⁽¹⁾ Appendix II gives a fuller account of some of the problems of assessing reliability. Appendix III gives sources and methods, and a basic table of growth rates.

⁽¹⁾ See Appendix I, page 39.

⁽²⁾ See table 12, page 45.

⁽³⁾ Jul Bjerke, Some Aspects of Long-term Economic Growth of Norway since 1865, paper presented to the 6th European Conference of the International Association for Research in Income and Wealth, August 1959.

⁽⁴⁾ See Appendix I, page 40.

⁽¹⁾ See S. Kuznets, Levels and Variability of Rates of growth “Economic Development and Cultural Change, vol. 5, no. 1, October 1956. Kuznets found small positive correlations for various periods and groups of countries, but only one (comparing only the seven larger countries) was signific antly different from zero at a 5 per cent probability level.

⁽¹⁾ Colin Clark, Financial Times, 8 June 1960.

⁽²⁾ S. Kuznets, Levels and Variability of Rates of growth’, Economic Development and Cultural Change, vol. 5, no. 1, October 1956.

⁽¹⁾ A rank correlation is a measure of the correlation between the ranking or order (as opposed to the actual values) of two series of numbers. In this article the Spearman coefficient was used throughout.

⁽²⁾ The rank correlation is −0.27.

⁽¹⁾ As growth was again interrupted in a number of countries by the recession in 1938, the post-depression period has been taken only up to 1937 for all countries except Japan. Two alternative definitions of national income for Japan produce different movements between 1937 and 1938. The movement over the whole period from 1933 until 1938 appears, however, to be moderately well established.

⁽²⁾ If countries are ranked first according to the severity of their fall in output during the depression, with those least affected at the top, and secondly according to their growth rates over the whole inter-war period, there is a rank correlation of +0.5 between the two series.

⁽³⁾ If countries are first ranked according to the severity of their fall in output during the depression, with those least affected at the top, and secondly according to the rate at which their total output rose from the trough of the depression up to 1937, there is a strong negative correlation of −0.9. This high correlation might suggest a higher degree of catch-up in recovery from depression than in recovery from war. But this is probably due almost entirely to the fact that a four to five year recovery period is included within each wartime period (since the wartime periods are taken as 1913-1922 and 1938-1950). Growth rates after the depression, however, were measured from the bottom of the trough.

⁽⁴⁾ See Simon Kuznets, Six Lectures on Economic Growth, (Illinois, 1959), page 55, where it is established that income differentials are narrower in developed than in underdeveloped countries.

⁽⁵⁾ See Appendix I, page 39.

⁽¹⁾ It can be argued that this sort of difficulty, though it may lead to an overestimate of the standard of living, does not lead to an overestimate of productivity : the workers in transport and distribution are certainly producing more services. But it is also arguable that their services should be considered not as final products, but as intermediate products—analogous to the higher internal costs of transport in a large factory compared to the costs in a small one. There is a similar difficulty in the convention of treating all Government expenditure on goods and services as final output. In fact, part of the large increase in Government expenditure— such items as the increase in the size of the Board of Trade and the Ministry of Food and Agriculture—should properly be regarded as intermediate production, because its main purpose is to facilitate the output of other goods and services. But this particular difficulty is probably present as much in the twentieth century as in the nineteenth.

⁽¹⁾ For a number of countries our series start very near the time identified by Rostow as take-off’. See W. W. Rostow, Stages of Economic Growth, Cambridge University Press, 1960.

⁽²⁾ The rank correlation between 1922-1929 and 1950-1959 is +0.7; between 1922-1929 and 1954-1959 it is +0.8.

⁽³⁾ See Appendix I, page 41.

⁽¹⁾ Except for the few years of recovery from the great depression.

⁽²⁾ Kiyoshi Kojima, Capital Accumulation and the Course of Industrialisation, with Special Reference to Japan’, Economic Journal, December 1960.

⁽³⁾ The upward bias to the figures which derives from the exclusion of the non-market product and from the under estimate of some slowly-growing sectors (page 34) probably affects the Japanese figures more than those of the other countries.

⁽⁴⁾ Since, however, we expect these high rates to be due largely to an extension of the most productive sector, it does not follow that Japanese productivity within the modern factory sector will increase more rapidly than that of other countries.

⁽⁵⁾ New Long-range Economic Plan for Japan, Economic Planning Agency, published by Japan Times, Tokyo 1961.

⁽¹⁾ National Income and Expenditure, 1960, HMSO.

⁽²⁾ D. C. Paige and G. Bombach, A Comparison of National Output and Productivity of the United Kingdom and the United States, OEEC, 1959.

⁽³⁾ J. Marczewski, Some Aspects of the Economic Growth of France 1660-1958’, Economic Development and Cultural Change, vol. IX, no. 3, April 1961.

⁽⁴⁾ The difference in table 8 between series I and II is between 1905/13 price weighting, and a chain index; on 1850 weighting it would be greater.

⁽⁵⁾ K. Bjerke, The National Product of Denmark, 1870-1952, in Income and Wealth series V (Bowes and Bowes, 1955), page 126.

⁽⁶⁾ Productivity, Prices and Incomes, Materials prepared for the Joint Economic Committee, (Washington, 1957), page 19.

⁽⁷⁾ M. W. Towne and W. D. Rasmussen, Farm Gross Product and Gross Investment in the Nineteenth Century’, Trends in the American Economy in the Nineteenth Century, Studies in Income and Wealth, vol. 24, NBER 1960.

⁽¹⁾ K. Bjerke, page 125 of source quoted in footnote (5), page 39.

⁽²⁾ O. J. Firestone, Canada's Economic Development, 1867- 1953, Income and Wealth series VII, Bowes & Bowes, 1958, page 223.

⁽³⁾ Industrial Statistics, 1900-1959, OEEC, Paris 1960.

⁽¹⁾ As this weight is affected by the price base used, any attempt to separate these factors in series constructed on different bases and by different methods would have a very high margin of error.

⁽²⁾ H. Barger, Distribution's Place in the American Economy since 1899, NBER, 1955, which shows that over the period 1899-1949 output per person engaged in commodity production in the United States increased nearly fivefold, while output per person engaged in distribution barely doubled.

⁽¹⁾ In algebraic terms this is demonstrated by the fact that a current weighted (Paasche) price index must be applied to obtain a base weighted (Laspeyres) quantity index, viz :

⁽²⁾ M. Gilbert and Associates, Comparative National Products and Price Levels, OEEC, 1958.

⁽³⁾ If one assumed that a change of this order occurred over, say, 50 years instead of 80 years, the spread between the two differently weighted annual rates would be slightly greater.

⁽⁴⁾ Sveriges Nationalprodukt, 1861-1951, Konjunkturinst itutet, (Stockholm, 1956), page 43.

⁽¹⁾ This is presented as 1954 in the OEEC series, but in some cases conversion from a different national base year may have been made rather crudely.

⁽¹⁾ Statistisches Jahrbuch für die Bundesrepublik Deutschland 1960, Statistisches Bundesamt, (Wiesbaden, 1960), page 486.

⁽¹⁾ A. Maddison, Economic Growth in Western Europe 1870-1957’, Banca Nazionale del Lavoro Quarterly Review, March 1959.

⁽²⁾ 15-59 for Japan.

⁽³⁾ In those periods during which hours were sharply reduced, it is also doubtful how much of the productivity rise per man-hour which occurred was due simply to increases in personal efficiency directly resulting from the reduction of hours.

⁽⁴⁾ The series for France and Italy were adjusted for unem ployment from 1950-59, but the effect of the adjustment in this period is small.

⁽⁵⁾ Walter Galenson and Arther Zellner, International Comparison of Unemployment Rates' in The Measurement and Behaviour of Unemployment, NBER, Princeton, 1957.

⁽¹⁾ The base year of the original series is not generally given, because this frequently does not constitute a true base, but merely the year in whose average values the figures are presented.