointed out in the earlier session, if you are homeless, you do not get an interview, and a lot of homeless people are victims of crimes. I gather there is some progress addressing under‑16‑year‑olds. A lot of domestic violence and violence against women is not picked up. The whole catastrophe of female genital mutilation is completely invisible to the Crime Survey.
Norman Baker: You are right to say it is not complete, but you cannot get to a complete picture. We have a picture that is better than most countries have, and having two parallel statistical methods that roughly go in the same direction should give us confidence that we are not far off. That is why I rejected the word “catastrophe”, or whatever word was used earlier on about the police figures. We are now seeing 10 to 15‑year‑olds interviewed for the Crime Survey. We are doing more to pick up female genital mutilation, which is a high priority for the Government. I am working closely with my colleagues Jane Ellison at the Department of Health and Lynne Featherstone at DFID on a strategy for that. We are determined to get prosecutions for that. She is independent, of course, but Alison Saunders, the DPP, is very keen to get prosecutions on that front. Yes, there are categories of people, including the homeless, who are not picked up in the Crime Survey; that is terribly true.
Q593 Chair: We must draw to a close, Minister. The last point I want to raise is about what occasioned this inquiry, which, by coincidence, was a constituent of mine, PC James Patrick, coming to my surgery and alarming me sufficiently for me to put it to this Committee we should do an inquiry into crime statistics. His evidence was extremely compelling and seems to have changed the course of political discourse around this issue. What discussions have you had in the Home Office about how the Metropolitan Police have treated this whistle-blower?
Norman Baker: I understand that PC Patrick is subject to a disciplinary hearing. That is a matter for the Met. Insofar as PC Patrick made allegations or suggestions about how crime figures were not accurate, Tom Winsor has said that he has taken on board those comments and will factor them into his work with the 43 forces, which is quite right.
Q594 Chair: When the chief of the Metropolitan Police first responded to the evidence session at which PC Patrick gave us his evidence, he said he wanted to meet and talk to PC James Patrick. When he came before us last week, he said he had not done so because of the disciplinary hearing. That is understandable, but how happy is the Home Office that this thoroughly credible whistle-blower, in our view, still is not being heard openly and dispassionately at the top of the Metropolitan Police Service?
Norman Baker: His comments have been heard, not least of all because of your intervention, Chairman, and your invitation to him to give evidence. As I say, Tom Winsor, as I understand it, is certainly factoring into his work PC Patrick’s comments, and it is entirely right that that should occur. As you recognise, there is a disciplinary hearing under way; it is not for me to get involved in that, as you will appreciate.
Q595 Chair: Indeed, I am refraining from commenting in any way that might prejudice that matter, because we should treat it as sub judice, but one of his complaints to us is that when he was unable to pursue his concern about the quality of police-recorded crime in the Metropolitan Police, he was unable to persuade his line management to make a complaint on his behalf to the IPCC. Have you given this matter any thought?
Norman Baker: Yes. I welcome that question; it gives me a chance to make this comment. All police officers are required to challenge and report any conduct that falls below the standards of professional behaviour. I expect police officers to do that, and I expect others, when they hear those complaints, to act upon them and take them seriously. All of the forces have professional standards departments, to which officers can raise concerns. The Independent Police Complaints Commission also runs a confidential hotline for police officers to contact them with concerns. I welcome the fact that if police officers have concerns about this—or any other matter, for that matter—they do exercise their duty and report any conduct they believe to be inappropriate.
Q596 Chair: But my understanding is that a police officer cannot make a complaint to the IPCC about his own force.
Norman Baker: Apart from anything else, the IPCC can take anonymous complaints.
Q597 Chair: So you are saying that it is now the policy of the Government to allow police officers to make a direct complaint to the IPCC.
Norman Baker: If a serving officer makes a complaint anonymously, then he or she makes a complaint anonymously.
Q598 Chair: Why should they have to make it anonymously?
Norman Baker: I am not saying they do; I am just saying that it is one of the answers to your question that they can do so.
Q599 Chair: To what extent is this a change of Government policy?
Norman Baker: I am not sure it is a change of Government policy. I want police officers who believe something inappropriate is occurring to report that. They are required to challenge and report any conduct that falls below standards, and I hope they will do so. There are at least two ways of doing so. One is through their own forces and one is through the IPCC. In my view, if an officer honestly and sincerely believes something has occurred that is inappropriate and reports that, they should be respected for that.
Q600 Chair: Am I incorrect in saying that a named police officer cannot complain to the IPCC about his own force?
Norman Baker: I am not aware that there is a problem in that regard. If there is, and I am subsequently advised by officials that that is the case—
Q601 Chair: How concerned are you that PC Patrick believes that he was prevented from making such a complaint?
Norman Baker: I do not believe any officer should be prevented from reporting any behaviour that they regard as being worrying or inappropriate in their force, either to their professional standards department or the IPCC.
Q602 Chair: I have to say I am very confused by the answers you are giving, because they are not consistent with what we have understood before and yet you are saying that there has not been a change in Government policy.
Norman Baker: I am not aware there is change of Government policy. I am personally relaxed about the idea that police officers—
Q603 Chair: Well personally, what is the Government’s policy? Is it okay for a named police officer to complain to the IPCC about their own force?
Norman Baker: I am satisfied that they should do so. If the Home Secretary has taken a different view, I will write to you subsequently and tell you that.
Chair: That would be very helpful. Thank you very much indeed. It has not been an easy session, but you have been very frank with us. I do not think we are going to agree with everything you have told us, but I very much hope the Government will take our recommendations seriously.
Norman Baker: We will certainly look at them very seriously, Chairman.
Chair: I am most grateful.
Oral evidence: Crime Statistics, HC 760 2