THE TRUTH ABOUT CRIME STATISTICS

Chris Lewis, Institute for Criminal Justice Studies, University of Portsmouth 1

Introduction

Crime statistics seem to be discussed more than figures for other social phenomenon. Some feel they are
little more than government propaganda. Some feel the police count what they choose to and put their
faith in the figures obtained from surveys. Some look to the ‘golden past’ when figures were much lower.
Some look to figures for other countries and take reassurance that English figures are lower.

Others avoid using figures at all and look at their own experience. When did they last lose anything to a
robber, thief or burglar? When was any friend of theirs last attacked?

Most people rely on the media for their information. Few actually read the statistics themselves, published
in great detail on the internet2. Most rely on highly summarized versions from their favourite source. This
confuses further the message from crime statistics. Different media sources often have agendas which
drive the particular aspects of crime figures that they choose to comment on. Nearly all current media are
characterized by their reliance on ‘sound bites’.

Hardly anyone appreciates the complex mix of data collection methods, technical input and expert advice
that lie behind crime figures. Probably most people are just thoroughly confused about them. This is made
worse by the recent economic crisis that has changed the way common statistical series are behaving, not
simply in the area of crime. Although there has been virtually no growth for the last five years, crime
figures continue to fall, unemployment has not risen as expected and the stock exchange is booming!
Experts cannot understand what is happening so how can the average person expect to?

This paper attempts to improve public understanding of crime figures. It is based on experience, as the
author was in charge of crime statistics at the Home Office for 14 years in the 1980s and 1990s. It takes a
historical perspective, looking back over the last 60 years and forward for a further generation. It takes a
technological perspective as crime is constantly changing with the development of new technologies 3,
goods and services. It also takes a crime prevention perspective as it acknowledges that some recent
reductions in crime have followed improvements in crime prevention.

What is crime?

Many of the problems with crime statistics rest on this basic question. Most people feel that the answer
intuitively. But defining what is a crime is not an easy question, as any policeman will know. And it is
further confused by the next question: what sort of crime should be recorded in a particular case?

An example will help to develop this point.

1
Address for correspondence chris.lewis@port.ac.uk. Chris was Chief Statistician at the Home Office in charge of Crime
Statistics from 1983-1996 and continues to research, teach, and write on the topic.
2
The most recent figures at the time of writing were available in a 106-page Statistical Bulletin from the Office of National
Statistics at http://www.ons.gov.uk/ons/dcp171778_307458.pdf accessed on 15 May 2013
3
This has been highlighted most recently by the report of the Home Affairs Select Committee on E-Crime, Fifth report 29 July
2013 downloaded from http://www.publications.parliament.uk/pa/cm201314/cmselect/cmhaff/70/7002.htm

1
It is generally accepted that killing someone, homicide, is a crime. But not all homicides are the same.
We all agree that killing a policeman in an armed robbery is a crime: but was the killing of Osama Ben
Laden a crime? Is assisted suicide a crime? Is a doctor who performs an abortion a criminal? Or a
doctor who gives more morphine to a terminally ill patient than is strictly necessary to alleviate the pain?
Does a rapist whose offence leads directly to his victim committing suicide then become a murderer?

It is clear that there is much scope for differences of opinion. In practical terms, the relatives of dead
person have to decide whether to report the death as a possible crime: police have to decide whether they
will investigate the death as a homicide and include it in the crime statistics: prosecution have to decide
whether there is enough evidence to take the case to court: the judge and the jury will also have their
views. That is why we have the body of law, including common law, statute law and precedent of
previous cases to assist law enforcement officers as to what to do in a particular case.

Law and practice changes from time to time: to take another example:

Up to 1991 it was impossible in England for a man who had forced his wife to have sex when she was
unwilling to be prosecuted for Rape4. This was based on the accepted belief that a wife gave up the
possibility of any such accusation when she married. The House of Lords 5 changed this law in 1991.
Because most Rapes occur between couples who are, or have been in close relationships, this change has
caused the number of reported Rapes to increase dramatically over the last 20 years. Police will now
investigate such a report, whereas before 1991 they would have said no crime had been committed.

More complications arise when it is clear that a crime has been committed but not exactly what crime it is
or how many crimes there were. Again let us take an example:

Considerable publicity has been given to recent cases of sexual exploitation of young teenage girls. 6
Everyone agrees these were horrendous but how many crimes are actually committed when cases go on
for many months or years and involve a number of offenders and victims. Most people will sensibly say
that the actual number of crimes is not particularly important. But we need to know whether the number
of such cases is rising or not. Therefore there is a need for a clear set of rules to record the number of
offences in complex cases: rules every policeman in the country knows about.

What is a crime is not defined in a natural way, such as weight or height or numbers of people living in a
town. Crime is what the authorities of a country decide to count. There is a series of rules for counting
crime and this can change from time to time and differ from country to country. In the case of England
and Wales they are decided by the Home Office after consultation with all relevant stakeholders including
the Association of Chief Police Officers (ACPO). Counting rules are available for all to see 7. They form a
national standard for the police to operate. In recent years the Inspectorate of Constabulary (HMIC) has

4
The marital rape exemption was abolished in England and Wales in 1991 by the Appellate Committee of the House of Lords, in
the case of R v R. http://www.bailii.org/uk/cases/UKHL/1991/12.html accessed 15 May 2013.
5
This change occurred before the setting up of the current Supreme Court which would decide any future similar changes.
6
See press discussion of the verdicts given to a group of men in Oxford accused of grooming and sexual exploitation at
http://www.bbc.co.uk/news/uk-england-oxfordshire-22438623 accessed 15 May 2013
7
Home Office counting rules can be found at https://www.gov.uk/government/publications/counting-rules-for-recorded-crime
accessed on 15 May 2013. This document contains the 2013 revised set of Rules for counting and classifying crime.

2
included an inspection of the processes for collecting statistics within its routine inspection procedures
and does not hesitate to criticize forces that fall below the standards laid down by the Home Office. 8

Having the Home Office set standards is not the only way crime statistics could be collected: other
methods are possible. Crime statistics could be left to each police force – but that would mean national
figures would be misleading. It could be a power given to the European Commission to define a set of
European counting rules9 but this would be inappropriate because of the different legal systems in each of
the 28 countries. The logic of allowing the Home Office to set the rules for counting is that the law is the
same for each police force in England and Wales.

However, such pragmatism has its disadvantages. It completely depends on the police being the sole
organization to collect the data on crime. But research has shown and the police acknowledge that they do
not hear about all crime. There are two ways of dealing with this:

a. to collect crime numbers through surveys: such as asking a sample of households or

businesses about crimes they experienced and whether they reported them to the police.
b. to collect information from other authorities on crimes that know about and how they dealt with them.

We shall come to these other sources later when the author will content that the main current deficiency
in crime figures is that, although there are good figures from household and business surveys, the Home
Office does not take full account of the information available to other authorities.

What do the official figures claim?

Latest England and Wales figures are for 201210. Police recorded 3.7 million crimes in 2012, 8% down on
2011. In the household survey 8.9 million crimes against adults were reported (5% down on 2011) and,
for the first time 0.9 million against children aged 10-15. In the same year the survey of businesses 11
reported 9.2 million crimes and police recorded around 2.3 million incidents of anti-social behaviour.

Each of these figures can be divided into different types of crime: even though crimes as a whole fell,
there will be some types that rose. Depending on the interests of the media reporting such statistics, they
may give the overall (quite positive) picture or concentrate on those more negative parts.

The picture over a longer period also shows falls in crime, although it is not possible easily to compare
crime against children aged 10-15 ( only collected once so far) or against businesses ( because of different
coverage of previous surveys.)

The standard longer term view of published police figures is that they rose continuously from around
1950, when they were around half a million, to over 5 million by the first half of the 1990s, since when
they have gradually fallen. There have been several occasions when the counting rules have been changed

8
In an inspection report on Kent Police in 2013, HMIC stated that ‘the force is under-recording approximately one in every ten
crimes.’ See ‘Crime Recording in Kent’ http://www.hmic.gov.uk/publication/crime-recording-in-kent/ accessed June 26 2013
9
This is not such a silly suggestion. The 50 states of the USA, with their different legal systems, adopt counting rules set by the
federal authorities in Washington. See http://www.fbi.gov/about-us/cjis/ucr/data_quality_guidelines accessed 15 May 2013.
10
These figures can be found on the web site of the Office of National Statistics, an independent body at
http://www.ons.gov.uk/ons/rel/crime-stats/crime-statistics/period-ending-december-2012/index.html accessed 22 May 2013.
11
This survey concentrates on the four of wholesale and retail, manufacturing, accommodation and food, and transportation
and storage.

3
over this long period but despite these the broad picture is one of a large rise, followed by a more recent
fall. Notwithstanding this recent fall, it is a fact that crime numbers now are some 7 or 8 times their level
in 1950. Only a small amount of this rise will be due to increased population.

Household victimization surveys were first conducted in 1981. In broad terms households report
something like four times the number of crimes than the police know about. This ratio changes from time
to time. However, the overall trend in the thirty years from 1981 to 2012 has broadly tracked the police
figures, showing a rise to the mid-1990s and a substantial fall ever since then.

We go into possible reasons for these falls in the next section.

It is also worth noting that the official figures have never claimed to include all crimes that the police
know about. They are a subset of all crime, defined by the Home Office because it would be impossible
and/or unrealistic to expect police to record or households to report every single crime: eg every misuse
of drugs or all road offences.) Typical crimes not included but which can result in a criminal sanction,
are:

 Most motoring offences such as speeding, drink driving, parking, driving without various forms of
authority (eg. MOT, licence, insurance) although the more serious motoring offences are included.
 Most public order offences ( eg, public drunkenness, prostitution)
 Most drugs offences of possession and use.
 Most regulatory offences ( eg. TV licence evasion, offences against by-laws)
 Most cybercrime, fraud and bribery

What lies behind the falls in published crime statistics?

First, it is necessary to consider the main drivers for the levels of crime over the last generation:

Crime Prevention

In the author’s view that one important driver for crime levels has been the greatly increased effort put
into crime prevention over the last 20 years. Up to the 1970s police concentrated on crime investigation
and public protection: house protection was generally poor: local authorities and manufacturers left crime
prevention to others. However, the 1980s and 1990s saw much more emphasis on crime prevention:
manufacturers recognized the commercial advantage in making their products safer: cars became more
secure: perimeter security for houses, schools and hospitals became much more effective and local
authorities were encouraged - and in 1998 forced through legislation 12 - to take into account the
implications for crime prevention of all that they did and form partnerships with others to reduce the
likelihood of crime occurring.13 It is not surprising that such concentration of effort into crime prevention
has led to a real fall in crime. We also find that other countries that have had similar crime prevention
initiatives have also seen a fall in their crime figures, and that the EU Statistical Office has reported that:

‘In most EU countries, crime levels have been decreasing consistently since about 2002.’14

12
Crime and Disorder Act,1998
13
See eg. Lewis (2013) The Safer Sutton Partnership Service: Blueprint For the Future?. The Police Journal: March 2013, Vol. 86,
No. 1, pp. 15-28.

4
Technological changes15

Developments in technology have also affected crime levels. The last 20 years have seen great strides in
the use of computers, which are now integral to nearly all consumer goods and household appliances.
Security is also integral to any appliance which has a power source: the use of passwords, etc to be able to
enter buildings and use computers, TVs, DVD players, etc has grown so that it is not surprising that
traditional crimes of stealing have reduced as many modern electric goods can be effectively useless
without their electronic key. The use of physical money has been substantially reduced, so that few
people now carry large sums of cash around for day-to-day transactions: card use has replaced cash and
cheque use. To steal a modern car you need to steal the keys first or force the owner to open it up for you.

These technological developments have substantially altered crime. Traditional crime such as theft,
burglary, car theft, robbery have declined because there is less to steal in the way of cash and perimeter
security is so much greater than it was.

But the criminal has not gone away: he may find it less worthwhile to try to break into your house or pick
your pocket. What does seem to be happening, however, is that the criminal is now moving his target to
your computer, your web site or to your bank account.

Old crimes come in new guises, as well. Computer hackers have devised a new crime whereby they
threaten to deny service to web sites and demand ‘protection money’ from local firms in return for not
hacking into their web sites. Not surprisingly, many firms do not report this to the police.

Increased prosperity

Whatever we may feel about the last 5 years, it is accepted that the second half of the 20 th century saw a
great increase in personal wealth and possessions. However, whereas this wealth was in the form of
valuable possessions it is now more likely to be held in savings for the purchase of services such as
holidays, eating out, or increased leisure. Criminals, in the past were likely to try to steal these
possessions. Today they target savings by committing frauds that are characterized by the making of
thousands of telephone calls to persuade ‘victims’ to put their money into dodgy investments or the
sending out of millions of e-mails in the hope that a small percentage of people will be persuaded to
respond to non-existent lottery wins, free holidays, easily obtained qualifications, high-yielding
investments or other similar temptations.

Reactions to new forms of crime

14
See Eurostat Statistical Bulletin Crime Trends in detail, January 2012 accessed 20 May 2013 at
http://epp.eurostat.ec.europa.eu/statistics_explained/index.php/Crime_trends_in_detail
15
For Cyber Crime the need for the government and law enforcement agencies to respond to technological developments and
the need for better public education has been covered most recently by the Home Affairs Select Committee in their 29 July
2013 report downloaded from http://www.publications.parliament.uk/pa/cm201314/cmselect/cmhaff/70/7002.htm

5
Traditional reactions to crime were to record the event, and then investigate it in the hope of prosecution.
Counting was a by-product. Reaction to more modern crimes is very different.

If a credit card is cloned or a bank account hacked into, the police tend not to get involved. Financial
institutions spend large sums of money on electronic security but they acknowledge that breaches do
occur. Once they do, however, the response is primarily to restore the customer’s financial position.

The attitude of financial institutions to investigation can be very different from the ‘traditional’ police
attitude. Police investigate to catch and prosecute a criminal. Financial institutions primarily look at
patterns of offending so that they can modify their systems to reduce the likelihood of such offences
occurring in the future. They give lower priority to investigating the specific crime and pursuing the
specific criminal. This reasons behind this are primarily economic: the money available for security in the
financial institution is usually better spent by building more secure firewalls or other form of security into
their systems: moreover, to catch a specific computer hacker is often virtually impossible, as the offender
might well be physically in another jurisdiction.

New forms of Fraud

Fraud is a very old type of crime, but modern technologies and systems have greatly increased the
likelihood that criminals will get illegal access to money through new forms of fraudulent activity. The
development of these new forms has led to new structures for dealing with them. 16 Many of these frauds
follow developments in the way people use computers, as well as targeting new types of victim: eg older
people who have access to savings, as mentioned above. Other frauds target institutions such as insurance
companies where the making of false claims for injuries from non-existent car crashes has become an
industry of considerable size and is inflating each single car insurance premium in the UK.

Fraud by an employee is also causing considerable problems to many large institutions: ranging from
false claims for expenses or remuneration: false granting of overdrafts in a bank: false issuing of
contracts, especially to overseas suppliers. Bribery is also a considerable new crime, as trade and aid
become more global and dealers in other countries demand back-handers before they will facilitate an
export contract.

Much such behaviour is on the borderline of crime, deceit, and appealing to people’s greed. Many of
these actions are ones the police would find it difficult to know how to prosecute. New laws such as the
Bribery Act 2010 have been brought in to counter such actions. New decisions of the courts are clarifying
the position almost on a daily basis.17

This section has given a flavour of how technological advances have both led to reductions in traditional
crimes and to an increase in new forms of crime, mainly those that are difficult to define and count.

Limitations of Police and Survey Statistics

16
Eg. see the web site of ActionFraud, which encourages victims of fraud to report direct to them rather than to a local police
force and produces national fraud statistics of its own. http://www.actionfraud.police.uk/ accessed 22 May 2013.
17
For example, as this paper was being finalized, a High Court decision ruled that a ‘Tweet’ which the author claimed to be
‘conversational and malicious’ was in fact libellous and hence potentially a crime.

6
The Police will accept that they do not know about all crime. Household and business surveys help but
there are many other potential sources of statistics. For example:

Financial institutions will know about illegal use of credit cards, bank accounts and internet banking
Hospital A&E departments will know more about violent injuries than the police do
Insurance companies will know or at least suspect that many claims are fraudulent.
The RPSCA and the NSPCC will know about cruelty to animals and to children.
Fraud: a new central source of data of crimes of fraud and forgery has been set up. This shows 440,000
frauds in 2012 compared to the police figure of 136,000.
Central and Local governments and various regulators will know about crimes that are committed in
their area

To develop crime statistics in a more comprehensive way, these sources of data and many others will
need to be used. Indeed, more sources may have to be developed, along the lines of the new centralized
Fraud data.

Has crime really fallen?

It is interesting that this question is actually asked? There is a general acceptance that other measures of
social phenomena both rise and fall over time: unemployment figures go up and down: inflation is known
to rise and fall: the number of babies born will vary from year to year. No-one expresses disbelief or hints
at a statistical conspiracy when these happen.

The main reason for suspicion about whether crime figures are really falling is that, in most person’s
memories, crime is one of those things have been rising for so long – like the number of motor vehicles -
that it has almost seemed a natural law that just as the sun rises each day, crime and motor car usage rises
every time new figures are published. But people are now recognizing that motor car usage has leveled
off. Vehicle first registration figures in 2012 were 20% down on 2001. 18 In the same way it is now
becoming accepted that crime as measured by the police and surveys has actually fallen.

Could the figures be fiddled?

If you define crime in the same way as the Home Office and the police, then you can only say crime has
not fallen by claiming that the figures are somehow fiddled in some way.

There is very little evidence that the figures are fiddled. Home Office statisticians from 1980 to 2010
worked hard to improve the effectiveness of the statistics and their independence from politicians, leading
to the production of the National Crime Recording Standard (NCRS). 19 This was not always easy: the
author recalls being leant on by a political advisor before a 1980s election asking how figures could be
massaged downwards by slipping them from one month to another. However, gradually the National
Statistics label was applied to crime statistics leading to a higher technical standard and to ministers only
18
See vehicle licence statistics bulletin for September 2012, accessed May 22 2013, at
https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/9290/vls-q2-2012.pdf
19
See history of this development in National Crime Recording Standard (NCRS): an analysis of the impact on recorded crime by
Simmons et al. accessed on 24 May 2013 from:
http://webarchive.nationalarchives.gov.uk/20110218135832/http:/rds.homeoffice.gov.uk/rds/pdfs2/rdsolr3103.pdf

7
seeing the figures a day before publication, with no time to influence their preparation. Finally, in April
2012 the responsibility for crime statistics was moved from the Home Office to the independent Office of
National Statistics, with a committee of academics to oversee their production.

In what ways is the system for counting crime deficient?

As well as not including all potential sources, there are other deficiencies resulting from new technologies
noted about. This is partly because such new forms of crime are very difficult to count: eg. how many
crimes were committed in each of the following cases:

a. 10,000 phone calls are made, 250 people respond by making ‘unsafe’ investments and 45 complain to
the police they have been defrauded:
b. A criminal network of accountants, lawyers, doctors and inspectors group together to encourage
1,000 false personal injury claims as a result of motor accidents, of which 350 succeed?
c. 40 out of 500 academics respond to an e-mail invitation to a non-existent conference in the USA, ‘all
expenses paid’, each sending a $30 ‘facilitation’ fee in advance.
d. A computer hacker gets in touch with 100 local firms demanding to be paid £500 for not causing
denial of service through hacking into their web site. 30 pay up.
e. The manager of a small shop which is part of a bigger chain steals a small amount from his store’s
takings each week over several years, through false accounting

As can be seen the very idea of there being a countable number of crimes is becoming outdated.

Media Comment on crime falls

This paper has concentrated on what might be called technical issues but a counter weight to this
approach is to look at media and academic comments on recent crime figures.

When commenting recently on the UK Peace Index, which showed the UK had a very low rate of crime,
the Daily Mail20 claimed the following had led to recent reductions in crime: Technology advances
deterring criminals: Internet use exposing people to the judgement of others: Reduced alcohol
consumption: Fewer young people: Poor people and young people off the streets due to raising of the
school leaving age: Internet use exposing people to the judgement of others: Ban on lead in petrol:
Abortion law reform. No evidence was quoted for any of these.

When looking at causes, one criminologist, Professor Hough was quoted by the BBC as saying that:

The absence of a 1980s-style "loadsamoney" culture may be a key factor behind the austerity-defying downward
trend in nearly all types of crime in England and Wales21.

Others have queried whether crime is really falling. Berman in a UK parliament briefing paper says 22:

Perhaps crime is not actually falling but crime reductions have been confined to those crimes covered by surveys.
Administrative police data and victimization surveys tend to concentrate on well-known crimes…t. It may be the
20
See detailed discussion at http://www.dailymail.co.uk/news/article-2313942/UK-Peace-Index-Rate-murders-violent-crime-
falling-faster-Western-Europe.html#ixzz2TNNoKfZg accessed on 15 May 2013
21
See discussion on the Guardian web site at http://www.guardian.co.uk/uk/2013/apr/25/crime-rates-falling-austerity
accessed on 15 May 2013
22
See discussion at www.parliament.uk/briefing-papers/SN06567.pdf accessed on 15 May 2013

8
case that offenders have diverted attention away from these crimes and are committing more crimes that are not
covered by such statistics such as internet-related offences and e-commerce

Professor FitzGerald has gone further and while claiming there is still a strong relationship between crime
and the economy criticizes the current ways of counting it. In a recent Police Foundation seminar 23 she
has challenged the new Police and Crime Commissioners (PCCs) to oversee a cultural shift in the way the
police detect and record crime, based on the police proactively looking for crime: she told a mainly police
audience that:

The police service has been driven for nearly 15 years by the imperative to demonstrate year-on-year reductions in
crime. So it has developed a mindset… resistant to recognizing new types of crime (such as card fraud and the many
scams perpetrated over the internet) and it has been more than happy for the Home Office to absolve it of taking any
responsibility for them. Yet proactive police work …using the internet and social media, is uncovering large
amounts of unreported crime.

Conclusion

This article has explained the main characteristics of police recorded crime figures and the
complementary survey data. It has examined the long standing fall in both figures and the reasons which
might lie behind this fall. Readers can make up their own minds as which explanations they prefer.

The current author has, for some time, taken the line that the fall in the figures is real but that the figures
themselves are inadequate. He agrees with Berman and Fitzgerald. The current measures are inadequate
to cope with a changing trend in criminal activity, away from ‘conventional’ crimes such as robbery, theft
and burglary, towards a greater concentration on internet crime, frauds of all kinds and the various kinds
of corruption associated with the global economy and the growth in electronic commerce.

Crime has always responded to new technologies. No crimes associated with motor cars were committed
before the start of the 20th century: mobile phone theft was not possible in the 1950s: computer hacking
was unknown before the development of the world-wide web. What is also not new is the speed at which
criminals move their activities once technology changes. Typically criminals move long before the
investigation authorities react, or the statisticians devise ways of counting new criminal activities. Just as
generals are said to be always fighting the last war, government tools for measuring social phenomena
such as crime, unemployment or health always lag a generation behind.

The answers are not straightforward. Many currently held assumptions about crime statistics may need to
be jettisoned. Specifically, the following aspects will need to be carefully considered:

 Law enforcement should acknowledge they are only a small part of the system to combat crime. In
the future prevention, measurement, investigation and sanctions for crimes will be spread across
many more authorities than just law enforcement and coordinated in a loose way, if at all.
 Statistics should no longer be a by-product of the investigation process or the individual reporting
process for surveys.
 Some international agreement will be needed to cover crimes that cut across national boundaries.

23
See ‘Rethinking the ‘fall’ in crime’ at http://www.kent.ac.uk/sspssr/postgraduate/news/view.html?view=1378 accessed on 15
May 2013

9
 In the area of measurement, the following is likely to happen
o Many more sources of crime figures should be used.
o The idea of a simple global national total of crime should be abandoned.
o It will need to be recognized that many crimes exist only in cyberspace
o More organizations need to be involved in setting the rules for crime measurement.
o Such organizations will also need to become part of the government data collection systems.
o Many of these will be government departments, financial institutions or traders who may not
wish to acknowledge publicly the full extent of crime to which they are subjected.
o Some central authority, probably the Office for National Statistics should coordinate all this

10