Written Evidence from the Association of Police and Crime Commissioners [CST0016]
Introduction
The Association of Police and Crime Commissioners (APCC) is the national body which supports Police and Crime Commissioners (‘PCCs’) and other local policing bodies. This document sets out the key views that have been expressed by PCCs and their officers. Six PCC offices responded through us to the questions set out by the Inquiry (although it is possible that others responded direct without copying us in). The following overview summarises the key comments and issues they raised, and their individual responses are set out in the Annex submitted with this document.
Background
The Public Administration Select Committee (PASC) will no doubt be aware of some of the key background material to this Inquiry, including the Home Office proposals and consultation on crime outcome, the UK Statistics Authority consultation on aligning the Code of Practice, and recent HMIC Inspection reports on police crime recording – the most recent ‘The Crime Scene’ dating to January 2012. This latter report in particular has guided some of the answers set out below in relation to the understanding of the accuracy of crime data. We believe that HMIC will shortly start a Data Quality inspection, which will look again at crime recording data in forces and may provide some interesting insights into some of the questions posed by the Inquiry.
Consistency and comparability has emerged as a key common issue raised by the PCCs that responded. This is a complex area and there are anomalies caused in part by the system and in part by the nature of statistics. In the case of the former, for instance, Anti-Social Behaviour incidents do not count as crimes under the rules, and the line between ASB, public disorder and low level violence is often blurred. There is also some local flexibility in how rules are applied (for instance the time during which a crime must be recorded), which can also cause some anomalies. There are a number of caveats that apply to statistics and raw data no matter how accurate it is, such as the need to interpret and triangulate with other evidence, before it can be coherently understood. These issues are explored in more detail below in the context of the questions set out by the PASC.
Specific Points
Q1 Are crimes being recorded by the police when they should be? Are crimes being categorised correctly?
The general feeling which PCCs expressed, based on internal and external checks, was that crimes were being recorded and categorised correctly on the whole, within their areas and within the limitations imposed by the system and the nature of statistics themselves. However, human error and a number of the other issues addressed at Q2 do lead to discrepancies, but as noted at Q3, there are a number of checks and balances incorporated in the system already to try to ensure that recording and categorisation is accurate, although the nature of checks and balances can vary. Taken together this means that there are some inconsistencies nationally and some dangers inherent in direct comparisons of raw data. Some PCCs are also concerned that budget cuts might impact on resources dedicated to quality control, and some PCCs feel that the whole system is too process based and does not taken enough account of victims’ views and needs, or allow for a common-sense approach.
Q2 What are the factors which can influence police mis-recording of crime?
A number of factors have been suggested by PCCs. One of the key issues identified is the short time slot allowed under the rules for officers to record an incident as a crime, which puts pressure on them to make a judgement, without sufficient time to consider all the circumstances, and especially the victim’s wishes.
Other factors suggested include human error, subjective interpretation, complexity of HO Counting Rules and lack of knowledge (particularly in relation to rarer types of crime, or recording of multiple crimes arising from one incident), the bureaucracy of the process itself (which can result in short cuts being taken), organisational culture (particularly around compliance issues, such as reluctance to challenge fellow officers), and level of scrutiny. Several respondents also noted that targets can encourage perverse incentives, but also noted that the trend generally both from Government and PCCs is to move away from targets which has helped the problem of ‘gaming’ to meet ‘quotas’. But there is still some feeling that the general pressure to show reduced rates of crime might still be leading to some down-grading of crime incidents to non-crime incidents, such as ASB.
Q3 Are the right checks in place to ensure that the systems for recording crime function effectively and accurately?
There are a significant number of quality control checks already in place, although the nature and depth of scrutiny can vary from force to force. However, typically as a minimum, these would include internal audit, crime registrar, external checks, such as periodic HMIC inspection, triangulation with other data, and a degree of oversight from force management board and PCC. Much depends on the quality of the audit, although some forces also include additional checks, such as external audit, dip sampling of crime recording and categorisation examples, additional levels of internal control process, or IT systems which flag erroneous data.
Greater transparency can also play a role in raising standards and highlighting areas of concern – although it should be noted that there remain concerns about the extent to which data is truly comparable, for reasons already noted, and whether comparison of this nature is desirable, given the more local focus of PCCs. Ultimately, it is for PCCs to hold the force to account (through the Chief Constable), and they will be concerned to ensure that crime recording meets local needs and expectations (see Q6 below).
Q4 Has enough been done to ensure the integrity of crime data? What more should be done?
There is a balance to be struck between accuracy and the cost of quality control mechanisms, and some PCCs raised concerns about whether existing checks are sustainable in the current financial climate. The anticipated HMIC Data Quality Inspection should throw some interesting light on what more could be done, and some PCC respondees suggested that new mobile technology may help, or perhaps more training.
However, a more fundamental debate may be needed about whether discrepancies from place to place are an inevitable consequence of greater local accountability and responsiveness to victims (provided the differences this generates are transparent, understood and accounted for), or whether consistency and comparability are paramount, which would involve greater centralisation, stricter rules and probably higher costs flowing from the greater bureaucracy involved.
Q5 To what extent can policy-makers have confidence in the statistics which result from the recording of crime by police forces?
Most PCCs who responded felt that some confidence could be placed in the data (given its limitations), and that it was a key factor in how they held their force to account. They acknowledge there are problems with comparability (for reasons already discussed), but at least some PCCs felt that the key issues was whether it met the needs of local communities, not necessarily national policy makers.
There was some feeling that the real question here was whether the public had confidence in the data, and whether something should be done to improve their perception of statistics. There was also some concern that there is anecdotal evidence that some crimes, including serious crimes are not reported by the public and some thought is needed as to how communities can be encouraged to fully report crime.
Q6 Should recorded crime statistics be classified as National Statistics?
There were mixed feeling on this amongst the PCCs that responded. Some thought they should as it would help with perceptions of independence, one thought this could only really work for long term trends, but others thought this should be substantially a local issue and therefore not a National Statistic.
Annex
TO APCC Submission to PASC Inquiry into Crime Statistics
Individual Responses from PCCs
A. Devon and Cornwall
Please find below our responses to the questions posed by the above inquiry.
The fundamental dilemma highlighted here is the focus on accuracy rather than reliability. In answer to the questions posed (i.e. is the data accurate?) the response would be that most forces endeavour to meet the National Crime Recording Standards and the Home Office Counting Rules and adhere to high standards of ethical recording practices. However the more pertinent question seems to be whether data can be compared between forces with confidence (i.e. is the data reliable). The answer to this is almost certainly not. There is a degree of flexibility in the system currently that whilst recording in accordance with the guidance set out in the National Crime Recording Standards and within agreed time frames there are significant differences in how the resulting data is interpreted. For example with regard to when a crime is recorded, in Devon and Cornwall we direct record (as soon as an offence is disclosed it is recorded on the crime system and is not necessarily recorded first on the incident management system). This means that if we subsequently discover that no offence was committed we have to provide evidence that no offence took place before we can no-crime. In terms of comparison to other forces this means that not only do we have a higher no-crime rate than might otherwise be expected but also that more offences remain on the record than might be the case for other forces who conduct some early investigation to corroborate the occurrence of an offence. In addition how each force interprets the NCRS can vary slightly but significantly, for example the line between anti-social behaviour (a non-crime incident), public disorder (an offence against society) and low level violence (a victim-based crime) is very blurred.
Politics – an issue or offence that is high on the national or local political agenda may influence the tendency of either individuals or organisations to record an offence in a certain way. This is not to suggest that figures are deliberately manipulated but rather that there is a human tendency to want to conform.
Scrutiny – whether local performance frameworks or HMIC inspections, scrutiny can influence compliance. In addition to the potential for perverse incentives through performance frameworks there is also potential for forces to adopt recording practices that are not necessarily compatible with other forces following an adverse inspection.
Human error – whether systematic as a result of competence or the result of a distraction or an off-day the police crime recording system is heavily reliant on people, their abilities, attention and attitudes, and as such is vulnerable to associated limitations.
Organisational culture – an organisation that promotes integrity, personal responsibility and has a sound supervision structure in place is likely to achieve a high level of compliance and to identify non-compliance issues earlier.
3. Are the right checks in place to ensure that the systems for recording crime function effectively and accurately?
Current checks include internal force audits, external inspections, publication of data, the Crime Survey for England & Wales and crime comparator systems. The Police & Crime Commissioners have a role to play in monitoring trends and holding forces to account where the information emerging from existing checks indicates an issue relating to recording. What seems to be missing is any means of identifying and accounting for the intrinsic differences in how forces record offences. Interestingly the questions posed do not address explicitly the recording of outcomes against offences, which potentially raises further issues.
4. Has enough been done to ensure the integrity of crime data? What more should be done?
A balanced approach is needed to ensure that the data is as accurate and reliable as possible, and that limitations in the data are openly recognised and as far as possible accounted for without compromising operational delivery or imposing a significant financial burden.
There are a number of potential solutions to the issues raised. Perhaps the most extreme would be to consider centralising crime recording such that forces record and manage incidents from the incident log but a national or regional team would determine whether the incident was a crime and record this appropriately. An even more robust recording framework, mandating the point of crime recording and outcomes might offer some solution however it runs the risk of becoming operationally unwieldy and unworkable. A further alternative is to accept that there will always be difference and instead to identify a means of measuring more reliably, and accounting for, variability in compliance between forces.
5. To what extent can policy-makers have confidence in the statistics which result from the recording of crime by police forces?
This question speaks more to how the data is used and interpreted than to either the accuracy or the reliability of the data. Police recorded data in isolation is often not the best data source and even police forces will use other available data sources including other partners and proxy data to identify and address policy issues. The confidence of policy-makers needs to be in the capability and competence of analysts to use the data appropriately, to recognise its limitations and to address these through other valid sources.
6. Should recorded crime statistics be classified as National Statistics?
It isn’t entirely clear what this would deliver to address the issues raised. If this would add weight to forces to improve compliance or support a more robust recording framework then there might be some value.
B. Dorset
The Committee is interested in receiving responses to the following questions:
a) In my view yes to the first question. The NCRS has become more and more proscriptive, and Force Registrars and HMIC regularly audit the crimes. There is a gap between ASB reports and recording, but as these are not technically crimes, I think the Forces in general comply well. I know Dorset has, I scrutinise this closely, and HMIC have praised the Force in 2011. However, the cuts are reducing governance and oversight staff, so this could get worse.
b) The second question is the same, once the call enters the system, various balances and checks exist to ensure categorisation is appropriate. One thing that does need sorting is the “72 hour rule” that most Forces adopt to serious sexual offence allegations. Research shows that what is reported at 3am on a Saturday night as a rape is often downgraded in the cold light of day, hence Forces adopting the 72 hour rule. I think this should be stopped, BUT it will lead to a massive increase in rape allegations that, when scrutinised later, are actually indecent assaults.
a) Lack of knowledge of the hundreds of crime categories, and failure to embrace more than one offence. For example, an ex partner harassing a partner could be DA and harassment, but only one or the other would be recorded. Likewise with trafficking of humans, often the different offence in relation to CSE, forced labour or prostitution is recorded, not embracing the bigger picture. I have dealt with a case where a constituent had her foot run over by an offender in a parking dispute. The police recorded this as an RTA, when it was in fact an assault. And RTA v Assault brings lots of issues for the victim, especially in relation to disclosure and compensation.
b) Lack of time, and lack of correct oversight, as the cuts bite, Crime Registrar staff and CMU supervisors are being cut
Yes, please see above
I am happy the checks and balances are in place, but we cannot be complacent.
Pretty well, except for the overriding issue of – What happens if crime isn’t being recorded. I remain convinced that lack of faith in police action and lack of faith in the 101 service leads a massive amount of crime being unreported.
Yes they should.
C. Humberside
Many thanks for affording me the opportunity to submit written evidence in relation to your inquiry into crime statistics.
This inquiry is timely as Police and Crime Commissioners are one-year into their roles in holding their Chief Constable to account for the delivery of policing on behalf of the public. This inquiry must consider the impact of this significant development. It should take into account the improvement in local accountability, and the transparency, contextualisation and timeliness of information to the public by Police and Crime Commissioners, especially when compared to the time-lag in publishing national statistics which generally have little in the way of local context.
Question 1: Are crimes being recorded by the police when they should be? Are crimes being categorised correctly?
We have all seen the continued reductions in recorded crime over the last several years and the ‘gap’ between recorded crime statistics and the crime survey results. This ‘gap’ is arguably one of the reasons behind some of the dissatisfaction we hear from the public about published crime statistics.
Locally, this ‘gap’ has been relatively small and consistent over the years due to the ethical approach adopted locally, following the National Crime Recording Standard (NCRS) to the letter.
What has not happened over this period is a proper and informed debate about the victims ‘needs’ and a more common-sense approach to recording for certain crime categories. For example, there is a tendency to ‘crime’ criminal damage or the more minor ‘violent’ incidents when the victim may merely want some form of restorative solution and does not feel a crime has been committed. Police officers are not always afforded this kind of discretion and the NCRS rules sometimes take an element of this away from them.
Locally, we have tended to take a risk-averse approach to the recording of crime. This has arguably been a contributory factor in our higher levels of crime when compared to our peers. This makes comparison difficult, but also raises the question over whether it is possible or even necessary to have national comparative statistics, especially as the focus of Police and Crime Commissioners is predominantly on local provision. It also raises the question over whether a national standard for crime recording should continue to be viewed as achievable or even necessary.
The other thing we have to remember with recorded crime statistics is that whilst the ‘traditional’ crimes (burglary, vehicle crime, and so on) are falling, fraud is replacing such crimes as the offence of choice for an increasing number of offenders. The clear message needs to be that whilst other reported crime decreases, fraud in the UK is growing. We also know that a large volume of fraud remains unreported, due to a number of factors such as victimisation of vulnerable and hard to reach groups, and we do not know the full scale of the threat and its impact on our local communities. By focusing on ‘traditional’ crime recording, we are also in fear of putting our resources not in the places where they are most needed and will reap the greatest benefit to victims and our communities.
It terms of the categorisation of crimes, my biggest concern is in relation to elements of violent crime. The category of violent crime without injury can at times be a ‘catch all’. Although it is right that we somehow capture minor assaults, as they could be an indicator of escalating activity by a person, whether we should ‘mix’ these incidents in with all violent crime is questionable.
There is a further anomaly around violent crime. Locally we are viewed as having one of the highest levels of recorded violent crime and yet a very low homicide level. This undermines the validity of the data to some extent and an overall high rate of violent crime doesn’t display the local context, where we don’t have a culture of more serious and gun crime that you see in other areas of the country. People make the incorrect assumption that they live in a violent place according to the statistics, whereas the reality is somewhat different in most parts of Humberside.
Question 2: What are the factors which can influence police mis-recording of crime?
Following on from Question 1, they key issue is the time-frame for criming.
I believe we need to question whether the police are sometimes too quick to crime an incident before taking into account the victims wishes and feelings. The NCRS rules have to some extent taken away the common-sense element of policing. There is no doubt in my mind that recording rules are a necessity, but we need to look potentially at a ‘cooling off’ period for certain crime types and the ability for discretion by police officers to take what they consider to be an appropriate response (which might not include criming), whilst ultimately record what is right after taking such time and accounting for the victims needs.
The removal of the target culture has undoubtedly helped in reducing the possibility of ‘gaming’ by officers categorising crimes incorrectly, e.g. criming a dwelling burglary as a theft from dwelling when they may have reached their ‘quota’ of dwelling burglaries for the month. This type of activity must be discouraged and the removal of a target culture helps in this respect.
In addition, the over reliance on quantitative data combined with simplistic and inappropriate centrally set targets in the past has, in my opinion, corrupted policing and led to perverse outcomes, devoid of common sense that has undermined public confidence.
Question 3: Are the right checks in place to ensure that the systems for recording crime functions effectively and accurately?
The Police and Crime Commissioners are there to hold their Chief Constable to account on behalf of the public.
The question we need to ask is whether public perception of crime recording reflects the reality. It is important that systems are in place locally and that they are effective for local circumstances.
In Humberside, as part of my Police and Crime Plan, I will be rolling out mobile technology to ensure officers are visible in our local neighbourhoods and in the places where they are most required. We need to recognise the risks in rolling-out such technology, ensuring the correct impact on crime recording, consistency and the use of common-sense by officers to individual situations.
I am aware of the work by Her Majesty’s Inspectorate of Constabulary (HMIC) around data integrity. This relies on a limited sample and doesn’t fully take into account the work undertaken by local Crime Registrars in their audits, findings and learning points. This local element to me is important to ensure that our systems locally are as effective and accurate as they can be. We need to create an atmosphere where we do things right for the victim in the first instance and correct any issues quickly if things do not go as expected.
I will give an example where it could be questioned whether crimes are being recorded accurately. Locally we have a high incidence of criminal damage (a crime) and a low incidence of anti-social behaviour (an incident). It could be argued that we have taken a risk-averse approach to recording, hence the high-level of criminal damage. Less risk-averse approaches could produce a more ‘balanced’ level of crimes and incidents. Our approach indicates that we could be over-recording criminal damage, although our data integrity audits show that NCRS compliance is excellent, i.e. we’re not over-recording. But this raises two questions: (i) whether others are following a similar line and (ii) the comparability of figures between forces.
Question 4: Has enough been done to ensure the integrity of crime data? What more should be done?
See answer 3 above.
Question 5: To what extent can policy-makers have confidence in the statistics which result from the recording of crime by police forces?
It is important for Police and Crime Commissioners to have confidence in their statistics locally in order to hold to account the Chief Constable on behalf of the public. This is important for me at a local level.
I am uncertain as to the level of confidence within other police forces and the ability for national policy-makers to have confidence in this data. However, the duties of the Police and Crime Commissioner are clear, and if they are undertaken adequately then confidence in the data should be high.
Question 6: Should recorded crime statistics be classified as National Statistics?
It is important that the statistics are accurate at a local level as part of my statutory duties. The focus on local solutions through local statistics must be the way forward in the first instance.
I believe that the National Statistics can be unrepresentative of the local picture. In addition, the Office of National Statistics (ONS) now tend to produce quarterly figures which are devoid of context and tend to merely restate those figures presented to them in the first instance. This is due in part to the risk-averse nature of crime statistics publications. However, the ONS has produced guidance on the publication of local data which is useful to follow.
We also need to bear in mind that we should all move away from the current culture of the recorded crime data being seen as the ‘result’ which is then often used as the stick. We need to move to a place where the figures are used as ‘guidance’, for what they tell us as part of an overall holistic assessment of performance. This includes a mixture of local qualitative as well as quantitative data.
Finally, we need to publish statistics with context and in a manner which incentivises people to reporting and/or forces to record. The clearest examples are the increasing rates of sexual crime and domestic violence that are evident throughout many parts of the country at present. Such rises are depicted by the media and others as failure, whereas what they actually demonstrate is success through partnership working and the increase in the confidence of victims to report the crime by the variety of means we make available to them.
The more serious crimes tend to be grossly underreported and we must make all efforts to encourage reporting of such crimes.
D. Merseyside
18 forces correctly recorded crimes in 95% of cases or more;
15 forces correctly recorded crimes in 90% to 94% of cases; and
11 forces between 86 and 89% of cases.
2. What are the factors which can influence police mis-recording of crime?
3. Are the right checks in place to ensure that the systems for recording crime function effectively and accurately?
4. Has enough been done to ensure the integrity of crime data? What more should be done?
5. To what extent can policy-makers have confidence in the statistics which result from the recording of crime by police Forces?
6. Should recorded crime statistics be classified as National Statistics?
E. Surrey
Our comments on the questions posed on crime statistics:
1. Generally police are recording crime when reported. The main area of current concern we have is over cybercrime. Fraud cybercrime has been taken out of police crime statistics, leading to an impression of crime going down. But we know this is a large growing area of criminality often going un-reported. We need to ensure that Action Fraud data is included in crime statistics in the future.
2. Setting targets on specific crime types can influence mis-recording – potentially leading to mis-categorisation of crimes e.g. recording burglary as criminal damage in order to meet crime reduction or detection targets. Generally the move by many PCCs away from targets should lead to more ethical behaviour and better integrity in crime recording.
3. Forces all have force registrars and internal audit processes, but little is currently done on ensuring that there is consistency across forces in how they interpret Home Office counting rules and record data. However, data is compared in national reports so there should be consistency.
4. We need to balance bureaucracy involved in audits and checking with the benefit given by confidence in data. In recent years there has been a lack of audit and checking but hopeful that the forthcoming HMIC inspection into crime data integrity will provide information on data quality.
5. Crime data provides one data set that can be used with some confidence by policy makers, bearing in mind the comments about on lack of quality assurance on consistency. But crime data should be put into context with other information such as the crime survey of England and Wales, changes in Home Office counting rules which affect the data and other demands on policing when making policy decisions.
6. Yes
F. Thames Valley
As PCC for Thames Valley, I am reasonably assured that the Force (TVP) is recording crime in an appropriate and ethical manner. The evidence for this is founded on two key sources. Firstly, HMICs thematic review (late 2011) of TVPs crime and data quality. This showed that the Force had the best accuracy rates for crime recording. Additionally, TVP was the only force in the country with 100% accuracy in relation to ‘No Crime’ decisions. HMI concluded that ....‘The force has high standards when recording crimes and incidents in a consistent and accurate manner to correctly reflect the sequence of events as described by victims’.
Secondly, an Internal Audit of TVP’s Performance Management function was carried out in April 2013. The key controls for the collection and recording data were given full and/ or majority assurance.
Pressures to reduce crime rates could act as perverse incentives for some crimes to be downgraded, or, downgrading ASB related incidents (which are not captured in data returns to the Home Office).
The move to Neighborhood Policing may have also led to more low level crimes being dealt with informally and outside the formal crime recording system.
3. Are the right checks in place to ensure that the systems for recording crime function effectively and accurately?
As mentioned above, both HMIC and our own OPCC Internal Audit Section undertake periodic reviews of the Force’s performance systems – this assurance informs our confidence of TVP’s crime recording systems.
However, it would be short-sighted not to recognise some of the wider issues affecting crime recording systems. Anecdotal evidence across wider policing circles suggests that league tables and performance targets have created a culture of ‘Gaming’ – manipulating crime data to meet targets and measures.
External checks need to better focus on Police culture, particularly around high ethical standards to ensure practices are conducted in a consistent manner. Other key areas are the skills and training of staff, backed up with sound quality assurance processes.
Earlier this year, the Kent PCC commissioned HMIC to examine Kent Police’s crime and resolution figures to ensure that local people had confidence in the accuracy of such figures. Findings from this inspection found that the force had under recorded crime, and, that crime was not being resolved in accordance with national requirements.
There need to be further examples in which HMIC ‘name and shame’ Chief Constables/Forces where similar sharp practices may be happening elsewhere. Failure to tackle this issue may, at best, suggest not enough expertise within HMIC to shine a light on such practices; at worse, highlight a more cynical complicity in this type of culture.
5. To what extent can policy-makers have confidence in the statistics which result from the recording of crime by police forces?
Policy makers do have confidence in the statistics – they are able to triangulate quantitative and qualitative sources of data/information to create an informed picture. Statistics in isolation can be misleading and can be used incorrectly. Great care needs to be used when quoting statistics to ensure the context is properly explained.
Perhaps the key question here is not whether policy-makers have confidence in the statistics, but ‘do the public believe the statistics they are being told? (Difference between the perception of crime and the actual crime).
6. Should recorded crime statistics be classified as National Statistics?
Yes, classifying crime statistics as National Statistics would go some way to demonstrate better independence and more effective governance and accountability (as displayed by the Home Office who transferred this function over to the ONS).
However, given all government departments are operating in a climate of challenging austerity measures, there is a counter-argument which questions what the true cost-benefits of ‘National Statistics’ are. Who benefits the most: policy makers and statisticians, or the public? Are the public interested - how can they be better engaged in such matters? If you were a victim of crime, what would matter the most to you?
November 2013