the British welfare state. W. Beveridge and J. M. Keynes greatly in¬‚uenced his

approach to economics and statistics. During World War II he ¬rst (in 1940) worked

in the statistical section of the Prime Minister™s of¬ce. Then Prime Minister Winston

Churchill asked his “think tank” to explore questions about home and enemy

resources, and the unit™s head, Professor F. A. Lindemann (later Lord Cherwell)

became, in turn, extremely demanding of his staff. Champernowne did not get on

well with him, and in 1941 he moved to the Ministry of Aircraft Production, staying

there for the remainder of the war.

In 1944 Champernowne got involved with J. M. Keynes of the Cambridge

economics faculty in establishing a separate Department of Applied Economics. In

1945 he became Director of the Oxford Institute of Statistics and a Fellow of

Nuf¬eld College, attaining a full professorship in 1948.

In 1948 Champernowne married Wilhelmina (“Mieke”) Dullaert who had had the

good grace to lose to “Turochamp” in their chess matches. They had two sons. That

same year he also presented a pioneering paper at the Royal Statistical Society on the

Bayesian approach to the time series analysis of autoregressive processes. A related

paper presented in 1960 dealt with the spurious correlation problem, anticipating the

work of Granger and Newbold (Journal of Econometrics, 1974) on this topic.

However, Champernowne came to regret his change of university and began to look

for a way to return to Cambridge. This he could achieve only through a highly

unusual step, by assuming a lower academic rank, which he did when he became a

reader in economics at Cambridge and a teaching fellow at Trinity from 1959 “1970.

During the subsequent decade Champernowne combined his work on income

distributions with investigations of uncertainty in economic analysis in a three-

volume book: Uncertainty and Estimation in Economics (1969). In this magnum

opus he attempted to integrate quantitative analysis with decision theory. This unique

(but currently overlooked) study resulted in Champernowne™s being given a personal

chair at Cambridge in economics and statistics and his election as a Fellow of the

British Academy in 1970. During 1971 “ 1976 he was a co-editor of the Economic

Journal. As mentioned above, his economic studies in the early 1930s were

supervised by J. M. Keynes and he made a valuable contribution in his published

275

REFERENCES

review of Keynes™s General Theory (1936) to the debate between Keynes™s approach

and upholders of the classical theory of employment, speci¬cally whether money or

real wages are the proper subject of bargaining between management and workers.

He courageously (at the age of 23) took a controversial stand, respectfully

disagreeing with his eminent supervisor and history proved him right.

From 1935, for over 50 years, Champernowne carried out research on income dis-

tributions, producing a rigorous stochastic model using the technique of the Markov

chain (see Chapter 3 for a brief description of his model leading to the Pareto

distribution). His 1937 work earned him a prize fellowship at King™s College and

was widely discussed and consulted by researchers worldwide. However, this work

was only of¬cially published some 36 years later (Champernowne, 1973). His last

book Economic Inequality and Income Distribution he wrote jointly with a former

student, Frank Cowell; it was published in 1998.

Champernowne retired in 1978 but remained an emeritus professor for almost

20 years. When he fell ill in his ¬nal years, he moved with Mieke to Budleigh

Salterton in Devon, close to his family roots”the birthplace of Sir Walter Raleigh

and to be near one of his sons.

REFERENCES

Champernowne, D. G. (1933). The construction of decimals normal in the scale of ten.

Journal of the London Mathematical Society, 8, 254“ 260.

Champernowne, D. G. (1936). Unemployment, basic and monetary: The Classical analysis

and the Keynesian. Review of Economic Studies, 3, 201“ 216.

Champernowne, D. G. (1969). Uncertainty and Estimation in Economics. 3 Vols. London:

Oliver & Boyd.

Champernowne, D. G. (1973). The Distribution of Income Between Persons. Cambridge:

Cambridge University Press.

Champernowne, D. G., and Cowell, F. A. (1998). Economic Inequality and Income

Distribution. Cambridge: Cambridge University Press.

Harcourt, G. C. (2001). David Gawen Champernowne, 1912 “ 2000: In appreciation.

Cambridge Journal of Economics, 25, 439 “442.

Obituary in Daily Telegraph, Sept. 4, 2000.

Obituary in Guardian, Sept. 1, 2000.

APPENDIX B

Data on Size Distributions

Below we provide a list of data sources, mostly of data in grouped form, along with

information on the distributions that were ¬tted to these data. Our list is in no way

exhaustive and merely pertains to the references presented in this book.

Sources for large data sets on individual incomes include the U.S. Panel Study of

Income Dynamics (PSID), maintained by the University of Michigan (Hill, 1992)

and the German Socio-Economic Panel (SOEP), maintained by the Deutsches

Institut fu Wirtschaftsforschung (DIW), Berlin (Burkhauser, Kreyenfeld, and

¨r

Wagner, 1997). Further actuarial data sets may be found in textbooks on

actuarial statistics such as Klugman, Panjer, and Willmot (1998).

Table B.1 Individual Data (Incomes)

No. of

Source Description Observations Distributions Fitted

Aggarwal and Kenyan annual earnings 200 No distribution ¬tted

Singh (1984)

Arnold (1983) Lifetime earnings of 50 Par, Par(II), Fisk, SM

professional golfers

Arnold (1983) Texas county data (1969 157 Par, Par(II), Fisk, SM

total personal incomes)

Dyer (1981) Annual wages of production-line 30 Par

workers under age 40 in large

U.S. industrial ¬rm

277

278

Table B.2 Grouped Data (Incomes)

Source Description No. of Groups Distributions Fitted

Anand (1983) Malaysia 1970 32 Par

Aoyama et al. (2000) Japan 1998 18 Par

Bordley, McDonald, and United States 1970, 1975, 1980, 9 15 distributions of gamma and

Mantrala (1996) 1985, 1990 beta type

Bowley (1926) UK 1911“ 1912 11 Par

Bowman (1945) United States 1935“ 1936 25 No distribution ¬tted

Brunazzo and Pollastri (1986) Italy 1948 21 Generalized LN

Champernowne (1952) United States 1918 18 4-parameter Champernowne,

Davis

Champernowne (1952) Norwegian townsmen 1930 11 3-parameter Champernowne

Champernowne (1952) United States 1947 12 3-parameter Champernowne

Champernowne (1952) Bohemia 1933 16 3-parameter Champernowne,

Davis, Par, LN

Champernowne (1952) UK 1938“ 1939 12 3-parameter Champernowne

Champernowne and Cowell UK 1994“ 1995 16 No distribution ¬tted

(1998)

Cowell and Mehta (1982) Sweden 1977 21 No distritbution ¬tted

Creedy, Lye, and Martin (1997) United States 1986 15 LN, generalized LN, GG,

generalized Ga

Dagum (1980a) United States 1969 10 LN, Ga, D

Dagum (1983) United States 1978 21 LN, Ga, SM, D I-III

Dagum (1985) Canada 1965, 1967, 1969, 1971, 16 D I, II, and III

1973, 1975, 1977, 1979, and

1981

Dagum and Lemmi (1989) Italy 1977, 1980, and 1984 15 D I, II, and III

Davies and Shorrocks (1989) Canada 1983, 1984 18, 12* No distribution ¬tted

Davis (1941b) United States 1918 73 Davis

Espinguet and Terraza (1983) France 1978 11 D II

Fattorini and Lemmi (1979) Italy 1967“ 1976 19 Champernowne, D I

Gastwirth and Smith (1972) United States 1955 26 2- and 3-parameter LN, Fisk

Hayakawa (1951) Japan 1932 34 Par

Iyengar (1960) India 1955“ 1956 12 LN