by R. G. D. Allen, rev. John Bosnell
© Oxford University Press 2004 All rights reserved
Bowley, Sir Arthur Lyon (1869-1957), statistician, was born at 12 King's Square, Bristol, on 6 November 1869, the son of the Revd James William Lyon Bowley (1826-1871), vicar of St Philip and St Jacob, and his wife, Maria Johnson. Bowley spent nine years at Christ's Hospital (1879-88), which left a lasting impression on him; in later life he served as a governor of the school for more than ten years. He went to Trinity College, Cambridge, with a major scholarship in mathematics, was bracketed tenth wrangler in 1891, and later obtained both the Cobden and the Adam Smith prizes. He was awarded the ScD by his university in 1913.
On leaving Cambridge, Bowley seemed destined to teach mathematics in schools and he was on the staff, briefly at Brighton College, then at St John's School, Leatherhead, until 1899. His own interests lay in the application of economic analysis to problems of social reform. He published in 1893 his Cobden prize essay, on England's foreign trade in the nineteenth century, then began an extensive series of research projects, mainly with G. H. Wood, on the relationship between movements in wages and prices; he read his first paper to the Royal Statistical Society in 1895. He published many further papers on this subject, all of them put together with great historical and statistical care.
These interests led, in 1895, to a complete and permanent change in Bowley's career. In that year Sidney Webb and others founded the London School of Economics (LSE) and assembled a small staff of part-time experts. Bowley was chosen to take charge of the teaching of statistics, and taught at LSE without interruption for more than forty years. He was never a socialist in Webb's sense but, as a good Liberal, he found the senior common room a congenial and stimulating background to his teaching, research, and public service.
In 1904 Bowley married Julia (1871-1959), daughter of Thomas Hilliam, land agent; they had three daughters, one of whom, Marian Bowley, became professor of political economy at University College, London. Bowley also taught at the University College at Reading from 1900, as lecturer in mathematics (1900-7), in economics (1913-19), and professor in both subjects (1907-13). Meanwhile, at the LSE, he became part-time reader in statistics in 1908 and was given the title of professor in 1915. When the University of London created a full-time chair in statistics in 1919, tenable at the LSE, Bowley became the first occupant. Although he retired from the chair in 1936, he continued many of his activities, both at the LSE and elsewhere, until the early fifties. He acted as director of the Oxford University Institute of Statistics during the war years (1940-44) and received an honorary DLitt in 1943. He was elected FBA in 1922, appointed CBE in 1937, and knighted in 1950.
As a mathematician Bowley was competent but rather old-fashioned. He was one of the founders of the international Econometric Society in 1933 and served later as its president. However, first and foremost he was a practitioner in statistics applied across the whole of the social sciences. He was highly regarded by official statisticians but it was not then the custom for the British government to call upon outside experts for advice, and he exercised his main influence therefore through his teaching and research on the one hand and his extensive international contacts on the other. He was called upon as an expert witness in cases such as the inquiry into dockers' wages (1920).
Two of Bowley's pioneer activities were in the economic field. One comprised a number of studies on the definition and measurement of national income which occupied his attention, on and off, for more than twenty years before the first official estimates were made under Lord Keynes during the Second World War. Without Bowley's careful and precise work the official computations would scarcely have been possible. His other pioneer activity in this field was from 1923 with the London and Cambridge Economic Service, the first venture of this kind in Britain. Bowley was its first editor and he served in this capacity continuously until 1945, remaining a regular contributor until 1953. His editorship was characterized both by the skill with which he pulled together various views in his own succinct assessment of the current economic position for publication in the Monthly Bulletins of the service, and by the statistical techniques he used in designing economic series and in devising ways of presenting them. He showed how economic analysis depends on long runs of comparable series, presented in graphical form (for example by the use of ratio scales) and adjusted where necessary for seasonal variation.
The major contribution which Bowley made--and it was one of the utmost importance to statistics--lay in the development of sampling techniques in their application to social studies. He explored, largely for the first time, the appropriate design of sample surveys, the proper formulation of sampling precision, and the ways of interpreting the results in their application. He devised and conducted sample surveys of working-class households in four English towns and, in presenting the results in Livelihood and Poverty (with A. R. Burnett-Hurst, 1915), he was far ahead of his time both in explaining the methods used and in formulating the precision of the results. He distinguished four sources of error: incorrect information, loose definitions, bias in selection of samples, and calculable errors of sampling. He may not have been entirely correct in his use of what is now known as cluster sampling, but what he wrote in 1915 remains relevant.
It was natural that Bowley should dominate the committee set up in 1924 by the International Statistical Institute to report on the representative method. At the 1925 session in Rome Bowley's influence was clearly visible in the main recommendation of the committee that:
Bowley was effective, if rather dour, on committees, and held many high offices in the British Association, the Royal Statistical Society, the Royal Economic Society, and the International Statistical Institute. He was shy and retiring, never happier than when talking quietly to his research students or playing Bach with his family. He was respected by all his colleagues and students but intimate with few. One of his close friendships was with the distinguished economist Edwin Cannan, who shared his enthusiasm for cycling. Bowley died at the Otara Nursing Home, Fernhurst, near Haslemere, on 21 January 1957. He was survived by his wife.
R. G. D. ALLEN, rev. JOHN BOSNELL
Sources
A. H. Bowley, A memoir of Professor Sir Arthur Bowley (1972)
R. D. G. Allen and R. F. George, Journal of the Royal Statistical Society: series A, 120 (1957), 236-41
Nature, 179 (1957), 398-9
E. P. Hennock, 'The measurement of urban poverty: from the metropolis to the nation, 1880-1920', Economic History Review, 2nd ser., 40 (1987), 208-27
The Times (23 Jan 1957), 12a
The Times (12 Feb 1957), 10c
CGPLA Eng. & Wales (1957)
Archives
BLPES, corresp., lecture notes, and papers
Royal Statistical Society, corresp. and papers | BLPES, letters to Edwin Cannan
King's AC Cam., letters to John Maynard Keynes
U. St Andr. L., letters to Sir D'Arcy Thompson
Likenesses
S. Bowen, oils, 1936, London School of Economics [see illus.]
photograph, c.1939, repro. in Bowley, Memoir of Professor Sir Arthur Bowley
photograph, repro. in The Times (23 Jan 1957)
Wealth at death
£10,549 7s. 10d.: probate, 13 March 1957, CGPLA Eng. & Wales
[http://www.oxforddnb.com/view/article/32006]
GO TO THE OUP ARTICLE (Sign-in required)