Income comparisons between persons or groups of persons in different countries are a special field of inquiry because there are different currency units. For the most part, the other theoretical and empirical problems encountered in international income comparisons are similar to those of within-nation comparisons between different persons at the same time or different groups of persons either interspatially or, what is more common, intertemporally (e.g., constant price series of national income). The qualifications mainly and the most part are included because some writers hold that international comparisons are complicated or even invalidated by differences in consumption patterns. Sometimes differences in tastes are held to underlie these variations in consumption patterns, variations that are often much larger between countries than those found between regions within a country or between different periods in the same country. This essay focuses, in its methodological aspects, on these special problems. The basic problems that are common to international and within-nation comparisons are left to the standard literature on national accounts.2 We turn now to the currency unit problems and reserve the question of tastes for a later section.

I. Objectives of the United States reciprocal trade program, 273. — II. A program based on tariff quotas and the growth criterion, 276. — III. Advantages and problems of a policy based on tariff quotas and the growth criterion, 281.

N this paper the distribution of before-tax income among consumer units in ten other countries is compared with that of the United States, and an attempt is made to explain the differences that are observed. The comparisons are made by selecting, from the relative wealth of American data, distributions that match those for other countries as closely as possible with respect to the strata of society covered, the concept of the incomereceiving unit, the definition of income, and general technique (e.g., whether tax returns or sample surveys or both were the basic source of information). The results are presented in a highly summarized form in Table i. The measures of inequality, the source materials, the way in which different bodies of data were matched, and the biases affecting particular comparisons are discussed in detail elsewhere.' Although we shall summarize some of the general sources of bias affecting the comparisons, our main attention in this paper will be devoted to the explanation of international differences in equality. While the comparisons in Table I still contain unknown margins of error, it seems likely that Denmark, Israel (Jewish population only), and the Netherlands have less inequality than the United States (with more certainty about Denmark than the others); Great Britain, Japan, and Canada about the same degree of inequality (with the first probably having a little less and the last a little more inequality than the United States); and Italy, Puerto Rico, Ceylon, and El Salvador more inequality than the United States (in most probable order of increasing inequality). The position of the last four countries tends to confirm the results of earlier comparisons indicating greater inequality in underdeveloped countries than in developed ones.2 Indeed, the remaining biases in the comparisons probably work in the direction of understating the relative equality in the distribution of income in the United States vis-a-vis the other countries, and, more generally, in the developed vis-a-vis the underdeveloped countries. Among the factors that tend to bias the comparisons so as to underestimate the extent to which the underdeveloped countries have less equality are (i) the frequent inclusion of non-money incomes in the data of developed countries and their exclusion in many distributions of the undeveloped countries, (2) the possibility that the lengthening of the accounting period beyond one year might reduce inequality more in the developed than in the underdeveloped countries, (3) the effect of old age pensions prevalent in the rich but not in the poor countries -in splitting off older individuals from units containing economically active persons and thus increasing the relative number of lowincome units in the rich countries, and (4) the likelihood that high incomes tend to escape measurement to a greater degree in underdeveloped countries which tend to have less efficient tax administration. The major factors working in the opposite direction are (i) the omission of incomes accruing mainly to high income units in the form of capital gains, expense accounts, and other tax-free forms, and (2) the existence of international differences in price structure of such a character that interclass differences in prices reduce the observed inequality in the distribution of money incomes more in poor than in rich countries. Considering the varied aspects of inequality measured by our five indexes, the results ' See Chapter VII of the author's Structure of Income, a forthcoming volume in the monograph series of the University of Pennsylvania's Wharton School Study of Consumer Expenditures, Incomes and Savings. The Study has been supported by a grant from the Ford Foundation. The author wishes to acknowledge helpful comments on an earlier version of this paper made by R. A. Easterlin, I. Friend, L. R. Klein, and S. Kuznets. Mr. Manoranjin Dutta did the statistical work. 2T. Morgan, Distribution of Income in Ceylon, Puerto Rico, the United States and the United Kingdom, Economic Journal, LXIII (December I953), 82I-34; and S. Kuznets, Regional Economic Trends and Levels of Living, F. M. Hauser, ed., Population and World Politics (Glencoe, Ill., I958), 79-II7.

THIS paper summarizes the results of an exploratory study of wage levels in relation to the foreign trade of the United States.' Attention is concentrated on manufacturing industries, although mining and agriculture are also included. The broad conclusion is that export industries tend to pay higher wages than importcompeting industries (i.e., the industrial classifications within which the largest volume of competitive imports fall). Within the manufacturing sector, the difference is associated with the higher wages paid by durable goods industries as compared with nondurable goods industries; that is, a large part of the difference is connected with the fact that the proportion of durable goods industries is higher among the leading export industries than among the leading import-competing industries. Even greater wage differences are found when the leading export industries are compared with industries producing commodities most vulnerable to further import competition or with industries producing goods that are competitive with highly protected imports. Relatively higher wages have characterized exported manufactures for over half a century. The same tendency for average wages in the export group to exceed those of the import-competing group is found in the mining and agricultural sectors, but for agriculture the analysis of wages in relation to foreign trade is much more tentative. The reconciliation of foreign trade statistics with data on wages, employment, and production for the years I947 and I952 had been accomplished by the Bureau of Labor Statistics in connection with its input-output studies.2 Average hourly earnings, which are used as the measure of wages,3 are either calculated from Census data or taken from B.L.S. estimates. For agriculture, however, there was no source of hourly earnings data, and, as is explained in the section on agriculture, the author has employed a makeshift method based on data of the Bureau of Agricultural Economics.

J OOD prices in Moscow state stores in 1, 1larch I950 were approximately 2.25 to 2.5 times higher than those prevailing in January I936, according to various price indices coimputed below. These indices indicate that the March I950 food prices were, however, about 30 per cent lower than those established in connection with the currency reform and the abolition of rationing in December I947. The construction of consumers' price indices for the Sovi.et Union is subject to a large margin of error because of (i) the paucity of price and consumers' expenditure data and (2) the country's multiple price system. Each of these complications is discussed briefly. The available price data are then tabulated, and, finally, the methods and results of the index number computations are presented.

This paper compares the industry characteristics that determine U.S. export competitiveness with those that affect the export competitiveness of U.S. multinationals. Higher R&D and human capital intensities are associated with high U.S. shares in exports and, more strongly with high U.S. multinationals' shares, and higher labor content with low shares. Within the multinationals, high R&D intensity leads to a larger share of the firms' exports being supplied from the U.S. parent. The higher the labor intensity and the level of advertising expenditures in an industry, the larger the proportion of the firms' exports supplied by the foreign affiliates.