Navigation – Plan du site

« Lies, damned lies and criminal statistics » : Reinterpreting the criminal statistics in England and Wales

Robert M. Morris
p. 111-127


Ce qu'on est en droit de tirer des statistiques pénales établies en Angleterre et Galles, depuis leur création en 1803, n'a pas soulevé un grand intérêt, ni controverse, parmi les historiens. Bien que les défauts de ces données soient bien connus, on a eu tendance à s'accorder pour les utiliser pour comprendre les phénomènes que, de l'avis général, elles visaient à représenter. En se fondant sur des sources publiées, le Dr. Howard Taylor a remis en cause ce consensus et a proposé une relecture des données sur un siècle, jusqu'en 1960. Il soutient qu'on ne peut pas considérer ces données comme une production, certes défectueuse, mais réalisée de bonne foi ; il y voit au contraire une entreprise délibérée visant à dissimuler le véritable niveau de la criminalité, meurtres compris. En analysant les données dans une perspective de l'offre, il affirme qu'à certaines périodes, on a cherché à atteindre ce but par un sous-enregistrement, une pénurie budgétaire organisée, et en jouant sur les poursuites.
Le présent article soutient que l'argumentation du Dr. Taylor fait trop peu de place au caractère décentralisé et émietté des politiques et des pratiques pénales ; qu'elle attribue trop de signification à des phénomènes statistiques fortuits ; et qu'elle voit à tort, un projet prémédité là où n'existait pas la capacité d'en contrôler le résultat que l'on peut, du reste, convenablement expliquer d'une autre façon.

Haut de page

Texte intégral

  • 2  Taylor (1998a, 1998b and 1999).
  • 3  The principal examples of modern historical accounts are perhaps Gatrell and Hadden (1972), Gatrel (...)

1This paper responds to a series of recent articles by Dr Howard Taylor2 in which he has, single-handedly, revitalised the approach to unravelling the historical significance of data published in the Criminal Statistics for England and Wales3. Broadly, he has maintained that the outputs of the criminal justice system from the 1850s to the 1960s were consciously manipulated by the executive and its agencies to misrepresent the true incidence of crime.

2The paper adopts two strategies – the first general and contextual, and the second to engage directly with Howard Taylor's arguments. The first is a necessary preliminary to establish the essential operating background to criminal justice policy development in the last 150 years. Unless that is done, discussion cannot comprehensibly identify what is, and what is not, significant in interpreting the data under review.

I. The contextual background

Central and local government in England and Wales

  • 4  Official Statistics Committee (1881), memorandum of November 1876 by Robert Giffen, head of statis (...)

3At the outset it should be understood that England and Wales falls to be viewed as a discrete entity within the United Kingdom (UK). Despite the existence of powerful UK-wide institutions such as the monarchy and the Westminster Parliament, the UK is a union as opposed to a unitary state. Administratively, it consists of three distinct parts viz. Northern Ireland (and, until 1922, the whole of the island of Ireland), Scotland, and England and Wales. Although devolution has given assemblies and executives of varying kinds to Northern Ireland, Scotland and Wales, for criminal justice purposes England and Wales continues as a single unit. It follows that the separate operations of the criminal justice systems within the UK have never been treated statistically as a single unit. Although there was some initiative in that direction in the nineteenth century, it did not prosper4.

4In addition, one of the cardinal facts to be borne in mind in approaching historical accounts of criminal justice in England and Wales is the extent to which its delivery was locally organised, locally financed and locally run until very recent times. Indeed, in many of its current aspects the delivery of criminal justice is still a local affair : 96 % of all criminal cases are disposed of by magistrates ’courts where 30 000 lay, and a small band of stipendiary, magistrates sit in locally run and partly locally financed courts (in their administrative capacity, magistrates controlled county government until 1888, and the fact that they for long shared the social background of the legislators – and, as well as being frequently related, were often actually the same people – meant that Parliament did not lack sympathy for the local perspective).

5The higher criminal courts were nationalised only as recently as the Courts Act 1971. There are 43 police forces which are now all accountable to local police authorities. For long the sole exception – the Metropolitan police which since inception in 1829 has been answerable solely to (and subject to the orders of) a minister, the Home Secretary – has become answerable to a new police authority following the creation of the new Greater London Authority in 2000. Apart from the limited remit of the Director of Public Prosecutions originating in 1879, there was no national system of public prosecution operating until 1986 (under the Prosecution of Offenders Act 1985). Even then, no state monopoly of prosecution was created because the way was still left open for private prosecutions in addition to those undertaken by a considerable range of statutory bodies.

6At the same time, of course, the role of central authority should not be discounted. Royal judges have been circumambulatory since the early twelfth century. Since 1833 government has contributed on an increasing scale to the costs of local justice in respect, first, of policing costs (until 1856 in London only) and, secondly, from 1835 of prosecution costs. Throughout, government appointed all the judges in the senior courts and controlled the system which appointed the magistrates. In more recent times, government has embarked on financial control policies which have imposed monetary ceilings on local expenditure, and successive Acts in 1991 (Criminal Justice Act) and 1994 (Police and Magistrates’ Courts Act) have, for example, imposed cash limits on the permitted annual expenditure of magistrates’courts, and of the police and probation services.

  • 5  At the same time it has to be borne in mind that private for long outnumbered public legislation i (...)
  • 6  The Bill introduced flogging for « garotters », and was memorably castigated by Grey as « a panic (...)

7The government has also throughout usually been the dominant partner in achieving systemic change by means of statutory enactment, for example in adjusting the powers of the criminal courts5. Successive statutes, particularly from 1855, redistributed business downwards from the upper courts to the magistrates’courts by allowing certain kinds of indictable offences to be triable summarily in the lower courts rather than in the upper courts alone. In contrast, the powers of the police received relatively less legislative attention until very recent times, largely because they were left for long subjected to the common law and judicial oversight rather than statutory regulation. But nineteenth century governments did not necessarily sweep all before them and always get their way. The Constabulary Act 1856 was preceded by the rejection of Palmerston's attempt to legislate on the same subject. Conversely, in 1863 the Home Secretary (Grey) was unable to dissuade Parliament from passing Adderley's Security from Violence Bill6.

  • 7  Maitland (1885, p. 148).

8Accustomed as the UK is in present times to a Parliament where the executive and government legislation dominate, it takes a special effort perhaps to recall that the principal reason why prosecution practice was not modernised lay in the absence throughout the Victorian period (and for some time afterwards) of a workable, abiding consensus in Parliament about what should be the basis of reform which did not increase the reach of government patronage. As Maitland put it : « It became the mark of enlightenment to demand the immediate creation of a complete system of public prosecutors ; but beyond this demand there was extremely little agreement »7.

  • 8  The rejection on 20 January 2000 by the House of Lords of the government's Criminal Justice (Mode (...)
  • 9  Described as « the institution of a centralised police force », an unusual misdescription in the c (...)

9The fact that recent governments have been perceived as, and criticised for, following policies of centralisation is itself an indication of the persistence of localised forms of criminal justice administration8. This is not of merely wishful or partisan retrospection : it underlines the need to appreciate the real and pervading distance between central and local decision making for practically the whole of the period effectively under discussion. Whilst policing outside London would not have assumed a common form and coverage when it did had not the government agreed after an earlier failure of legislation based on the Report of the Select Committee of 1852-1853 to contribute one quarter of the costs from 1856, the system established was essentially a locally controlled one in the shires as it remained in the boroughs – where the borough police authorities insisted on keeping the greater degree of influence over operational matters that the original 1835 legislation had given them as opposed to that available in the shires from 18399.

  • 10  It has been pointed out that one of the principal reasons why magisterial rule in the English and (...)

10The fact that the UK is a relatively small landmass does not mean that centralisation was inevitable. On the contrary, its relative inviolability behind what was for long a challenging moat meant that central government did not have to insist on administrative forms wholly under its control. In addition, before the cumulative effects of extending the suffrage, there was – as already indicated above – a good deal of common view between the office-holders at the centre and those in the localities : the men who participated in national politics were essentially at one with – and not infrequently the same as – those who led local political life10. In Ireland, the same conditions did not exist under the rule of an alien protestant Ascendancy and, as a result, different forms developed – notably, centrally controlled police forces, prosecution control by a single professional authority, and a largely stipendiary magistracy.

The development of the statistical series

  • 11  See note 3 above.

11The development falls into five phases, of which the first four have been amply described before11. The fifth commenced in 1981 with the inauguration of the crime surveys i.e. national surveys of victims’experience of crime.

12In practice, Howard Taylor is concerned with the third and fourth phases. These originated in reforms respectively of 1857 and 1893. That of 1857 (under the direction of the Home Office official, Samuel Redgrave) inaugurated the publication of offences whose commission came to the attention of the police, an initiative made possible only by the Police Act of the previous year. The 1893 reform divorced the criminal and judicial statistics into separate series, dropped some of the less worthy material, and tightened up the counting rules to produce more uniformity of reporting amongst the nearly 200 separate police forces. Thereafter, the annual series increasingly became the base for more detailed individual studies and were complemented by additional series e.g. on probation and road traffic. Although now more sophisticated, and complemented also by the crime surveys, it is essentially this system that persists in the annual publications of today.

13Perfection was not reached in 1893 or subsequently, nor did contemporaries claim otherwise. As has been observed :

  • 12  Radzinowicz, Hood (1986, p. 107).

It is not sufficiently appreciated by modem students of criminal statistics that all the major limitations of this source had already been perceived and expressed with force and clarity by nineteenth century inquirers12.

14To what extent was the reformed data of better quality ? In his preface to the 1894 Statistics, Troup, the Home Office official responsible for completing the 1893 reform, discussed and showed full awareness of the stratagems available to, and used by, the police in their recording and reporting practices :

The figures showing the numbers of crimes and consequently the proportions must be taken with a great deal of caution. The returns of the numbers of crimes committed depend to a considerable extent on the discretion of the police ; they are not certain and definite figures in the same way as the numbers of prosecutions and convictions are certain and definite.

15Going on to explain why the police returns in this case had been re-entitled « Crimes known to the police » because the term « Crimes committed » was misleading, he said :

  • 13  Criminal Statistics 1894 (1896), preface dated 1 May 1896, p. 30 and pp. 34-35. He repeated the he (...)

Not only do the figures fall short of the real number of « crimes committed » by the enormous number of unreported or unknown cases ; but there seems much reason to think that, though the instructions as the mode of collecting them have been made as definite as possible, there is still a tendency on the part of some police forces to adopt a very high standard of what constitutes a « crime committed » or a « crime reported to the police », and by this means further to reduce the number of cases entered into this column... o doubt it is natural... that they should seek to minimise the amount of unpunished crime existing in their district, but such a tendency detracts so much from the value of the returns of crime that it almost raises the question whether it worth retaining the returns at all... t should be clearly stated that they represent only the crimes known to the police, and do not even approach the real total of crime13.

  • 14  Kitsuse, Cicourel (1963), p. 137 « Thus, rates can be viewed as indices of organizational processe (...)

16His understanding preceded what is usually regarded as its classic statement in the criminological literature by 70 years14.

  • 15  Minute of 3 May 1929 by Arthur Locke, whom the Home Secretary appointed to chair the committee set (...)
  • 16  Perks Report (1967). For recent changes in the counting rules, see Home Office Statistical Bulleti (...)
  • 17 Times and Guardian reports, 1.8.2 000.

17Throughout efforts have continued to be made to develop standardised recording practices. Troup gave much early attention to the problem which was particularly intractable when there were well over 100 separate police forces. The problems continued and were the cause of a Home Office initiative in 1929 to investigate « what is behind the extraordinary variations between different parts of England and Wales, and even between different Divisions in the Metropolitan Police District »15. Subsequent attempts were made to improve the quality of the data. For example, the Perks Committee undertook a thorough view during the 1960s. There have been continuing exchanges since between the Home Office's now thoroughly professionalised statistical services and the police, and important developments in the counting rules whose effect will be to give a clearer account of the victim's experience and allow greater comparability between notifiable offences recorded by the police and the product of the victim surveys16. More recently still, Home Office Ministers have been prepared to accept the recommendations of a study by the Inspectorate of Constabulary concerned with the accuracy of police recording of crime and which could result in raising recorded crime figures by as much as a quarter17.

II. Historical significance

Historians and the criminal statistics

18Historians have come late to crime, and it is only in the last three decades that they have shown any considerable interest in crime phenomena. In another sense, of course, they will always come late because their interest is backward looking. It is also utterly dependent on pre-existing data in whose construction they cannot have had any hand and whose interpretation, like other historical records, is fraught with difficulty. It is never enough to look at the data without understanding the motives and working methods of their compilers.

  • 18  Downes, Rock (1998, p. 50). On the other hand, it has been pointed out by a criminologist that the (...)

19The extent to which historians and the sociologists of deviance seem to have coexisted in parallel universes has on occasion been striking. Whilst some differences of perspective are to be expected where the former operate with retrospection and the latter are more preoccupied with explaining the present and predicting the future, historians have seemed relatively unfamiliar with the extensive sociological literature about data sources in which they could be expected to have at least a similar interest. Thus, discussion of crime trends has often started and finished with analyses of the criminal statistics, especially those of police recorded crime. This has led some sociologists to conclude that « Historians are particularly prone to defer to police and court records as indisputably sound »18. In their effort to account for evident changes in public behaviour during the nineteenth century, they have sometimes leant heavily on the official statistics and have treated them with a respect and lack of scepticism, which probably would have surprised Redgrave and Troup.

Howard Taylor's investigations

  • 19  See Note 2 above.
  • 20  Taylor (1998a, p. 571).

20In his recent articles19, Howard Taylor has reinvigorated the debate about crime trends by focusing on the supply side of criminal justice from the nineteenth century up to the Second World War. Far from showing the criminal statistics unthinking and exaggerated respect, he has been properly determined to show them no unearned respect at all, and look closely amongst other things at what they reveal about the impact of criminal justice services on those affected by crime, including the victims of crime. To enlarge on his statement « the crime statistics largely reflected supply side policies »20, the perspective he has explored may be encapsulated in the question « what if the official data represent not the product of a free flowing, open-ended series of criminal justice processes but merely record the output of a system where, despite managed appearances to the contrary, it was in fact the inputs that were controlled in the first place ? ».

21This hypothesis is certainly challenging and fruitful, but how far has it been demonstrated to hold true ? As to fruitfulness, for example, the hypothesis draws useful attention to just how little some of the inputs have been studied, and how far the division and re-division of criminal business between the courts was manipulated. Police behaviour comes in for particular criticism ; and « bureaucracy » is given the character of the unseen hand. Although the perspective is novel, Howard Taylor argues from published sources only and risks walking into the fire of the sociologists ’criticism of historians’ naivety already mentioned in so far only, of course, as that criticism is valid.

22Howard Taylor has covered a wide field in a series of articles to which on this occasion it is not possible to respond in the same detail. Space requires selection, and what follows will therefore concentrate on two important limbs of his argument : the extent to which acknowledgement of crimes of murder was rationed by the authorities, and the extent to which the prosecution process as a whole was controlled by Treasury supply.

Howard Taylor and murder

  • 21  Morris, Blom-Cooper (1964). The authors were respectively a Professor of Criminology at the London (...)
  • 22  Taylor (1998a, p. 85).

23The argument about murder is that the numbers prosecuted were deliberately limited in the period 1880-1966, that is, between when the office of DPP was first created (before its more effective re-establishment in 1884) to the year immediately after the death penalty for murder was abolished. Moreover, not only were the numbers limited but they were limited, it is maintained, to an average of just short of 150 a year within a range of 120 to 179 : « In only five of the 104 years between 1862 and 1966 were there either fewer than 120 murders or more than 179, i.e. the number of murders was kept within a tight band of 20 per cent (30 murders) on either side of the average of 150 ». Dismissing as esoteric the belief of Professor Terence Morris and Louis Blom-Cooper21 that the aetiology of murder is distinct from that of other crimes, it is claimed that « a far more likely explanation is that prosecutions for murder were amongst the most strictly rationed of all crimes »22.

  • 23  Gurr (1981, p. 99).

24This conclusion challenges what has hitherto been a fairly settled view viz. that the relative rarity as well as the seriousness of murder meant that it could be regarded as the offence whose recognised incidence was most reliably recorded in the police returns. As one major comparative study of trends in violent crime put it after reviewing the character of official records of crime : « Official records thus are in varying degrees inadequate for assessing the true incidence of violent offences, but closer to the mark for homicide than for others »23. Of course, a settled view is not necessarily synonymous with the truth, and the question therefore is which view falls to be preferred ?

  • 24  For example, by Havard (1960).
  • 25  Broderick Committee (1971).

25First, it must be agreed that not all homicides were then, as now, invariably discovered or acknowledged. Coroners ’inquests were no doubt particularly fallible before the availability of modern forensic and scientific capacities. It has, indeed, been argued that they remain fallible, and serious doubts raised on that score in 1960s24 led to the review of death certification in England and Wales undertaken by the Broderick Committee which reported in 197125. Subsequent cases show on occasion that there still can never be any room for complacency about the certainty of discovery.

26Secondly, however, examination of Howard Taylor's presentation of the recorded figures raises some questions about his conclusions. The first stems from contemplating variations of as much as 20 % above and below his average annual rate. This suggests either a very ineffective form of control, or raises the question of how that degree of variation could be compatible with conscious manipulation. If there was manipulation, then it would have had to be of a very high degree of sophistication sustained by an alliance of well over 100 separate police forces for a period of more than a century to keep to a pre-arranged and unchanging mark of an average of not more than l50 cases a year. In the latter case, granted the highly disaggregated character of the criminal justice process already observed, that such a degree of agreement could be reached at all, still less achieved and sustained, is inherently implausible.

27Moreover, it is not necessary to reach for an explanation of that character when there is a more plausible one suggested by the figures themselves, and acknowledged by Howard Taylor viz. that the variation extended by as much as 20 % above and below the arithmetic average. This is a remarkably high degree of volatility not less significant for being acknowledged by Howard Taylor. In practice, it must seem much more likely, because of the range of incidence, that the average, far from being a target, can only have been an arithmetical accident devoid of anything other than arithmetical significance. What is significant, in other words, is not the average but the volatility. These points taken together, it may be judged, do not so much qualify Howard Taylor's interpretation as invalidate it so far as the quantitative data are concerned.

  • 26  Taylor (1998a, p. 586).

28This nonetheless leaves his two other arguments : that budgetary controls at least inhibited or at most prevented open-ended approaches to murder prosecutions ; and that « because the discovery of a suspicious death and its subsequent investigation and prosecution could make a large dent in a police authority budget, and the chances were that it would not be prosecuted as a murder, it was an open secret that most murders and suspicious deaths went uninvestigated »26. Both these assertions assume that contemporary financing entailed unbreakable budgetary ceilings.

  • 27  The exchange – Official Report, Commons, 27 July 1889, column 888 – concerned whether the heads of (...)
  • 28  The reference (Taylor 1998a, p. 587) to an anecdote in Robert Mark's memoirs seems a mistake. The (...)

29There is in fact no evidence for this. Howard Taylor has taken a Treasury Minister's answer to Parliament in 1889 to imply the impossibility of resort to a Supplementary Estimate – that is, an application to Parliament for a sum additional to that originally sought and granted27. Further, it is far from clear from the material upon which he relies in connexion with the suggested manipulation of the independent coroners’jurisdiction is capable in fact of bearing the construction that he places upon it. Rather, the material more convincingly demonstrates a lack of effectiveness and an unacceptable degree on occasion of complacency but no conspiracy with the police and the Director of Public Prosecutions28.


30The extent and character of budgetary control is, of course, the most important theme running through the whole of Dr Taylor's argument. I shall now therefore turn to deal with it in the context of the important points he seeks to establish about the prosecution system as a whole. It should first be accepted that Howard Taylor has somewhat modified some of the positions he took in his original thesis, and he does not now seek to maintain that there was an explicit regime of cash limiting in anything like the modern sense. However, the ghost of that view does in some respects remain to haunt his argument.

  • 29  Taylor (1998a, p. 571).

31Howard Taylor hypothesises that the processing of crime in the nineteenth century and beyond was influenced by the manipulation of supply : « the crime statistics largely reflected supply-side policies. In this light, [he sought] to find political explanations for two interrelated questions : first, why prosecutions stopped rising and, secondly, why the statistics fluctuated around such an astonishingly constant average for 70 years »29.

32It is notable, however, that sometimes eye-catching assertions are balanced by hesitations and scrupulous reservations. For example, in his Economic History Review article, Treasury cash limits are claimed for 1856 and 1889 at pp. 574 and 587, general budgets at p. 588, but more research deemed necessary at p. 574 to determine the full effect of Treasury parsimony, and a direct relationship between the size of the prosecution vote and the statistics of prosecutions thought unlikely at p. 577. A very considerable volatility is also displayed, if unremarked, in the data displayed in Figure 1.

33It is contended that these hesitations in fact indicate that it is not possible to press home the charge of wilful suppression. Further, it is argued that the article mistakes a long-running argument between local and central government about how costs should be shared for an argument about how often they should be incurred. Whilst it is undoubtedly true – as contemporaries recognised – that prosecution arrangements left much to be desired, there is no evidence that central government sought to impose overall expenditure controls, or used their powers to suppress prosecution rates. On this issue, the Royal Commission of 1836 was clear and never challenged :

  • 30  County Rates (1836, p. 8).

Any expedient founded on the principle of reducing the fair measure of compensation for expense, time and trouble allowed to the prosecutor and the witnesses, would, in our opinion be opposed to justice and to sound policy. The course of legislation on this subject has been already shown to have lain uniformly in the opposite direction. It seems to have been the object of the later enactments to give a complete remuneration to the individuals engaged in prosecuting crime, for the whole money, time, and labour reasonably bestowed by them in the performance of that public service ; and we think it right and expedient that nothing short of this should be awarded30.

  • 31  Oddly, the literature seems devoid of any comprehensive account. The closest approach is probably (...)
  • 32  See Philips and Storch (1999) for a thorough discussion of these processes in the context of polic (...)

34It is not simply a question of emphasis or perspective, though the latter is important. For example, the article offers no continuous account of the tortuous evolution of prosecution policy and legislation in a still limited administrative machine31. Nor does it place the problems in other contemporary contexts, for example changing assumptions and disagreements about the extent to which crime control should continue to be perceived as a local responsibility locally funded ; and whether procedural and management change at local level should be wholly optional within a statutory framework (the Criminal Justice Administration Act 1851, the Public Prosecution Bills of 1871-2) or compulsory (the Public Prosecutors Bill of 1873)32.

  • 33  Taylor (1998a, pp. 574-575).
  • 34  Prosecutions Royal Commission (1859), Evidence, Q 120.
  • 35 Ibid., Appx 7, letter of 17 April 1858.

35The article rightly highlights the evidence33 of Wilmot Seton, the Treasury clerk, to the 1859 Royal Commission but draws questionable conclusions. For example, Seton made clear that from 1846 (when the Treasury assumed responsibility for 100 % of prosecution costs) local costs were paid on demand but that, following uncovenanted increases in local authorities’demands, from 1854 vouchers were required and Criminal Law Accounts Examiners were appointed to oversee the demands and advise on their admissibility. They maintained that the Treasury had often been paying costs that were entirely a local responsibility or not properly incurred. The result had been, the Treasury believed, « ... relief to the local jurisdiction far beyond the contemplation of the legislature »34. Overall, there appeared to be an inverse relationship between prosecution costs and the number of committals, especially in some Welsh counties (the Examiners calculated that universal application of Denbigh rates would double total expenditure35).

  • 36  T28/74-81.
  • 37  T28/81 – « Rules to be observed in relation to the Costs of Criminal Prosecutions which are repaya (...)
  • 38  1859, op. cit., Evidence, Q 148. See HO 347/9 for Home Secretary's Order of 9 February 1858, under (...)

36The relevant Treasury letter books for 1858-187136 contain circumstantial explanations of disallowances, sometimes offering meetings with the Examiners. (A circular was issued in 1869 summarising the Treasury criteria37). Whilst it is true that Seton advocated paying expenses at the lowest rate, he also explained that he was not upheld and the Treasury endorsed the more generous allowances of the Home Secretary's tariff of 185838. A single tariff was bound to create losers as well as winners and this was, of course, the reason for the outcry that led to the 1859 Commission – whose majority upheld the tariff, as in effect did the 1862 Select Committee. The tariff stood until 1904, but at no time extended to the fees for counsel or attorneys which remained locally determined.

  • 39  1859, op. cit., Q120. Joseph Hume, the Parliamentary apostle of administrative and financial effic (...)

37It is relevant to bear in mind that the expenses claimed by the local authorities had already been disbursed by them before they applied to the Treasury for their reimbursement : that is, the Treasury did not take away what had already been given. Rather, the Treasury's aim was to secure a regime of reasonable economy : it was not, and Seton did not claim it was, applying a cash limit – a device first applied to criminal justice expenditure only in the Criminal Justice Act, 1991. As Seton saw it, the principal mischief derived from the fact that the variety of lavish fees allowed locally was « a natural result of the absence of interest in those who superintended the allowances in the general fund from which those costs were ultimately to be defrayed »39.

  • 40  HO 45/9305/11932 has a copy of the judgement (« ... Nothing can be more anomalous, nothing can be (...)
  • 41 Ibid., Treasury Minute of 29 January 1875. See T165/1/3, p. 28, for private 1880-1 Treasury Blue Bo (...)
  • 42  Prosecution allowances (1903, p. 27).

38The friction between the Treasury and the local authorities led in 1872 to a mandamus action brought by Lancashire to force the Treasury to pay on demand. The judges side-stepped the issue by maintaining that, because it related to a Vote, the matter could be determined only by Parliament40. Private members’Bills for public prosecutors were introduced from 1870, just as J.G. Phillimore (the lawyer M.P. who chaired the 1854 and 1856 Select Committees) had introduced Bills in 1854/5. In response, and after extensive consultation and calculating that a compulsory, salaried system would not cost much more than the existing system, the Home Office introduced in May 1873 its own thorough-going Bill of 32 clauses and 3 schedules which, if successful, would have put in place a regime not wholly dissimilar from that finally established in 1986. There was insufficient time for the Bill to make progress, and the subsequent change of government resulted in a change of tack. From 1874 the Treasury paid full Assize costs after only local taxing, and settled for paying Sessions’costs at the average allowed over the previous 3 years41. Later, as noted by the Committee in 1903 set up to review the tariff of 1858, responsibility for prosecution costs were returned to the local level « in consideration of the Exchequer contribution grants which were made under the Local Government Act, 1888, and which were much in excess of the Treasury payments and subventions up to that date »42. Finally, the 1908 Costs in Criminal Cases Act rationalised the statutory lumber from 1752.

39How may the effect of all this on the number and efficiency of prosecutions be judged ? A crucial point is that then, as now, whether a prosecution should occur at all was, and is, as much a subjective as an objective question. As one modern review of past practice has pointed out :

  • 43  Davis (1989, p. 400).

... the growing willingness of the state to underwrite the costs of private prosecutions made prosecutions easier... of key importance in understanding prosecutions is the recognition that prosecuted law breaking represented only a tiny proportion of similar behaviour which for various reasons never reached the courts43.

  • 44  HO 347/11 pp. 563-612. This lengthy printed memorandum appears to have been prepared to assist wit (...)

40Looking back from 1873 in the Home Office where he lacked enthusiasm for a Treasury regime based solely on a brief clause in the annual Appropriation Acts, the then Legal Under-Secretary (Lushington) wrote : « There is no ground for believing that criminal cases have, on the whole, been less effectively prosecuted since the system of Treasury re-taxation has been established. Quite the contrary. But in some instances the disallowances may have been harsh, and here and there steps necessary for the proper conduct of prosecution may have been omitted by prosecutors for fear lest the costs would be disallowed by the Treasury »44. Of course, the Home Office was not an entirely disinterested party, and there must have been some inhibitory effect even if Treasury estimates were supplemented as necessary to defray actually incurred local expenditure. No-one seems, however, to have regarded as satisfactory a situation where they were compelled to act in an environment which continued to find difficulty in devising adequate mechanisms for recycling resources between local and national levels.

  • 45  Prosecutions Royal Commission (1859), Note of dissent by R. Upperton.
  • 46  Public Prosecutors (1854/5), evidence of Horatio Waddington, Home Office Permanent Under Secretary (...)
  • 47  Stephen (1883, Vol. I, p. 497).

41It is doubtful whether further research can establish how exactly funding regimes affected what is fundamentally a discretionary process. Even an early advocate of more generous allowances could do no more than claim « ... here is no sufficient evidence to justify the assertion that there is no failure of justice... his assertion must always be difficult of proof under any circumstances... » – an argument which cuts both ways45. In muddling through, improvements did occur but the most important was the incidental effect of putting into place a permanent, paid constabulary throughout the country which, without anyone planning it (though there was clearly Home Office encouragement)46, became the public prosecutor in default of anyone else. Writing like Maitland in the 1880s, Stephen summarised the situation as arising from « a series of omissions on the part of the legislature to establish new officers for the administration of justice as the old methods of procedure gradually changed their character »47.

Extending the time frame

42This analysis has so far concentrated on the first of Howard Taylor's three recent articles, and now turns to look finally, and more briefly, at the two later articles which carry the supply side perspective further into the present century. The first of these deals with the rising crime statistics during 1914-1960, and the second focuses more narrowly on the balance between types of non-indictable crime (where motoring offences eclipsed drink and similar public order offences). In both cases, the behaviour of the police is scrutinised with particular reference to professional self-concern.

  • 48  Taylor (1999, p. 117). Similar claims are made on pp. 118, 120 and 121.
  • 49  Taylor (1999, p. 117). Similar claims are made on pp. 118, 120 and 121.
  • 50 Ibid., pp. 126 and 128.

43The thesis of both articles is that the police chose to exercise their crime recording discretion to increase the level of reported crime in order to bolster the case for maintaining or increasing police manpower. It is argued that this was achieved through the regional conference areas of chief constables where « the conference area was the basic statistical unit... ther evidence suggests the statistics were coordinated on a national scale with such precision that they could be raised or lowered by the number of serving officers »48. Additionally, it is claimed that « Management targets appear to have been set »49. Trends towards motoring offences already apparent beforehand were much increased following the introduction of the Road Fund in 1921 which diverted fines for motoring offences from the magistrates’courts to the Exchequer : « From this point of view the shift towards policing the motorist rather than the drunk and the vagrant was clearly in the national if not the local interests Moreover, there was a direct relationship between police strengths and motor accident figures : « In the period between 1918 and 1934 the accident figures appear to have been created simply by adding, year on year, an additional number of accidents equivalent to one sixth of the police establishment »50.

  • 51  For example, Wiener (1994).

44Again, the problem with this species of argument is that it relies more than seems credible on supposition, and on arithmetical relationships that are more accidental than probative. (More generally, it may also reflect a fashionable iconoclasm about the administrative state to which attention has been drawn elsewhere51). In a situation where there are more sufficient alternative explanations of the same phenomena, there seems little reason why the supply side theories as advanced should be preferred, granted in addition the implausible degree of concerted conspiracy they posit. In a situation where the number of road vehicles increased by over 350 % between 1920 and 1938, and road deaths ran at annual rates more than twice today's rates with more than forty times the 1920 number of vehicles on the road, it is not difficult to see why traffic increasingly took up police attention. But the data shows that the police had not ceased to deal with drunkenness, still less had switched away from it. On the contrary, after the figures for drunkenness convictions had fallen during 1920-1932, they rose continuously from 43 000 cases to 52 600 in 1938, the last fully peacetime year. In other words, as is common in the real world, change was not then, any more than now, an entirely linear process.

  • 52  Martin and Wilson (1969), Chapter XI.
  • 53  Gatrell, Hadden (1972 p. 368) ; Field (1999). The perception of such linkages is not, of course, n (...)

45Of course, there was much police anxiety from time to time about the service's increasingly prominent position in criminal justice expenditure, just as there was perplexity in the Home Office about trying to explain changing statistical trends and address awkward issues about police productivity – which one estimate has calculated as declining by over 30 % 1921-196552. But the long term upward trend in reported crime of 5 % per annum was established after the 1914-1918 War, and has been impervious to every vicissitude in police strength, deployment or rhetoric. The trend has yet clearly to change direction (the most recent figures at the time of writing show an increase) and the downturn of the past six years appears to have been corroborated by the British Crime Survey only after a delay. Stronger than possible relationships with the self-interested goals of Treasury and/or Home Office officials on the one hand, and police officers on the other, have been the enduring relationships between certain kinds of crime and the economic cycle, first in modern times identified by Gatrell and Hadden, and for more recent times by Simon Field53.

Lies, damned lies, or just criminal statistics ?

  • 54  Maguire (1997, p. 141).

46Police recorded crime statistics deal in and with behaviours. In that sense, crime is « ultimately, a social construct. Looked at as an abstract formal category, it consists of a diverse set of behaviours which have in common (perhaps only) the fact that they are proscribed by the criminal law »54. As Troup tried to emphasise a century ago, they are not facts in the same sense that court appearances and decisions are facts. They reflect the experience and decisions of a multiplicity of actors in the interactive social process of criminal justice. They are of their time, and are appealed to, and employed for, the uses of that time. It follows that they will not be internally consistent over time. In his 1857 reforms, Redgrave constructed them so that they could answer to his belief in the existence of criminal classes. In the current age of unbelief, we look to them for data about offences at least as much as we look to them to explain offenders. But no more now than in the past are they to be taken at face value. Indeed, a former editor of the Times, no less, has sermonised :

  • 55  Simon Jenkins, Times, 1 December 1999.

There are, we all know, three sorts of lie. There are statistics, crime statistics and police crime statistics55.

47Nonetheless, the criminal statistics are incontestably the most profuse and the most accessible data available. They are, therefore, to be taken seriously for what they aire, that is artefacts of contemporary social processes. To inquire whether they share the characteristics of observations of the physical universe, say, is less important than to establish what they mean and what they tell us about the human universe that created them. For that enterprise, the historian and the sociologist need each other, and both need the statistician.

48Howard Taylor's articles ask new questions about the data, and his approach does a great deal to challenge views which may have become too settled and comfortable. Whatever may be said about the detail of his argument – and, to draw on a road traffic analogy, it risks looking very like hit and run history at times – it is an argument abundantly worth having.

Haut de page


Aberdare Letters – Letters of Lord Aberdare, 1902.

Aubusson de Cavarlay, B., What and how do police statistics count ?, Penal Issues, 1998, 9, 10-13.

Cavanagh, T.A., Scotland Yard Past and Present ; Experiences of thirty-seven years, London, Chatto & Windus, 1893.

Davis, J., Prosecutions and their context, in Hay, D., Snyder, F., (Eds.), Policing and Prosecution in Britain 1750-1850, Oxford, Clarendon Press, 1989, pp. 397-426.

Devlin, P., The Criminal Prosecution in England, New Haven, Yale U.P., 1958.

Downes, D., Rock, P., Understanding Deviance, 3rd edition, Oxford, Oxford U.P., 1998.

Dunbabin, J.P.D., British local government reform : the nineteenth century and after, English Historical Review, 1977, XCIL, 777-805.

Edwards, J. L1. J., The Law Officers of the Crown : a study of the offices of Attorney-general and Solicitor-general of England with an account of the office of the director of public prosecution in England, London, Sweet & Maxwell, 1964.

Emsley, C, The history of crime and crime control Institutions, in Maguire, M., Morgan, R., Reiner, R. (Eds.), The Oxford Handbook of Criminology, Oxford, Oxford U.P., 1997, pp. 57-86.

Field, S., Trends in Crime Revisited, London, Home Office Research Study 195, 1999.

Gatrell, V.A.C., Hadden, T.B., Criminal statistics and their interpretation in Wrigley E.A. (ed.) Nineteenth Century Society : Essays in the Use of Quantitative Methods for the Study of Social Data, London, Cambridge U.P., 1972, pp. 336-396.

Gatrell, V.A.C., The decline of theft and violence in Victorian and Edwardian England, in Gatrell, V.A.C., Lenman B., Parker, G. (Eds.), Crime and the Law : The Social History of Crime in Western Europe since 1500, London, Europa publications, 1980, pp. 238-370.

Gurr, T.R., Historical trends in violent crime : a critical review of the evidence, Crime and Justice, 981, 3, pp. 295-353 (reprinted in Pease, 1999, pp. 27-85).

Havard, J.D.J., The Detection of Secret Homicide. A study of the medico-legal system of investigation of sudden and unexplained deaths, London, Macmillan (Cambridge Studies in Criminology, vol. 1), 1960.

Holdsworth, A History of English Law, Vol. XV, London, Methuen, 1965.

Kitsuse, J.I., Cicourel, A.V., A note on the use of official statistics, Social Problems, 1963, 1, 2, pp. 131-139.

Maitland, F.W., Justice and Police, London, Macmillan, 1885.

Mark, R, In the Office of Constable, London, Collins, 1978.

Martin, J.P., Wilson, G., The Police : A Study in Manpower. The evolution of the service in England and Wales, 1829-1965, London, Heinemann, 1969.

Morris, T., Blom-Cooper, L., A Calendar of Murder, London, Michael Joseph, 1964.

Philips, D., Storch, R.D., Policing Provincial England : The Politics of Reform, London, Leicester U.P., 1999.

Pease, K. (ed.), Uses of Criminal Statistics, Aldershot, Dartmouth, 1999.

Prest, J., Liberty and Locality. Parliament, permissive legislation and rate payer's democracies in the nineteenth century, Oxford, Clarendon, 1990.

Radzinowicz, L., Hood, R., A History of the Criminal Law, Vol. 5, Oxford, Clarendon Press, 1986.

Stephen, J.F. Sir, A History of the Criminal Law of England, 1883.

Taylor, H., Rationing Crime : the political economy of the criminal statistics since the 1850s, Economic History Review, 199 (a), LI, 3, pp. 569-590.

Taylor, H., The politics of the rising crime statistics of England and Wales, 1914-1960, Crime, Histoire etSocietes/Crime, History and Society, 199 (b), 2, 1, pp. 5-28.

Taylor, H., Forging the job, British Journal of Criminology, 1999, 39, pp. 113-135.

Wiener, Martin J., The unloved state : twentieth-century politics in the writing of nineteenth-century history, Journal of British Studies, 1994, 3, pp. 283-308.

Young, M., An Inside Job, Oxford, Clarendon Press, 1991.

Official Publications

Broderick Committee, 1971 – Report of the Committee on Death Certification and Coroners, Cmnd 4810.

County Rates – Report of the Commissioners on the County Rates in England and Wales, Parliamentary Papers (PP) 1836, XXVIL.

Criminal Statistics for England and Wales 1894, C 8072, 1896,

Criminal Statistics for England and Wales 1895, C 8352, 1897.

Home Office Statistical Bulletin 18/99, 2 October 1999.

Official Statistics Committee, Second and Third Reports, PP 1881, XXX.

Perks Report – Report of the Departmental Committee on Criminal Statistics, Cmnd 3448, 1967.

Prosecutions Royal Commission — Report of the Royal Commission to inquire into the costs of prosecutions etc., PP 1859, Session 2, XII.

Public Prosecutors – Select Committee Report on Public Prosecutors, PP XII, 1854/5.

Supplementary Votes, PP 1868-9, XLII.

Prosecution allowances – Departmental Committee on allowances to prosecutors and witnesses, Cd 1650, 1903.

Archives : UK Public Record Office

Home Office series H045, HO 329 and 347.

Treasury T28, and T165.

Haut de page


2  Taylor (1998a, 1998b and 1999).

3  The principal examples of modern historical accounts are perhaps Gatrell and Hadden (1972), Gatrell (1980) and Radzinowicz and Hood (1986, pp. 91-112). The same ground is more accessibly covered in Emsley (1997).

4  Official Statistics Committee (1881), memorandum of November 1876 by Robert Giffen, head of statistics in the Board of Trade, which amongst other things excoriated (p. 97) the Home Office Judicial, Prison and other related statistics as containing « duplicate and superfluous matter ».

5  At the same time it has to be borne in mind that private for long outnumbered public legislation in the UK Parliament. For example, it has been calculated that during 1800-1884 there were 18 497 private as opposed to 9556 public statutes. See Prest (1990, p. 4).

6  The Bill introduced flogging for « garotters », and was memorably castigated by Grey as « a panic measure after the panic had subsided », Official Report, Commons, 11 March 1863, col. 1311.

7  Maitland (1885, p. 148).

8  The rejection on 20 January 2000 by the House of Lords of the government's Criminal Justice (Mode of Trial) Bill (which extended changes initiated in principle by the Criminal Justice Act 1855 and would have further qualified the right of defendants to opt for jury trial) may be seen to represent, among other things, the outcome of continuing tensions between centre and locality in the administration of criminal justice in England and Wales. In Scotland defendants have never had any such rights in a system where the mode of trial is fixed by the Procurator Fiscal, that is the public prosecutor.

9  Described as « the institution of a centralised police force », an unusual misdescription in the compendious and thoughtful meditation on the criminal statistics, policing and social change in Gatrell (1980, p. 259). The most recent account of the decades before the 1856 Act pays particular attention to the political background : Philips, Storch (1999).

10  It has been pointed out that one of the principal reasons why magisterial rule in the English and Welsh counties lasted so long (i.e. until 1888) was because of « the reluctance of MPs to sanction an assault on their own class » – Dunbabin (1977, p. 79).

11  See note 3 above.

12  Radzinowicz, Hood (1986, p. 107).

13  Criminal Statistics 1894 (1896), preface dated 1 May 1896, p. 30 and pp. 34-35. He repeated the health warning in the case of the 1895 Statistics even more emphatically : « There is some basis of truth in the remark made by an experienced chief constable, the ’amount of crime in a county is a matter of book-keeping' » – Criminal Statistics 1895 (1897), preface of 6 February 1897, p. 35. Exactly the same point has been made with a wealth of contemporary detail in a modern police account – see Young (1991), for example pp. 377-381.

14  Kitsuse, Cicourel (1963), p. 137 « Thus, rates can be viewed as indices of organizational processes rather than as indices of certain forms of behaviour ».

15  Minute of 3 May 1929 by Arthur Locke, whom the Home Secretary appointed to chair the committee set up as a result – HO 329/108-R843/56B. Because of Locke's death, the committee never produced a formal report, though it was considered that its objects had been achieved in practice.

16  Perks Report (1967). For recent changes in the counting rules, see Home Office Statistical Bulletin 18/99. For an illuminating discussion of counting ambiguities in the comparable French system, see Aubusson de Cavarlay (1998).

17 Times and Guardian reports, 1.8.2 000.

18  Downes, Rock (1998, p. 50). On the other hand, it has been pointed out by a criminologist that the « standard of statistical sophistication in criminology is not high ». Pease (1999, p. xi).

19  See Note 2 above.

20  Taylor (1998a, p. 571).

21  Morris, Blom-Cooper (1964). The authors were respectively a Professor of Criminology at the London School of Economics and a leading criminal barrister and Queen's Counsel.

22  Taylor (1998a, p. 85).

23  Gurr (1981, p. 99).

24  For example, by Havard (1960).

25  Broderick Committee (1971).

26  Taylor (1998a, p. 586).

27  The exchange – Official Report, Commons, 27 July 1889, column 888 – concerned whether the heads of departmental prosecution units were able to retain for their own use any surpluses from the voted monies delegated to them for managing their offices. (The reply explained, of course, that they could not.) There is evidence elsewhere that Supplementary Votes to defray unexpectedly high prosecution costs were sought as necessary – see PP 1868-9, XLII, pp. 523-524 where the Treasury explained that a request for an extra £12000 (the second largest of the Supplementaries being sought) was necessary because an « increase has also taken place in the Amounts claimed for the Repayment of the Costs of Prosecution ». The Treasury Blue Book for 1880-1 reviewing the financing system for prosecutions also makes it clear that the normal Parliamentary regimes applied – T 165/1/3 pp. 325-329. Further, the quotation (Taylor, 1998a, p. 85) from an 1890 paper by the Home Office clerk, Grosvenor, then responsible for producing the annual criminal statistics (viz. « any marked variation observed from year to year are at once noted and an explanation required by officials ») does not imply that the Home Office role was exercised to keep prosecutions down rather than operating merely at a clerical level in checking returns and pursuing apparent discrepancies. It was essentially the same concerns that motivated the setting up of the Home Office committee in 1929 to produce more uniform returns (see note 15 above) as well as other attempts down the years to improve the quality of the statistics.

28  The reference (Taylor 1998a, p. 587) to an anecdote in Robert Mark's memoirs seems a mistake. The point Mark was illustrating – Mark (1978, p. 26) – was the steps taken to pass responsibility for investigation of a suicide on to a neighbouring force rather than trying to prevent investigation at all. Such boundary deceptions to evade an unrewarding responsibility are almost certainly as old as the police service – see, for example, Cavanagh (1893, p. 23) and a night duty story about another officer in 1850s Southwark moving a drunken woman from one side of the street on to Cavanagh's side so that he rather than the other would lose sleep by having to take the woman before the magistrate in the morning after his shift.

29  Taylor (1998a, p. 571).

30  County Rates (1836, p. 8).

31  Oddly, the literature seems devoid of any comprehensive account. The closest approach is probably still Chapter 16 (« Evolution of the Office of the Director of Public Prosecution ») in Edwards, J. LI. J. (1964). But this relied solely on published sources and is preoccupied more with the conduct of prosecutions than their initiation. Nor do other standard texts help e.g. Holdsworth (1965) or Devlin (1958).

32  See Philips and Storch (1999) for a thorough discussion of these processes in the context of policing.

33  Taylor (1998a, pp. 574-575).

34  Prosecutions Royal Commission (1859), Evidence, Q 120.

35 Ibid., Appx 7, letter of 17 April 1858.

36  T28/74-81.

37  T28/81 – « Rules to be observed in relation to the Costs of Criminal Prosecutions which are repayable from the Parliamentary Grant », 17 August 1869.

38  1859, op. cit., Evidence, Q 148. See HO 347/9 for Home Secretary's Order of 9 February 1858, under s.  of the statute 14&15Vic.c. 5.

39  1859, op. cit., Q120. Joseph Hume, the Parliamentary apostle of administrative and financial efficiency and rectitude had warned against this in 1835 – see his evidence at Q 852, County Rates Commission (1836).

40  HO 45/9305/11932 has a copy of the judgement (« ... Nothing can be more anomalous, nothing can be more unsatisfactory, than the present system with regard to the taxation of these bills of costs... ») of 29 January 1872. The Treasury's subsequent letter of 22 April 1872, attempting to reinforce its regime, was greeted (minute of 27 April 1872) by the Home Office Parliamentary Under Secretary : « This Treasury letter is as wrong as wrong can be... he whole scheme of the Treasury is clumsy and irritating » (he was no doubt influenced by the mood of the debate following the judgement – Official Report, Commons, 15 March 1872, cols. 51-72).

41 Ibid., Treasury Minute of 29 January 1875. See T165/1/3, p. 28, for private 1880-1 Treasury Blue Book recantation.

42  Prosecution allowances (1903, p. 27).

43  Davis (1989, p. 400).

44  HO 347/11 pp. 563-612. This lengthy printed memorandum appears to have been prepared to assist with finalising the 1873 Bill.

45  Prosecutions Royal Commission (1859), Note of dissent by R. Upperton.

46  Public Prosecutors (1854/5), evidence of Horatio Waddington, Home Office Permanent Under Secretary, Q 2072.

47  Stephen (1883, Vol. I, p. 497).

48  Taylor (1999, p. 117). Similar claims are made on pp. 118, 120 and 121.

49  Taylor (1999, p. 117). Similar claims are made on pp. 118, 120 and 121.

50 Ibid., pp. 126 and 128.

51  For example, Wiener (1994).

52  Martin and Wilson (1969), Chapter XI.

53  Gatrell, Hadden (1972 p. 368) ; Field (1999). The perception of such linkages is not, of course, new. Writing in 1875, Sir George Grey (Home Secretary on three occasions between 1841 and 1866 for a total of nearly 14 years) claimed that the phenomenon had been remarked on during the Irish famine in the late 1840s. Aberdare Letters (1902, Vol. 2, p. 3), letter of 9 October 1875, p. 33.

54  Maguire (1997, p. 141).

55  Simon Jenkins, Times, 1 December 1999.

Haut de page

Pour citer cet article

Référence papier

Robert M. Morris, « « Lies, damned lies and criminal statistics » : Reinterpreting the criminal statistics in England and Wales », Crime, Histoire & Sociétés / Crime, History & Societies, Vol. 5, n°1 | 2001, 111-127.

Référence électronique

Robert M. Morris, « « Lies, damned lies and criminal statistics » : Reinterpreting the criminal statistics in England and Wales », Crime, Histoire & Sociétés / Crime, History & Societies [En ligne], Vol. 5, n°1 | 2001, mis en ligne le 02 avril 2009, consulté le 19 juillet 2018. URL : ; DOI : 10.4000/chs.784

Haut de page


Robert M. Morris

The Open University, European Center for the Study of Policing, Walton Hall, MILTON KEYNES MK7 6AA (U.K.),
R.M. Morris studied History at Cambridge and worked as an administrator in the Home Office 1961-97, including on police, prisons and criminal justice issues. He is currently engaged (under Professor Clive Emsley at the Open University, UK) on a doctoral study of the London Metropolitan police 1860-1920. A version of the paper was originally given at the Seminar « Questions de polices/Policing Matters » (GERN-CNRS), Maison des Sciences de l'Homme, Paris, 10-11 December, 1999.

Articles du même auteur

Haut de page

Droits d’auteur

© Droz

Haut de page
  • OpenEdition Journals