^{(note 2 in page 13)} A. D. Smith, D. M. W. N. Hitchens, S. W. Davies, International industrial productivity: a comparison of Britain, America and Germany, Occasional Paper XXXIV, CUP forth coming autumn 1982.

^{(note 1 in page 16)} It is instructive to compare the above alternative US/UK productivity ratios with those measured by Paige and Bombach, (A comparison of national output and productivity of the United Kingdom and the United States, Paris, OEEC 1959, p. 21) for 1950.

The excessive transport differential is probably explained by the effect of the very different relative prices of freight and passenger traffic on the two railway systems.

^{(note 1 in page 17)} In the case of both the US/UK and Germany/UK comparisons, food industries figure prominently among the perverse results. This may well reflect the distortion which in these activities taxes and subsidies exert on the relevant product prices. Whilst alternative weighting systems disturb the overall pattern of results only to a minor extent, they may have a disproportionately large impact in individual cases.

^{(note 2 in page 17)} The impact of this adjustment raised the manufacturing ratio by as much as 27 per cent in Rostas' study of the late 1930s (Comparative Productivity in British and American Industry, CUP 1948, p. 29).

^{(note 1 in page 18)} A similar lack of pattern was evident in the findings of Paige and Bombach, A Comparison of National Output, who pursued the adjustment more assiduously than the present authors. Whilst the Paige-Bombach comparisons are not wholly on par with those given in table 5 they suggest that British agricultural productivity shows up worse with double deflation, extractive industries and the utilities rather better, and the impact of transport depends on the prices used.

^{(note 1 in page 20)} A comparison was made between the sectoral ratios obtained by Paige and Bombach (A Comparison of National Output) for 1950 and estimates for the same year based on our 1968 results taken back by appropriate output and labour force indices. Whilst the revealed discrepancies, at least in terms of rankings of sectoral ratios, were not too alarming (transport apart) the exercise confirmed that this approach is unsatisfactory for gauging long-term changes in productivity ratios. In particular the 1950 ratio for the extractive indus tries—where structural changes have been exceptionally important—emerges as 3.98 when ‘taken back’ from 1968, compared with the Paige and Bombach ratio of 7.69.

^{(note 2 in page 20)} Op.cit.

^{(note 1 in page 21)} A Comparison of National Output, p. 214.

^{(note 2 in page 21)} The fixed elements in transportation activities, loading, unloading etc.

^{(note 3 in page 21)} Comparative Productivity, pp. 42 and 79.

^{(note 1 in page 22)} It is more appropriate to apply this structural test than one based on the assumption that US output is distributed on the British pattern since conceivably we could improve our overall productivity performance by changing the structure of British manufacturing industry. The two tests will not in general yield the same result—though the differences should not be great—because industry patterns of output per man are not identical in two countries.

^{(note 1 in page 23)} A measure of the confusion in this picture is a correlation coefficient between the two sets of ratios in 1976 (US/UK and Germany/UK) of only +0.27. The coefficient between the two series in 1968 is +0.07.

^{(note 2 in page 23)} The correlation coefficient between the two series for 1976 is +0.60 and for 1968 is +0.72.

^{(note 3 in page 23)} Though our reliability assessment suggests that measure ment errors for this industry are perhaps the largest.

^{(note 4 in page 23)} International industrial productivity, op. cit.

^{(note 1 in page 24)} Although the measurement of international comparative productivity levels is a field of economic research less deve loped than many others, over the years there has been an accumulation of considerable experience in measurement methodology. These methods have recently been the subject of an admirable survey by Kravis (‘A survey of international comparisons of productivity’), perhaps the most striking feature of which is the variety of methodologies that have been attempted at various times by different people. The methods vary, inter alia, in their bases (physical and value) in the levels to which they pertain (national, sectoral, industry, firm, plant) and in whether or not they are concerned with single, or mul tiple, factor inputs. The basic methodology used in this study is well established, though it required further development in certain areas to deal with specific problems which arose.

^{(note 2 in page 24)} Labour productivity measures based essentially on physical units have also been adopted for instance in the case of oil and gas extraction, water supply, road haulage, ports, ship ping activities, airlines and airports.