Click for next page ( 182


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 181
a International Comparisons of Productivity Values placed on a nation's total output, its output per worker, or its output per hour worked mean little in themselves; they are significant in comparison with other times or places. Similarly, growth rates of output or productivity in a country during any stated period are of interest when they can be compared with growth rates in other periods or places. In the United States, productivity measurement and analysis has centered on comparisons over time within this country, and the results have been valuable. In this chapter, we focus on comparisons of U.S. output and productivity with those of other countries, of the U.S. growth rate with rates in other countries, and of growth rates at various times in a number of countries.) COMPARISONS OF LEVELS OF OUTPUT AND PRODUCTIVITY International comparisons of national income have a long history,2 but Colin Clark's pathbreaking The Conditions of Economic Progress (1960), first published in 1940, is a good starting point for this discussion. Clark compared total and per-capita real national income, and national income per hours worked, in all countries for which even fragmentary information was available, in a variety of time periods. Clark knew that a valid comparison of the national income or product of two countries requires price comparisons; it cannot be obtained by converting values in the two countries, expressed in their own national

OCR for page 181
182 REPORT OF THE PANEL currencies, by use of exchange rates because the exchange rate usually is a poor indicator of the ratio of the average price of all goods and services in one country to the average price in another. Clark also understood that the use of two countries' price weights would yield two different output comparisons and recommended Fisher's ideal index number (or, as Clark called it, the Fisher-Pigou index) to average them (Clark 1960, pp. 1~17~: To compare for instance the real value of $0.819 produced per hour worked in U.S.A. in 1929, and 1.34 Rm., or $0.320 at par of exchange, produced per hour worked in Germany in the same year, we must take account of the actual quantities of goods and services produced, or, in other words, what the money will buy. The average American over that period spent his income in a certain way, purchasing certain quantities of goods and services. If he had gone to Germany and had set out to purchase exactly the same goods and services, he would have found that they were 0.9 per cent cheaper in the aggregate than in his own country. The German with his income purchased certain goods and services, by no means in the same proportion as the American. He spent relatively less of his income on motor cars and rent, and relatively more on food. The German going to America and purchasing the goods and services which he was accustomed to consume would find that they were 19.8 per cent dearer. In comparing the real value of incomes in the two countries we must therefore allow something between 19.8 and 0.9 per cent for the difference in purchasing power of money. The ideal formula gives us the result that the comparative real income per head can be obtained from the geometrical mean of these two ratios, or Average real income in America Average real income in Germany = ~ / Cc,st of American ponds at (-,erman prices V (cost of German goods at American prices Average money income in America Average money income in Germany = 0819 x1/0991 0.320 V 1.198 The ratio is seen to be 2.33 as against the 2.56 which we obtained from a crude comparison of money incomes. By the application of this and other index numbers we can make comparisons of economic welfare of different times, places and groups of people without in any case having to use any more elaborate formulae than the one given above. In some cases even simpler comparisons will, in the present state of knowledge, be all that are worth attempting. Clark obtained estimates for many places and dates. He devised the "international unit" to bring them into a common unit of measurement, supplemented by the "oriental unit" for low-income countries. However, the information available to him was meager and often less appropriate than if it had been collected specifically for international comparisons. Studies by Gilbert and Kravis (1954) and Gilbert et al. (1958) for the Organization for European Economic Cooperation (OEEC) were the first really satisfactory international comparisons. They compared the United .^

OCR for page 181
International Comparisons of Productivity 183 States, Belgium, Denmark, France, West Germany, Italy, the Netherlands, Norway, and the United Kingdom in 1950, and estimates based on shortcut methods were provided for 1955. The OEEC group first adjusted available estimates of gross national product to improve comparability among the countries. They then divided the GNP estimates expressed in national currencies among about 250 uniform expenditure categories. For each category they established the ratio of the quantity in each European country to that in the United States and the ratio of the price in each European country to the price in the United States. (Knowledge of either the price ratio or the quantity ratio permitted the other to be computed.) A good deal of price information was specially collected, a difficult process because of the need to be sure that product specifications were the same or equivalent in the countries compared or, if equivalent products were not sold in both countries, that a basis could be found to adjust prices to allow for differences in specifications. The quantity data for the 250 components were then aggregated in various ways. First, the components of the national product of all the countries were valued in U.S. prices at factor costs, in dollars, and summed to obtain a comparison of U.S. and European national products with components weighted by U.S. relative prices. Second, the process was reversed and the components of U.S. national product were valued in the prices of each European country. This provided a comparison of the national product of each European country with that of the United States with components weighted by relative prices in the European country (termed "national European price weights"~. The ratio of the national product of each of the European countries to the national product of the United States was found to be much higher when the comparison was based on relative prices than when it was based on European relative prices. Like Clark, the OEEC group also computed Fisher ideal indexes. In addition, they computed price and quantity ratios based on use of the average weights for the European countries. They also repeated all the calculations with market-price (as distinguished from factor-cost) weights. The data, for the first time, provided acceptable comparisons of total and per-capita GNP and of GNP per person employed in the major and several smaller industrial countries. Comparisons of national income were also made by deduction of depreciation estimates from GNP at factor cost. It was almost two decades before funds could be assembled for a new study.3 In the meantime, various researchers and agencies, including OEEC itself, were forced to adopt the expedient of extrapolating the 1950 data by national income or product series for each country, valued in its own prices of some base year, or a variant of this method. But there was no assurance that the results were accurate, and uncertainties mounted as the

OCR for page 181
184 REPORT OF THE PANEL extrapolation period lengthened; neither could the detailed information necessary for analysis be obtained in this way. The next major study, the International Comparisons Project Acid, was a joint project of the Statistical Office of the United Nations, the World Bank, and the International Comparison Unit of the University of Pennsylvania (with additional financial support at various times from the Ford Foundation and the foreign aid agencies of several countries as well as cooperation of the statistical agencies of the countries included in the comparison). Kravis was again one of the principal investigators. Phase one of the project provided comparisons for 1967 among the United States, the United Kingdom, Japan, Hungary, India, and Kenya and comparisons for 1970 among the same six countries plus France, West Germany, Italy, and Colombia (Kravis et al. 19754. The authors state (p. 3~: the general framework of the binary comparisons follows the OEEC studies of the early 1950's, but several major improvements over these studies have been made. First, the more extensive cooperation of the countries has enabled ice to make a larger number of price comparisons with far more control over the comparability of quality than in the earlier studies. Second, for rents and some durable goods, so called hedonic regression methods of international price comparisons have been employed.... These methods, made feasible by the advances in economic theory and statistics and the advent of computers since the OEEC studies, allow us to hold constant across currencies a number of different quality variables and thus to improve price comparability. Third, for construction, reliance has been placed almost entirely upon puce comparisons for entire construction projects rather than for measured units of building operations (such as the laying of a certain number of bncks), upon which major reliance was placed in the earlier work. Although the experience gained in the earlier studies provided valuable guidance for the present work, new problems had to be solved. For one thing, the very heterogeneity of countries created some difficult problems. For example, a common list of items for which poce comparisons could be made for all countries had to be ruled out. In addition, ways had to be found to meet the problems posed by the existence in some countries of a large rural sector having a substantially different content of consumption from the more Westernized urban sector. Interest in productivity comparison in the United States centers on industrialized nations, for which the greatest accuracy is needed. The "new problems" concern comparisons with the developing countries. Of course, many old problems in comparisons remain even among industrial- ized countries, but the Ice study, like the OEEC study, is a major achievement. The Icy, like OEEC, provided bilateral comparisons between the United States and each of the other countries, based on U.S. prices, prices in the other country, and the ideal index. In addition, the investigators developed

OCR for page 181
International Comparisons of Productivity ~5 an "international" or "multilateral" system, which is based on linking comparisons between countries at adjacent levels of development. The Ice report provides comparisons of total and per-capita gross domestic product (GDP) at market prices (and their components). Table 9- 1 shows these comparisons for the industrialized countries, as well as of GDP per person employed, as calculated by Denison, and of per-capita consumption. The importance of the choice of price weights is apparent from this table. In addition, the substantial differences between compari- sons of output per person employed and output per capita, which result solely from intercountry differences in ratios of employment to population, make it clear that per-capita data are inappropriate for productivity comparisons. Phase two of the project (reported in Kravis et al. 1978b) added two more industrialized countries, Belgium and the Netherlands (see Table 9- 1) as well as Iran, the Republic of Korea, Malaysia, and the Philippines. Phase three, for which the statistical work was to be finished in 1978, will add nine more developing countries (Brazil, Jamaica, Malawi, Mexico, Pakistan, Sri Lanka, Syria, Thailand, and Zambia), bringing the total number of countries in the project to 30. The most important example of bilateral comparisons that are based on careful analysis of available data have been made for the United States and the Soviet Union, but without the cooperation of the statistical agencies of both countries or special collection of price and other statistics. Such comparisons have been made in both countries (in the United States by both government and private researchers).4 The Bureau of Labor Statistics, Office of Productivity and Technology (1978), has prepared estimates for the years 1950, 1955, and 1960-1976 that cover the United States, Canada, Japan, France, Germany, Italy, and the United Kingdom. In each year, total GDP,GDP per capita, and GDP per employed civilian are expressed as percentages of the corresponding figures. (The exclusion of military employment seems to us inappropriate.) Table 9-1 compares eight countries using each of the four types of weights shown. Except for Canada, and for Japan in 1950, the data are benchmarked to the OEEC and ice data and interpolated or extrapolated by series based on countries' own price weights. Because only countries with satisfactory benchmark data are included and extrapolations forward from the latest benchmark cover no more than 6 years (except for Canada), the procedures are reasonably acceptable but regularly recurring benchmark data will be needed to maintain the series. Table 9-2 shows the BES series for GDP per employed civilian that is based on the ideal index. In addition to the careful studies based on price and quantity data, comparisons have been published in which aggregates in national

OCR for page 181
186 o A, z Cat r ~ ~ O O ~ O ~ ON ~ ~ ~ ~ 00 t~ to ~A) O _ ~ _ O Cal O to [~ ~ ~ ~ A) ._ ~ 00 \~) to to 00 ~ A) to [A A) O U~ _ ~ O ~ O O ~ ~ ~ ~ O . _ Ct ~J ~ t- t- ~ C~ c~ t~ ~ ~ c~ [~ ~ ~ C~ ~ '_ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~_ _ C ~_ ~,_ ~ ~ ~ ~ a~ ~ ~ ~ ~ ~ 0 ~ ~ ~ ~ _ _\ ~ _ O ~ ~ ~ ~ O O 00 ~ _ ~ ~ `~C' ~ ~ ~ ~ ~ ~ =) ~;} ~ =) ~ _ _ ~ ~ ~:^ ~ ~ ~ O ~ O ~ O O ~ ~ ~ O ~ ~ _ ~ ~ ~ ~ ~ ~ 00 ~ 0- ~) C~ ~- 00 ~ ~ ~ (~ ~ ~ C~ _ C: C~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~D ~0 ~ _ o =4 C~ . _ ~: ~ oo oo ~ ~ O O - t~ ~ a~ => oo c-) CJ C~5 ~3 ~O ~ ~ ~ 0 oo ~ ~ ~ 0 ~ ~ o~ ~ oo ao ~ ~ oo ~D ~ t- oo ~ t- ~ ~ ~0 ~ ~-- _ _ V, ~ ~ & ~ C~ ~ 00 0 ~ O O 00 C: _ ~S' ~:} - ~ ~ ~ O ~ ~ ~ ~ ~ ~ ~ _ U) ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ C~ C~ O ~ ._ C) ~O O O O O O O O O O O O O O O O Ct ._ - O O O O O O O O O O O O O O O O != ~ ~O O O O O O O O O O O O O O O O ~cq ________________ _ C~ C~ ~C~ o _ _ C) (Q _ ) .') '__ ~ C~ 1 G,) _ U ~C) O C~ - U~ CN ~ 00 ~ ~ _ ~ ~ ~ ~ O t- 00 v0 C~ 00 C~ ~ ~ ~D ~ ~ u~ o o . . o o o o ' ~ x 3 O O C~ ._ . ~ - V) _ O <3 C~ ~ O O =4 C~ ~ ~ ._ C _ C) <) (d ~ Ct - X C ~ X ~ _ C) O _ C) O 0 ~ _ 0 5 - _ O _ ~ ~ o ~ _ - - O =,,~ _ .-~ =^a _ ~r~ _ ~ _ I~ ~_ C~ C) O ~O ~_ o ~ O r~ C~ O ~ ~ ~ O ~ Q) t~} _ _ . _ _ O ~ . ~ ~ ~ tS O ~~ . . _ ~ ~ ~ ~ t4 ) _ ~ ~ ~ ,-. O' ~ O _ _ ~ O ~ . O ~ ~ ~ _ ~ ~ ~ C~ 00 ~ ~ s0 .c ~ a~ ~ ~ 0 3: ~ V' Ct r ~ ~ O ~ ~ ~ cs ~ 3 (;_) .' .C <~) ~ ~ C 't S ~ C~ r-1 r . _ C O ~ ~ , Z ~ ~ O O O V ~ = Q) ~t r~ ~) ~ ~ ~ o. c~ t- ;) 0 't cr, ~ ~ ~ o ~ 3 ~ J, rc, E: _ ~ ' - . ~ o~ 0 O C, ~ CJ' - C.) ,- ~ s~ ~ ~0 _ _ CC ,~ ~ O O ~ ~ ^ ~ D. t :e ~ ~ _ _ CP, ~ ~ ~ O _ _ O - E~ ~ l- - Q) Ce ~ ~ o P~ O O l~ ~ ~ . _ ~ ~ C_) ~ '~ _ -~,, ,, 5 o ~ 2 ~ _

OCR for page 181
International Comparisons of Productivity ~7 TABLE 9-2 Real GOP per Employed Civilian, 7 Countries, 1950-1976a Output Based on Geometric Mean of U.S. and Own Price Weights (U.S.= 100) United United Year States Canada Japan France Germany b Italy Kingdom . _ . 100.0 81.8 16.2 42.8 36.6 28.5 52.3 100.0 84.2 19.1 45.0 43.3 31.4 50.2 100.0 86.6 24.5 54.3 5 1.6 36.0 53.0 100.0 85.8 27.0 56.4 52.4 38.0 51.6 100.0 85.6 27.4 57.3 52.3 38.8 49.6 100.0 86.2 29.3 58.0 52.4 40.7 50.3 100.0 86.0 31.8 58.8 54.4 40.9 50.9 100.0 85.6 31.9 59.1 55.4 41.9 49.9 100.0 83.5 33.2 60.3 57.2 43.7 49.1 100.0 83.4 36.4 61.8 56.4 45.9 50.7 100.0 84.6 39.7 62.7 58.6 48.0 51.5 100.0 86.2 43.5 66.3 62.2 5 1.2 52.2 100.0 88.6 48.3 70.9 66.0 54.2 54.5 100.0 90.3 50.4 72.4 66.5 54.1 55.8 100.0 90.2 53.4 74.0 67.5 55.4 55.5 100.0 90.0 56.0 75.5 69.2 57.6 56.7 100.0 91.8 57.4 79.2 73.2 60.4 58.0 100.0 90.3 58.8 79.0 73.7 57.8 57.4 100.0 90.3 60.3 80.7 76.6 59.1 57.6 aData for the latest year are based on preliminary estimates. bExcluding the Saar and West Berlin in 1950 and 1955. SOURCE: Bureau of Labor Statistics, Office of Productivity and Technology, "Com- carative Real Gross Domestic Products Real GDP ner Canita and Real GDP ner Em . . . played Civilian, Seven Countries, 1950-1976." Prepared June 1978. currencies are converted to a common basis by use of exchange rates. As noted above, such comparisons have never been acceptable. In the past, they tended to exaggerate international differences in output per capita or per worker, but the amount of the exaggeration was irregular. With the large and often erratic swings in exchange rates that have become common in recent years, use of exchange rates frequently yields year-to-year changes in the relative positions of countries that are inconceivable, and even the rank order of countries is often implausible. This appears clearly in Table 9-3, which shows per-capita output comparison based on both purchasing power parities and exchange rates, as compiled by Yang (1978~. The exchange rate comparisons indicate that in 1976 per-capita output was higher in Switzerland, Sweden, and Canada than in the United States, and almost as high in Denmark and Norway; the purchasing power parity comparison shows quite different results. The need for information covering all or most countries on a current

OCR for page 181
188 REPORT OF THE PANEL TABLE 9-3 International Compansons of Per-Capita Output, 1970, 1974-1976 (U.S. = 100) Based on Purchasing Power Parities Based on Exchange Rates Country 1970 1974 1975 1976 1970 1974 1975 1976 United States 100 100 100 100 100 100 100 100 Canada 83 91 93 91 80 100 100 105 France 75 81 NA NA 58 76 89 83 Belgium 72 79 79 NA 55 83 90 87 Sweden 78 78 78 77 86 103 118 104 West Germany 75 76 76 76 64 92 95 92 Netherlands 72 76 75 75 51 79 84 82 Denmark 72 72 72 72 66 90 98 96 Norway 63 67 71 71 60 87 99 98 Japan 62 68 71 70 39 62 62 62 United Kingdom 60 61 61 59 46 51 57 49 NA = not available. SOURCE: Yang, Jai-Horn (1978) Comparing per capita output internationally: Has the United States been overtaken? Review-Federal Reserve Bank of St. Louis, May, pp. 8-15, Table V. basis has also led to sets of estimates that are based on good use of all the available information, but are handicapped by the gross inadequacy of such information. The estimates of total and per-capita "planetary product" prepared annually by the Bureau of Intelligence and Research, U.S. Department of State, are in this category.5 Kravis et al. (1978a) have recently prepared a comprehensive set of estimates. The OEEC and Ice estimates, as noted above, were obtained by dividing GNP or GDP among detailed categories of final products, measuring the quantities of the products in the two countries in each category, and summing categories. This is the best way to compare the total production of two countries, and it also provides much interesting detail, including detail needed for analysis of productivity differences. It does not, however, yield comparisons by industry. A comparison of output per person employed in the United States and the United Kingdom in 1950, by industry, was made by Paige and Bombach (1959~; it is the only such comparison involving the United States that covers all industries. Earlier, Rostas (1948) and Frankel (1957) had compared productivity in the same countries in manufacturing, and Rostas had compared a few nonmanufacturing industries. West (1971) made rather similar comparisons between the United States and Canada, for 30 manufacturing industries and for all manufacturing. All these investigators tried to compare capital as well as output and employment. Yukizawa (1975, 1977) compared output per person employed in Japan and the United States in about 60 manufacturing industries. His studies are of special interest because they span a period in which the gap between

OCR for page 181
International Co. IS US Productivity ~9 Japanese and U.S. output per worker was narrowing with unusual rapidity. He finds that labor productivity was 2.7-3.2 times greater in American than in Japanese manufacturing in 1958-1959, 2.5-2.9 times greater in 1963, 1.7-2.1 times greater in 1967, and 1.3-1.6 times greater in 1972. These ranges do not refer to the variation among industries, which is far larger, but to the difference beween the results of using Japanese and American weights to combine industries. As always, low productivity is associated with low volume and high relative prices so that the comparison most favorable to each country is that in which the other countries' prices are used. For the steel industry, the Bureau of Labor Statistics provides annual comparisons of output per hour (as well as hourly and unit labor cost) in the United States, Japan, France, Germany, and the United Kingdom (see Chapter 2, Table 2-1~. The estimates are indexes on a 1964 base and cover the period 196~1976. INTERNATIONAL COMPARISONS OF CHANGES OVER TIME International comparisons of changes over time in output and productivity depend mainly on the quantity and quality of research within each country using its own data. Since the growth of the value of output and inputs in each country can be measured in its own currency, the comparison of growth rates of productivity does not require the use of purchasing power parities as does the comparison of levels of productivity. However, any study providing international comparisons must work to maximize consistency and to obtain an understanding of the strengths, weaknesses, and other characteristics of the series being compared. One type of comparative study has dealt with long time periods, starting from 1695 to the 1930s: Kuznets (1966, 1971) and Maddison (1964, 1970) are particularly identified with this type of investigation. These authors have assembled and evaluated a great deal of data and filled many gaps. Another type of study has confined comparisons largely to the years since World War II. Such studies benefit from the development of national income measurement in many countries. The interest of various interna- tional agencies in such data was also helpful; it led early in the postwar period to a standardized system of national accounts, a follow-up to an agreement on treatment of debatable items that had been reached earlier among national income accountants of the United States, the United Kingdom, and Canada. The United Nations, the Organization for Economic Cooperation and Development, and the European Economic Community regularly publish time series for their member countries. These are very helpful, although

OCR for page 181
190 REPORT OF THE PANEL they do not eliminate all difficulties of comparability. Neither have these organizations solved the problem of providing continuous series for periods of more than about 10 years, even when continuous series exist within the individual countries (which, quite desirably, revise their data from time to time).6 Major changes in the standardized system of national accounts a few years ago have made it difficult to extract continuous series for even such basic aggregates as national income and gross national product. BES compiles annual indexes of real GDP total, per capita, and per civilian employed in each country's own prices for the years and countries covered in Table 9-2 (see, for example, Boner and Neef 1977~. Series for various measures and a variety of countries are also published regularly or irregularly by other agencies including the Census Bureau, the Bureau of Economic Analysis, and the Federal Reserve Bank of St. Louis. BES provides an annual comparison of series for output per hour in manufacturing in the United States, Canada, Japan, Belgium, Denmark, France, Germany, Italy, the Netherlands, Sweden, Switzerland, and the United Kingdom (for a description, see Boner and Neef 1977~. For most countries, it covers all years from 1950 to 1977. Table 9-4 reproduces these data. BES also provides labor compensation (or cost) per hour, unit labor costs, total output, total hours, total compensation, employment, average hours, and real compensation. Manufacturing output is based on national accounts except for Japan and Switzerland. ANALYSES OF THE DATA International comparisons of levels and changes in output and productivi- ty are of interest in themselves, but their greatest value is as starting points for investigations of causes or consequences of international differences and of structural features of societies and economies that are typically associated with differences in productivity or per-capita income. Clark, Kuznets, and others cited in this chapter have not only compiled output and productivity data, they have analyzed those data. There have been a great many such analytic studies, some comprehensive, others concerned with a single aspect (such as tax systems, investment, resource allocation, or scale) of the causes of productivity differences or of the consequences of per-capita income differentials (for example, on consumption patterns or on population increase). Even studies with objectives unrelated to income or productivity often find it necessary to use information about these variables, if only to eliminate their influence. These include international comparisons of energy consumption to learn whether the United States is particularly wasteful in using energy, and if so, what might be done to

OCR for page 181
International Comparisons of Productivity 191 conserve it (see Darmstadter et al. 1977~. Such studies, it may be added, also require information on relative prices that can only be obtained by studies of the type done by the International Comparisons Project. Space does not permit a comprehensive examination of the literature, but this chapter briefly reviews three groups of studies that deal with the causes or sources of growth. The earliest was Denison (1967~. Denison developed estimates of the sources of growth of real national income for 195~1962, and for the 195~1955 and 1955-1962 subper~ods, for the United States, Belgium, Denmark, France, Germany, the Netherlands, Norway, the United Kingdom, and Italy. His classification of sources of growth, described in Chapter 7, and methodology were used for all countries. Denison also estimated the sources of the difference in the level of national income per person employed in 1960 between the United States and each of the other countries, based on United States price weights. Walters soon thereafter developed comparable estimates for Canada, with the sources of growth estimates eventually covering 195() 1967, with 195~1955, 1955-1962, and 1962-1967 subpenods, and a 1960 comparison of levels of income per person employed. Denison and Chung (1976) later developed estimates for Japan that cover 1953-1971, with 1953-1961 and 1961-1971 subper~ods, and a comparison of levels for 1970. Denison also revised and extended the U.S. estimates to cover 1929-1969 (Denison 19741.7 Table 9-5 presents estimates obtained from the studies of the sources of growth in 11 countries (Denison and Chung 1976~. Three points about this table should be noted. First, growth rates were adjusted to a standardized basis in order to improve statistical comparability of the output measures, to eliminate effects of irregular factors on output per unit of input, and to screen out some of the effects during the early 1950s of the recovery from wartime distortions. These adjustments were made to eliminate sources that would complicate or impair the validity of compansons. Second, the time periods are not uniform. Third, Table 9-5 includes a line, "gains from economies of scale associated with income elasticities," that was not used in the comparable Table 7-1, which is confined to the United States. According to Denison (1967~: Where per capita consumption abroad has risen markedly toward the U.S. level, the increase in consumption has been concentrated in products such as consumer durables for which demand is income elastic and relative prices abroad have been high (in comparison with United States prices) because relative quantities used have been small (in comparison with United States quantities). Consequently, the greater is the rise in per capita consumption in a country and the lower its initial level, the larger tends to be the amount by which the rise in consumption in each country measured in its own constant prices exceeds the rise measured in U.S.

OCR for page 181
195 \0 ~ ~D _oo o ... . . . ooo o o o V~ ~ ~ C~ _ ~ d. - ooo o o o o _ ~o oo _ ... . . . . ooo o o o o C~ o o ~_ ~ ~.C' ~ __o o _ o o _ ~o _ o _ o ~_ _ ~ ~ ~ ... . . . . _oo o _ o o U) ~ o _ ~D ~O O u~ ... . . . . ooo o _ o o oo_oo o~ <)e~_ o ooo o _o .. oo o V~_ ooo . . . o o o o ~ ~ ~ o ~ ~ o . . . . o o o o ~o ~ ~ oO ~t ~ 1 .. . . oo o o ~o_ ~ =, oo cr ~O ~ O ~ .... . . . oooo _ _ o o ~ Ct _ - o ,~ _ . Ct o o c ~ .~ C~ ~ , ~ C o C - c~ o $~ - - ~ D CC Ct > o - o o ~ o D . V~ C) ~ Ct ~ ~ - o - o . - ~: C) _ C~ c o ~ Ct C~ o c V) C) ~, o o C) , o C~ ~> C~ . - V) . C) cq _ . . ~ .o - .= ) ~V o C ^. _ D C) - ._ ~Ck _ ~ ~X ~C ._ O _ ._ ._ _ D ~ Ct ~C~ ~- O C~ bC ~._ O ~: O o .C : _ 3 a : C 0 O _ C C O ._ ._ ~c _ V3 :3 Ct ._ - co C) S c ._ u - c ._ C~ ._ o - C~ o C~ 3 C) o s. o e:: ~, ~o - - _~ c C~ 5 o C - 3 04 5 C . _ ~ C _ _ ~_ ,0 0~ 4_ C o 3 ~ -~ ~ DC C.> ' D Co C~ C). _ ~ - C,} C O - CL ~ ~ '= _ ~i _ o C~ ~ ~ _ C ~ ~ S_ ~ _ C 00 E,. 3 ~ _ , _ ~ e5 ~ ~ . . _ ~ c o ~ a ~ z ~ ~ ~ C~ ~ c o ._ c . - C O r ~ ._ _ ~ .' 'c 0 C,, U) CL

OCR for page 181
196 REPORT OF THE PANEL constant pnces. The "income elasticities" entry represents the difference that this systematic pattern introduces between the growth rate of N! when the components of consumption are weighted by U.S. prices and the rate when they are weighted by national pnces. This difference reflects concentration of increases in consumption in products where potential gains from economies of scale are particularly large. To estimate this growth component, it is essential to have comparisons of the levels of per-capita consumption in the United States and the other country based on price weights in both the United States and the other country. Table 9-6 summarizes the level estimates. The top line shows the percentage by which national income per worker in each of the other countries fell below that of the United States. The rest of the table divides this percentage among sources of growth, showing the number of percentage points ascribed to each source. Denison points out that Table 9-6 analyzes only differences in output measured in U.S. prices, which minimizes the gap between output per worker in the United States and abroad. It should also be noted that in the United States, 1960 and 1970 were both recession years in which productivity was unusually low, and this adversely affected the position of output per unit of input in the United States relative to the other countries (except Canada, which was even more severely affected by the 1960 recession). The amount is estimated in Table 9-6 in the line, "irregularity in pressure of demand." Denison has stressed the difficulties of"level" comparisons, and states that in almost every respect they are less satisfactory than the comparisons of the sources of growth over time. Comparisons of the sources of growth in a number of countries, computed on a consistent basis, permit one to distinguish the unusual features of any one country's growth experience from those that are typical of most countries. To evaluate such findings, however, and especially to draw conclusions from them, it is helpful and often essential to couple them with information on the relative positions of the countries. There is, in fact, a marked though by no means invariable tendency for countries in a weak position with respect to any output determinant to gain most from changes in that determinant. To draw an obvious example from the tables, countries in which overallocation of labor to farming has depressed the levels of output per person have obtained the largest contributions to growth from reductions in that overallocation. In 1959 the Committee on Economic Stability and Growth of the Social Science Research Council (SSRC) started a study of growth in seven industrialized countries: the United States, France, Germany, the United Kingdom, Italy, Sweden, and Japan. The SSRC arranged for a team of

OCR for page 181
International Comparisons of Productivity 197 investigators in each country to study growth in that country, with each team free to use its own methods, but with a list of minimum requirements. A central question in each study was, Why has the growth rate since World War II exceeded the previous long-term rate? Develop- ment of long-term series for output, employment, capital stock, and other magnitudes was an important feature of the studies. They were to be capped by a summary volume that was to draw generalizations from the individual studies. The reports on France and Japan have been published (Carre et al. 1975, Ohkawa and Rosovsky 1973~. They use various approaches to analysis, including Denison's techniques and, often, his estimates. Articles present- ing partial results for other countries have also appeared.8 These countries' studies have augmented statistical information about growth, especially about growth in the nineteenth and early twentieth centuries in a number of countries, including the United States. Jorgenson and Christensen, with earlier participation by Griliches, have developed a set of definitions and procedures for the measurement of output in the private domestic economy and its division between factor inputs and output per unit of input. They and their collaborators have prepared estimates for the United States, Canada, France, Germany, Italy, South Korea, Japan, the Netherlands, and the United Kingdom. Most series start with 1951 (although the starting range is from 1948 to 1961) and end with 1973.9 The Christensen, Cummings, and Jorgenson (1980) study (hereafter referred to as cc~) develops a production account within a complete accounting system and focuses on real output, real factor input, and multi- factor productivity. In comparing the results of these methods with those of Denison, it should be noted that the growth of real factor input accounts for a much larger proportion of the growth in real output in the cc estimates than in Denison's. The difference arises in large part from differences in the measurement of capital input (see above, and Chapter 7~. The ccs analysis stresses the importance of the growth of capital stock resulting from saving and capital formation and the associated growth in capital services as a component of factor input. The cc estimates are based on an internally consistent treatment of capital in four sets of accounts production, income, savings, and wealth ins current and constant prices and thus incorporate the determinants of capital forma- tion. The cc analysis uses index number formulas to represent particular production functions, and cc selected the Tornqvist index as an exact representation of the homogeneous translog production function. It is a log-change index using prices to weight input quantities. The Tornqvist index uses prices from both the base period and the comparison period,

OCR for page 181
198 Ct Ct :^ - 5 o ~ 5 _ o~ - c ~ ._ :^ 3 o z _ ~ 7 ce ~: _ ~ C~ $_ 3 v C~ Ct L~ ~ == o ~ ~ a' _ ~ o o .= Ct o o :: ~ ~: ._ ~ C~ C: ~C) L~ ~o V, V) o oo ~ ~ ~ ~ o ~ ~ ~ ~ ~ o o ~ ~ ~ ~ o C ~o o o ~ _ o _ 1 1 1 1 1 _ 1 o ~ ~ o~ oo ~ oo ~ ~ C~ . . . . . . . . . . . . . . . o ~ ~ ~ o oo ~ ~ o oo _ o _ ~0 ~ 1 ~ _ O O ~D - ~ O ~ ~ O _ _ o ~ o C~ ~ ~ o _, , ~ oo ~ o ~ . . . . .. o o o _ Co, o ~ ~ ~ _ C~ ~ _ o V~ ~ ~ ~ _ . . . . . . . . . . . . . . _ V) o C~ o ~ V) ~ _ _ o o 1 1 O oo ~ ~ cr~ ~ 0 ~ ~ ~ u~ c 4 . . . . . . . . . . . . . . et ~ o C~ ~ _ o ~ o o ~ o ~ 1 1 1 ~ 1 o o ~ ~ ~ - o ~ ~ o ~ o ~ ~t- ~ - ) ~ ~ - - o ~ - o ~ ~o ~4 0 0 0 _ _ 0 ~ _ ~ - crx ~ 0 oo ~t - ~t _ 1 o o oo ~ c~ - ~ oo u ~oo ~ ~ o - ~i _ ~i ~ c~i ~ ~ _ 0 ~ 0 0 - ~ - l ~ u~ - . ~ ~ c) s ~ - - o , ~ o ~ ~ Z 3 c~ Ct _) o ~ _ OS ~ oo ~ =~ ~, ~ oo - - - ~ - ~ cr~ _ 0 ~ 0 0 c~ _ 1 C~ ~o o ~ ~ ~ ~ ~ ~ ~ ~ ~ o ~ .. . . . . . . .. . . . . . o _ ~ ~ ~ oo ~ o~ _ _ _ 1 ~) V) q) , O _ 3 _ _ ,' C) ~ C) c:.= O ~='_,= o ~3 c, c ~ o ~ ~ , ,, _ te ,~ o O ~ = .~ 3 _ o :> ~ ~ ~ c c =, o o Ct ~ ~ =i ~ ~ ~ Z E~ ~ ~ C) ' ~ 3 _ ._ _ _ - ~ C ~ ~ ~ ~L) C '- C ~ - C,4 0 :~. Ct ~ O o O _ ~ O =, _ _ Ct D o _ _ ~ ~ _ o

OCR for page 181
199 ~ C ~ Z ~ ~_ z o _ o o - o ~o o ~ Z ~o _ o 0 cr . ~ o ~ Z oo O cr~ o ~ z , o c~ crx oo v~ o ~ Z I i ~ _ ~o oo o ~ ~_ o C~ Z i _ _ ~o _ ~d o ~ Z _ o ~ _ ~ o V) Z _' oo o o _ ~oo o o U~ ~_ o ~ Z 1 c-l - cr ~O o ~ Z I o C~ _ ~o o O c~ Ol q O ~ ' ,, ~ ~ ~ i = ~ ' 8 ~ o ~ C) ~o4 ~ ~ 3 ~ ~, o~ c . 3 C o ._ C - ~n C ._ y o o . . a - c CL c o c) ~c - o C~ .T cn C ._ U. co .' Ct C O C ~ O O C5 ~ C Ct ~ 34 C .- C C ~ c ~ ~ - = S _ O Ce ,~ ) (, C O Ct _ _ O 2 ~ ~ C} ~V ~ - - C C Ce o C C~ e~ ~ tC -o c) ~ = c,, ~ ~ 3 Co o _ ~ ~ c '~ ~C s ~ ~ . ~ o Y ~ L. C a ~ . ~ ~ ^ 0 - C Ct C ~ ~ O C - 0 ~ 'C ~ =,.~ ~ a ~ ~ - . C O ,~ ~ '~ O a C C z 0 ~ ~ ~ ~ ~ ~ & v - c ._ S 3 r ~o o~ o cc 3 ;` o o s. 3 - cr, - - 3 c ~c c -

OCR for page 181
200 REPORT OF THE PANEL combined as in a Fisher ideal index. The cc formulation takes account of changes in the composition of capital (by type of assets) and of labor (by skill). The cc analysis permits decomposition of the rate of growth of labor productivity into four distinct sources: the rate of growth of factor productivity; the rate of change of the composition of capital; the rate of change of the composition of labor; and the rate of growth of capital stock per hour worked (the capital/labor ratio). The cc estimates are limited to the private domestic sector of each country, except that government enterprises (economic activities similar to private enterprises, not govern- ment administration) are included for countries other than the United States. Imputations for the services of consumer durables are included in output and input, which raises the share of capital input in total services. The cc estimates for 196() 1973 are shown in Tables 9-7 and 9-8. According to these estimates, the growth in multi-factor productivity in the United States in this period was substantially slower than in any of the other eight countries for which estimates were made. In their analysis, cc point to the importance of increases in the capital/labor ratio in affecting the rate of growth of labor productivity. In their view, in addition to the decline in the growth of multi-factor productivity in the United States in the years since 1960, labor productivity has been dampened by a decline in the growth rate of the capital/labor ratio. They estimate the growth in the capital/labor ratio at 3.3 percent per year for 1947-1960 and at 1.2 percent per year for 196~1973. Although all analysts consistently show a decline in U.S. productivity growth for recent years, they emphasize different causes and they rate other countries differently in international comparisons. Additional analysis and further development and refinement of basic data are required for better understanding of recent trends. In addition to studies that compare the growth of productivity in the private domestic economies of different countries, there have been a number of studies making international comparisons of productivity in particular industries. These can sometimes be done using physical rather than value measures of output. An example of such a comparison for the steel industry was given in Chapter 2, Table 2-1. CONCLUSIONS AND RECOMMENDATIONS International comparisons extend the range of information available as a basis for generalizations about productivity, for the testing of hypotheses, and for policy analysis. They are therefore an important supplement to the study of changes over time in U.S. productivity.

OCR for page 181
201 ~ , ~ Ct to 3 o o 5 o ._ ~ ._ Cal by Ct , - o it: Cal :^ Ct V C) Cal Ct Ct Cal C: ~ ~ Us ~ ~ ~ ~ ~ ~ Cal ~ ~ ~ . . . . . . . . . . . . . _ ~ Cal ~ o ~ o ~ '_ 1 1 1 U) ~ ~ ~ ~ ~ ~ ~ o _ . . . . . . . . . . . . . o ~ ~ o ~ ~ Cal ~ fit 1 1 Cal ~ . . . . ~ ~ o ~ ~ ~ ~ ~ Do ~ ~ _ ~ ~ . . . . . . . . . . Cal o ~ ~ Cal 1 Go ~ ~ ~ Go ~ cry I. . . . . . . . . . . . . _ ~ ~ ~ o ~ ~ en ~ ~ ~ ~ ~ of Go _ ~ ~t ~ ~ no ~ ~ ~ o ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ o ~ oo ~ ~ oo ~ oo oo ~ ~ ~ . . . . . . . . . . . . . . ~ c~ ~ ~ ~ ~ ~ o ~ ~ ~ ~ ~ ~ ~ o ~ ~ ~ o ~ ~ ~ ~ ~ ~ ~ ~ . . . . . . . . . . . . . . ~ ~ ~ c~ ~ ~ c~ ~ ~ ~ ~ ~ c~ ~ oo oo ~ ~ ~ ~ ~ a~ . . . . . . . . . . . . . . o o ~ ~ ~ ~ ~ - ~ ~ - ~ ~ - ~ O _ ~ ~ ~ U~ ~D ~ co a~ 0 _ ~ ~ cr ~ ~ ~ ~ o~ ~ ax ~ - ~ ~ ~- ~ -~ - ~ ~ - ~- - . . . - .9 c ~ e ~ ~s ~ ~ c, o ch ~ c) ~ s ~ ^ ~ o ~ ~s ~ 4- 3 ~ ~ o ~ Co Ct .> ~ 0, .` , ~ ~ o . o =, _~ o ~ ;> S ~ ._ o 3 ~ ~ ~ C: o C~ ~ Q 04 b4 ~ ~ ~ ~ ~ ~ C-, =5 ~ o C~ Ct ~ Ct C~ Ct ~o C) C ~ . - Ce 3 Z ~ " Ct ; C) - m C) Cn C~ ~ -~ en ._ ._ ~ . 0 . O o~ ~ e~ . ~ ~ 0 C,) 0 _ 0 ~ Ct .> s:` ~; - 0 - C) O ~ Ct .. X .= C~ o ._ ._ ~5 .= C) e~ C~ ._ ._ U, ~> - Ct o 3

OCR for page 181
202 .~ Cd ._ - Ct .~ ._ o o Ct _. o Cd Cd Cal Hi* - o Ct o Cal Cd 3 _ ~ Cd ~ so ~ ~ o Ce Cal g ~ Cd C) so _ to or oo 1 -o ~ ~ . - ~ V) ,.O ._ I_ ._ Cal Ct A o Cal ;^ Ct I_ Cal C) Cal Ct Cal ~ o ~ o ~ . . . . . oo oo ~ ~ o . . . . . o o ~ ~ ~ . . . . . ~ ~ ~ ~ o rid ~ ~ ~ 0 . . . . . ~ ~ An ~ ~ . . . . . o ~ ~ '_ oo ~ ~ ~ 'a o 0 0 r . . . . . O 0 ~ ~ . . . . . ~ ~ ~ ~ 0 oo ~ 0 .` ._ c) _ ~_ C~ _ _ ~_ _ C~ Ct - Ct Ct o r s ~ ~ ~ ; r ~ ~ 13 - ,r, o ~ Oo ^t ~o '~.b ~-~; Ct S = ot =.S t. _ ~ ~ ~ 0 0 0 ~ ~ 0 ~ ~ ^ ~V ~ a, =: ~ ~ ~0 _ Ce U~ . z =: . ~ 0 . ~m O . .~: C: =.= 0 . _ _ ~ U' 0 ~ 0 E~ ~ ce ~ os -ct I ~ o ~ ~ ) ~ CL , x o.= .. c\, ~ ~ 3 C c, ~,

OCR for page 181
International Comparisons of Productivity 203 Because reliable comparisons of national income and product levels in different countries are a prerequisite for international productivity comparisons, provision must be made for their availability on a continuing basis. The same data serve several other important purposes. The International Comparison Project has the expertise, experience, and contracts with national governments necessary to provide the needed data. The United Nations (U.N.) Statistical Commission has expressed strong support for continuation and expansion of the work. The World Bank, while intending to continue participation in the ice, proposed that the United Nations should assume the role of principal sponsor. The U.N. Statistical Office (1978) has drawn up a proposal that includes the following statement: Before the intended system of international compansons of real GDP and purchasing power of currencies can be adopted on a global basis, the existing methodology needs further improvement and extension involving empirical work in a number of directions. Accordingly it is proposed to devote one more five-year period, Phase IV, during which time the Statistical Office intends to undertake the following substantive activities: (a) refinement of existing methodologies and expansion of benchmark compare sons to a total of about 60 countnes; (b) establishment of a system for repeating the full-scale compansons for these 60 selected countnes, probably at five-year intervals; (c) development and application of methods that can be used to bring into the system countries for which benchmark compansons cannot be made, since it is unrealistic to believe that it will be possible to carry out full-scale compansons for most of the 155 countries for which data are reported in the United Nations Yearbook of National Accounts Statistics; (d) further improvement and application of extrapolation techniques developed in Phase II to produce annual compansons of relative GDPS for all countries in the system for years between the more detailed compansons; (e) further development of methods to obtain valid intertemporal (i.e., in a constant international currency unit) comparison between bench-mark years and for intervening years as well; and (f) development of techniques for obtaining compansons within regions that reflect more closely the similarities of contiguous countries as well as compansons among regions. We support these proposals of the U.N. Statistical Office, of which the second is the most important from the standpoint of the Panel's interests. Recommendation 22. The Panel recommends continued and increas ing support for the work of the International Comparisons Project, along the lines suggested by the Statistical Office of the United Nations. We recommend that international organizations provide the

OCR for page 181
204 REPORT OF THE PANEL financial support necessary to implement the proposals and urge U.S. representatives to these organizations to support the proposals. The ice data allow for comparisons of expenditure categories, such as detailed components of consumption like food and apparel, but they do not allow for comparisons of output and productivity by individual industries. Some international productivity comparisons by industry have been made by private researchers, but these are limited in the number of industries or countries they cover. The next logical step in expansion of the UP would be to collect data to make industry comparisons for industrial- ized nations, where these comparisons are most meaningful. In the United States the analysis of international productivity differences has been almost entirely the work of private researchers, although the federal government has provided valuable small-scale support, through grants from the National Science Foundation and contracts from the Bureau of Economic Analysis and Bureau of Labor Statistics that were made possible by appropriations for economic growth studies and by activities of the National Center for Productivity and Quality of Working Life.~ Direct research by federal agencies has been confined (since the time of the Marshall Plan) to rather simple compila- tions of series on output per worker or hour, mainly by the Bureau of Labor Statistics. The Bureau of Labor Statistics undertook a few years ago to update Denison's (1967) estimates under a contract from the National Center for Productivity and Quality of Working Life, but had to suspend the project because the investigators were needed for other work; otherwise, reason- ably up-to-date estimates might now be available for Western Europe. At a time when the rate of growth of measured productivity in the United States has declined significantly for reasons that are not entirely clear, it would be instructive to have comparable series for other countries available for comparisons. Recommendation 23. The Panel endorses continued support for private research on international productivity comparisons, with some enlargement of its scale as opportunities for useful projects arise. The Panel also believes that research within government should be expanded along lines to be determined by the relevant agencies and adjusted on the basis of their experience. As an important component, we suggest, once the Bureau of Labor Statistics and the Bureau of Economic Analysis have established a program to provide growth- accounting series on a regular basis for the United States, they extend their series to other industrial countries.

OCR for page 181
International Comparisons of Productivity NOTES 205 1. Interregional comparisons within the United States, or within other countries, can also be made, but they are outside the scope of this chapter. 2. For a comprehensive bibliography, see Kravis (1976~. 3. During this period a comparison of the United States and Canada in 1965, based on a shortcut version of the Gilbert-Kravis methodology, was conducted by E. C. West (1968; see also appendix to Daly and Walters 1967~. The Economic Commission for Latin America did a study of its members, and a scattering of comparisons not involving the United States were also made. 4. See, for example, Central Intelligence Agency (1976), Cohn (1970), Bergson (1974), and Daly (1972) and, for Soviet estimates, the comment by Zoltan Kennessey in Daly (1972, p. 196~. 5. The most recent report is Block (1978~. 6. For convenient sources of long time series on national accounts in constant dollars, see the International Bank for Reconstruction and Development (1976), and for trade statistics, see the International Monetary Fund's International Financial Statistics. 7. A number of other studies have used, or are using, Denison's approach as a model for other countries, and shortcut studies have been made for many countries for which the basic data were very sketchy. 8. These include Abramovitz and David (1973a, 1973b), Abramovitz (forth- coming), Fua (1965, 1969), Silenstam (1970), Lundberg (1969), Bergstrom (1969), Ohlsson (1969), Aberg (1969), Wulf (1968), and Matthews (1973~. 9. Studies using this approach include Christensen et al. (1975a, 1975b, 1980), Brazell et al. (1975), Christensen and Cummings (1975), Christensen and Jorgenson (1970), and Jorgenson and Zaki (1973~. 10. The functions of the National Center were shifted to the U.S. Department of Labor and the U.S. Department of Commerce as of September 30, 1978.