Portal de Periódicos da CAPES

Statistics as the science of government: The stillborn British Empire statistical bureau, 1918–20

2005; Taylor & Francis; Volume: 33; Issue: 3 Linguagem: Inglês

10.1080/03086530500185860

1743-9329

Jean-Pierre Beaud, Jean‐Guy Prévost,

Census and Population Estimation

Abstract The 1920 British Empire Statistical Conference was the direct outcome of the Dominions Royal Commission's Final Report, which had spelt out the need to increase the uniformity and comparability of statistics originating from various parts of the Empire and had proposed setting up an imperial central statistical office. Over 24 days, delegates debated a large number of topics, ranging from the practical and empirical subject matters of statistical inquiry to more abstract issues such as the nature and object of statistical data collection and analysis, and to the problems raised by the establishment of a statistical bureau that would operate on an unprecedented scale. This article seeks to understand why, despite apparently favourable conditions, this project soon ended in complete failure. The reasons must be sought in the neatly distinctive outlooks held by the British government and Dominion representatives as regards the function of statistics for the purpose of government, in the quite different bureaucratic settings that embodied and sustained these views, as well as in the tensions and centrifugal pressures that acted upon inter-imperial relations following the Great War. The authors would like to thank Mrs Beth Wright from the Australian Bureau of Statistics for her help and comments as well as Canada's Social Science and Humanities Research Council for its financial contribution. An earlier version of this article was presented at the British World Conference, Melbourne, July 2004. Notes 1. On the international statistical congresses, see Eric Brian, 'Statistique administrative et internationalisme statistique pendant la seconde moitié du XIXe siècle', Histoire et mesure, 4, 3–4 (1989), 201–24. On the ISI, see J.W. Nixon, A History of the International Statistical Institute 1885–1960 (The Hague, 1960). 2. According to data gathered in W. Gregory (ed.), International Congresses and Conferences 1840–1937 (New York, 1967 [1938]). 3. A. Desrosières, La politique des grands nombres: Histoire de la raison statistique (Paris, 1993), 43–48. 4. S.N.D. North, 'Seventy Five Years of Progress in Statistics: The Outlook for the Future', in J. Koren (ed.), The History of Statistics: Their Development and Progress in Many Countries (New York, 1918), 20–21, 47. 5. These discussions can be traced back to the 1907 Colonial Conference, where a resolution 'to secure greater uniformity in the trade statistics of the Empire' was adopted – M. Ollivier (ed.), Colonial and Imperial Conferences from 1887 to 1937, Vol. II, Imperial Conferences, Part I (Ottawa, 1954), 222 – and from then on to the 1911 Imperial Conference which set up the Dominions Royal Commission, whose inquiries began in 1912 and whose final report was issued in 1917, and, finally, to the 1918 Imperial War Conference, from which emerged a resolution stating that 'having considered the correspondence as to the improvement of Imperial Statistics arising out of the recommendations of the Dominions Royal Commission, (the Imperial War Conference) is in favour of the proposal to hold a Conference of Statisticians after the war, and that such conference consider the establishment of an Imperial Statistical Bureau under the supervision of an Inter-Imperial Committee', ibid., 272. 6. Dominions Royal Commission, Final Report (London, 1918), 361–64. 7. Ibid., 365–69. 8. British Empire Statistical Conference, Report and Resolutions adopted by the First Conference of Government Officers engaged in dealing with Statistics in the British Empire (with four appendices) (London, 1920), 12–16. 9. Ibid., 18–42. 10. D.A. Worton, The Dominion Bureau of Statistics: A History of Canada's Central Statistical Office and its Antecedents, 1841–1972 (Montreal, 1998), 172–73. 11. A seminal work is this area has been Desrosières, La politique des grands nombres. 12. Among recent studies conducted along these lines and bearing upon quite different contexts, see B. Curtis, The Politics of Population: State Formation, Statistics, and the Census of Canada, 1840–1875 (Toronto, 2001); and A. Tooze, Statistics and the German State, 1900–1945: The Making of Modern Economic Knowledge (Cambridge, 2001). The case of Germany is relevant here, since the Imperial German statistical system was sometimes taken as a model of what the British Empire could – or should – achieve. 13. For instance, in an address read before the Royal Statistical Society on 19 Dec. 1916, its Vice-President, Geoffrey Drage, had devoted significant space to the 'lack of uniformity' in imperial statistics, 'The Reorganisation of Official Statistics and a Central Statistical Office', Journal of the Royal Statistical Society, 80 (1917), 31–64, Drage, 43–47. Drage was also among the individuals who had advocated such views before the Dominions Royal Commission (during its London evidence-gathering meetings in 1912) and is quoted as such three times in its final report. 14. P.J. Cain and A.G. Hopkins, British Imperialism: Crisis and Deconstruction 1914–1990 (London and New York, 1993), 109–10. 15. The term seems to have been coined by Richard Jebb, Studies in Colonial Nationalism (London, 1905), in order to describe the specific type of nationalism that emerged at the turn of the century in the 'white' Dominions (Australia, Canada, New Zealand and South Africa), the acceptance of which he considered a condition for maintaining and strengthening the Empire. 16. The minutes, which include a full verbatim report of discussions held during the plenary sessions of the conference, are gathered in 24 volumes, one for each day of the conference, under the general title British Empire Statistical Conference, London, England, Jan.20 – Feb.26, 1920, Minutes, Dominion Bureau of Statistics – Library. These documents were obviously not intended for publication (there is a 'confidential' label on the front page of each volume and a number of corrections in Dominion Statistician Robert H. Coats's handwriting) and, to our knowledge, the only surviving exemplar is kept in Statistics Canada's library (catalogue number HA12.5C5). Overall, the documents represent a total of some 1,500 pages. In his presidential address to the Royal Statistical Society a few months later, RSS – 16 Nov. 1920, Sir Henry Rew, 'The Organization of Statistics', Journal of the Royal Statistical Society, 84 (1921), 18, mentioned that reasons of public economy seemed to have prevented 'the publication of a full report of the proceedings, which would have been of great interest to all persons concerned in statistical progress', thereby confirming their strictly confidential character. 17. Ernest Godfrey, from Canada, who had had some previous experience at international meetings, contracted double pneumonia while travelling to Britain and could not therefore attend the sessions, Worton, The Dominion Bureau of Statistics, 372. 18. British Empire Statistical Conference, London, England, Jan.20 – Feb.26, 1920, Minutes (Ottawa, 1920), I, 43–46. From now on, we will refer to this document as Minutes, giving the volume and page. 19. On Knibbs's career, see S. Bambrick, 'Knibbs, Sir George Handley', Australian Dictionary of Biography (Melbourne, 1983), 20–21; and C.C. Heyde, 'Official Statistics in the Late Colonial Period Leading on to the Work of the First Commonwealth Statistician, G.H. Knibbs', Australian Journal of Statistics, 30 (1988), 23–43. Knibbs's quotation comes from his 1910 presidential address to the Australia and New Zealand Association for the Advancement of Science, 'The Problem of Statistics'. 20. G.H. Knibbs, Memorandum on the Proposed Conference of the Statisticians of the British Empire (London, 1920 [1918]). 21. J.K. Whitaker, 'Flux, Alfred William (1867–1942)', in J. Eatwell, M. Milgate and P.K. Newman (eds.), The New Palgrave: A Dictionary of Economics (New York, 1987), II, 395–96. 22. S.R.S. Szreter, Fertility, Class and Gender in Britain, 1860–1940 (Cambridge, 1996), 74. Since the RSS awarded only ten gold medals between 1892 and 1945, this was indeed a highly valued distinction. 23. Quite representative of Knibbs's organicism is his presentation to the Royal Colonial Institute on 9 Dec. 1920, 'Statistics and National Destiny', United Empire, 11, 1 (1920), 14–23, where he advocates a 'national eugenics' as well as 'a large measure of State interference' and the 'co-ordination of national industry'. 24. On Coats, see N. Keyfitz and H.F. Greenway, 'Robert Coats and the Organization of Statistics', Canadian Journal of Economics and Political Science, 27, 3 (1961), 313–22; and D.A. Worton, 'Robert H. Coats, Architect of Canada's Statistical System', in J.-P. Beaud and J.-G. Prévost (eds.), L'ère du chiffre: Systèmes statistiques et traditions nationales (The Age of Numbers: Statistical Systems and National Traditions) (Sillery, 2000), 87–104. 25. 'Statistics', Standard Encyclopedia of Southern Africa (Cape Town, 1974), 256. 26. D. Green, Statistics Count: An Illustrated History of Statistics New Zealand (Wellington, 2002), 6–8. 27. Rew, 'The Organization of Statistics', 15. 28. Szreter, Fertility, Class and Gender in Britain, 263. 29. G.H. Knibbs, Statistical Conference (British Empire): Report and Resolutions accompanied by Explanatory Memorandum and Observations by the Commonwealth Statistician (Melbourne, 1920), 6. 30. British Empire Statistical Conference, Report and Resolutions, 42–43. 31. In the address he delivered before the RSS on the first night of the conference, Knibbs was quite outspoken on this issue: 'it is not at all unusual to hear Canada, New Zealand, South Africa and Australia, still referred to as Colonies, and to find that the point of view – even of men of education and information – corresponds to the name', 'The Organization of Imperial Statistics', Journal of the Royal Statistical Society, 83 (1920), 204. 32. Knibbs, Statistical Conference (British Empire), 4. 33. His 1918 memorandum was the only elaborate document dealing in a comprehensive manner with the theme of the conference (the Board of Trade had laid no less than 11 papers before the delegates, but they were all related to specific matters) and the above-mentioned address, which can be described as a rhetorical version of the memorandum, kept surfacing throughout the conference as the main intellectual reference point for most delegates. Whether they agreed or not with his views, they could not but pay tribute (or lip-service) to his views. 34. British Empire Statistical Conference, Report and Resolutions, 44–45. (C) stands for chairman. 35. Minutes, V, 20. Comparative analysis of Canadian exports and imports to the United Kingdom and to the United States had become a standard feature of the Canada Year Book under Coats. 36. Minutes, I, 18–31. 37. All following quotations on the issue of classification come from Minutes, V, 9–20. 38. A public and systematic exposé of the Canadian views would be made in R.H. Coats, 'Classification Problem in Statistics', International Labour Review, 11, 4 (1925), 509–25. On this issue, see J.-P. Beaud and J.-G. Prévost, 'La classification canadienne des occupations pendant l'entre-deux-guerres: un cas d'indépendance statistique', Canadian Journal of Political Science, 25, 3 (1992), 489–512. 39. In his memorandum, Knibbs had taken a median position on this issue, deeming impossible 'the attainment of a really systematic classification', but maintaining at the same time that there was a 'necessity for standardization' of existing schemes and the conference would provide the opportunity for a compromise on this issue, Memorandum on the Proposed Conference of the Statisticians of the British Empire, 16. 40. Ibid., 7. 41. Biennial sessions of the International Statistical Institute (ISI) had been since the late 1880s quasi-ceremonial occasions of sharing these visions and language. As a matter of fact, writings such as Knibbs's 'Statistics and National Destiny' exemplify a minor genre to which contributed statisticians from a number of countries and which can be defined as a combination of positivism, reformism and self-promotion. The most elaborate expression of this vision is to be found in John Koren's The History of Statistics (1918), to which 20 authors representing 17 countries contributed (Knibbs wrote the chapter on Australia). 42. Minutes, I, 30. 43. 'The Organization of Imperial Statistics', 202. 44. Knibbs, 'The Evolution and Significance of the Census', Addresses and Proceedings of the Imperial Federal League of Australia (Melbourne, 1910), 16. 45. Minutes, II, 22. In the margin of his copy of the minutes, Coats wrote, besides Vivian's intervention, the following comment: 'Nonsense'. 46. 'The Organization of Imperial Statistics', 202. 47. Minutes, I, 13. 48. Minutes, II, 7. 49. Ibid., 8–13. 50. Ibid., 25. 51. Ibid., 20. 52. 'The Organization of Imperial Statistics', 203–08. This description – that must obviously have sounded to his listeners as a flattering self-portrait – was obviously setting the pass mark pretty high. It was at this point that Knibbs added: 'The method of appointing a figure-head who does not know his subject, hoping that some underling will make up for deficiencies – a method not entirely unknown in this world – should not be followed' (author's emphasis in all cases). 53. Knibbs, Memorandum on the Proposed Conference of the Statisticians of the British Empire, 8. 54. M. Ollivier (ed.), Colonial and Imperial Conferences from 1887 to 1937, II, Part I, 303. 55. Minutes, XVI, 24. 56. The mandate of the Imperial Mineral Resources Bureau, which was set up following the Imperial War Conference, was to collect, co-ordinate and disseminate all statistical information relating to mineral resources of the Empire as well as those of foreign countries. Even though he did not appear on the list of participants in the final report (though he is listed among members of the General Purposes Committee), its chairman, Sir Richard Redmayne, was present at the 1920 conference. Delegates had also received knowledge of the imminent establishment of an Imperial Agricultural Bureau that could equally be perceived as a source of potential duplication. 57. Minutes, XIX, 13–15. 58. Worton, The Dominion Bureau of Statistics, 170. 59. Ibid. Fraser had sent an elaborate report dated 26 April 1920. The report was later published together with a second one, dated 13 Aug. 1920, where he related also his visit to South Africa and Scotland and his mission to Canada and the United States (where he had been sent to inquire mainly about 'mechanical appliances for statistical analysis and accounting work'). M. Fraser, British Empire Statistical Conference (Report of Government Statistician on) (Wellington, 1920). 60. R.M. Ferguson, Letter to the Secretary of State for the Colonies, 9 Aug. 1920, National Archives of Australia, Series A 11804, Item 1920/714. As a matter of fact, this was a couple of weeks before, on 25 Aug., Canada nominated its high commissioner in London, Sir George E. Perley, to represent it in drafting the projected Bureau's charter. 61. Knibbs, Statistical Conference (British Empire), 2–7. 62. Beth Wright, from the Australian Bureau of Statistics, has expressed such a view (personal communication), which had also been that of some of those who listened to Knibbs's addresses to the Royal Colonial Institute in Dec. 1919 and to the RSS in Jan. 1920. See the discussions in 'Statistics and National Destiny', 22–23; and 'The Organization of Imperial Statistics', 214–24. Wright has called to our attention the fact that C. Forster and C. Hazlehurst concur with this view in their Australian Statisticians and the Development of Official Statistics (Canberra, 1988) and see the detailed character of Knibbs's memorandum, including specifications about machinery and library equipment, as proof of a personal as well as professional interest. 63. In a letter to his successor, Charles H. Wickens, written at a time when he had moved from his Commonwealth Statistician position to that of director of the new Commonwealth Institute of Science and Industry, Knibbs wrote: 'The trouble about the British conference was that even after having some sort of a decision, they went back on the whole thing owing to the water-tight compartment system of the British Public Service'. Letter to C.H. Wickens, 4 Oct. 1922, National Archives of Australia, Series A8510, Item 175/2. 64. Minutes, IX, 5. 65. Something confirmed later by Coats and which explains why Flux did not preside the census committee: 'we had actually at one stage to re-form the Conference, because the population fellows wouldn't attend a meeting presided over by a Board of Trade man …', Letter to Sir George Schuster, 5 Nov. 1934, National Archives of Canada, RG 31, vol.1405, File 1454. 66. Knibbs, 'Statistics and National Destiny', 18. 67. Fraser, British Empire Statistical Conference, 10–12.