CONCENTRATION. EXPORT EARNINGS. AND THE TERMS OF TRADE Thesis for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY MICHAEL RICHARD EDGMAND 1968 TH ESIS Date 0-169 This is to certify that the thesis entitled CONCENTRATION, EXPORT EARNINGS, AND THE TERMS OF TRADE presented by Michael Richard Edgmand has been accepted towards fulfillment of the requirements for Ph . D. degree in Economics WK /6fl£7[;v/Lp 7/, Major professor Nov. 22, 1968 ABSTRACT CONCENTRATION, EXPORT EARNINGS, AND THE TERMS OF TRADE by Michael Richard Edgmand There appears to be a widespread feeling among government offi- cials in underdeveloped countries that traditional exports, mainly primary products, should not be expanded. Instead, they believe investment should be channelled to other areas, usually in some form of industrialization, either to create new exports or to reduce imports. Although there are undoubtedly a variety of reasons for this feeling, only two are discussed. First, it is held that there is a tendency toward secular deterioration of the commodity terms of trade of underdeveIOped countries. Such a deterioration is presumed to indicate a welfare loss or, at best, an unequal distribution of the gains from international trade. As a consequence, officials in underdevelOped countries appear reluctant to allocate resources to the export sector. Second, officials in underdeveloped countries are reluctant to allocate resources to traditional export industries because it is said that concentration on a narrow range of products leads to greater year-to-year fluctuations in export earnings. Such fluctu— ations may lead to instability of national income, a mis—allocation of resources, and balance of payments problems. Hence, if exports are to be increased, the expansion should be in new rather than existing product lines. Nichael Richard Edgmand If these views persist and are unfounded, they could lead to a mis—allocation of resources; therefore, it is of importance to con- sider the following questions. First, has there been a deteriora— tion of the commodity terms of trade of the underdeveloped countries in the post—war period? Second, is diversification of exports likely to stabilize export earnings? To test the hypothesis that the commodity terms of trade of underdevelOped (developed) countries inevitably decline (improve), the commodity terms of trade of the countries involved were computed using export and import price indices published by the United Nations. After calculating the terms of trade of the sixty-two countries in the sample, the countries were divided into two groups--developed and underdeveIOped--on the basis of the baSis of per capita income. The commodity terms of trade of each group of countries were aggregated into a single index. The aggregate indices reveal no clear trend over the 1948-1964 period for either the twenty coun- tries classified as deveIOped or the forty-two countries classified as underdeveloped. As an alternative to dividing the countries into develOped and underdeveloped groups, the countries were considered individually and the commodity terms of trade of each country examined. In so doing, it was possible to test a hypothesis postulated by Charles P. Kindle- berger. According to the hypothesis, movements in the commodity terms of trade are related to the stage of a country's development with the most deve10ped countries eXpected to experience the most Michael Richard Edgmand favorable terms of trade over time and the least developed countries the most unfavorable. However, the results of this study show no significant relationship between the stage of a country's development and movements in its terms of trade. As a consequence, the hypothesis is taken as unproven. To test the hypothesis that export instability is related to export concentration, measures of export earning instability and the commodity concentration of exports were developed utilizing data published by the International Monetary Fund and the United Nations. Using regression analysis (with the measure of export instability as the dependent variable), a significant relationship was found be— tween export earning instability and commodity concentration of exports. This suggests that diversification of exports should, in general, lead to a reduction in export instability. However, as a practical matter, the results suggest that the amount of stabiliza— tion obtained from any given reduction in export concentration is likely to be minor. CONCENTRATION, EXPORT EARNINGS, AND THE TERMS OF TRADE By Michael Richard Edgmand A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Economics 1968 ACKNOWLEDGMENTS For his many helpful suggestions at all stages, I wish to thank my adviser and thesis committee chairman--Dr. Mordechai E. Kreinin. His interest, promptness, and efficiency in reading the various drafts are especially appreciated In addition, I am grateful to Dr. Peter J. Lloyd, Dr. John R. Moroney, and Dr. Paul W. Strassman for their valuable suggestions at different stages. To my wife, Judy, I would like to convey my gratitude for her patience and encouragement during the years leading to the completion of the thesis. Finally, I am indebted to the Ford Foundation for support of the study during the 1965-1966 academic year. However, I wish to add that the views expressed here are those of the author and are not necessarily those of the Ford Foundation. ii TABLE OF CONTENTS Page ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . ii LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . v LIST OF ILLUSTRATIONS . . . . . . . . . . . . . . . . . . . . . .Viii LIST OF APPENDICES . . . . . . . . . . . . . . . . . . . . . . . ix INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Chapter I. THE TERMS OF TRADE . . . . . . . . . . . . . . . . . . . 3 The Terms of Trade and Economic welfare Trends in the_Commodity Terms of Trade: Theoretical Foundation Trends in the Commodity Terms of Trade: Statistical Foundation II. POST—WAR TRENDS IN THE COMMODITY AND INCOME TERMS OF TRADE . . . . . . . . . . . . . . . . . . . . 36 -—The Data umThe Classification of Countries as Developed or Underdeveloped -—The Commodity Terms of Trade of the Developed Countries --The Commodity Terms of Trade of the Under— developed Countries The Commodity Terms of Trade and Stages of Economic Development Comparison with the Pre-War Period The Income Terms of Trade Summary III. FLUCTUATIONS IN THE COMMODITY AND INCOME TERMS OF TRADE . . . . . . . . . . . . . . . . . . . . 94 Measures of Instability The Commodity Terms of Trade The Income Terms of Trade Conclusion iii Chapter Page IV. INSTABILITY OF EXPORT EARNINGS . . . . . . . . . . . . . 113 Instability and Its Effects Causes of Instability The Empirical Evidence Sources of Export Earning Instability Summary V. EXPORT CONCENTRATION AND FLUCTUATIONS IN EXPORT EARNINGS . . . . . . . . . . . . . . . . . . . . 145 Export Instability and Commodity Concentration: Empirical Results Export Instability and Export Concentration: Commodity and Geographic Export Instability and Primary Products Conclusion VI. SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . . . . . 186 The Terms of Trade Hypothesis The Export Concentration Hypothesis APPENDICES . . . . . . . . . . . . . . . . . . . . . . . . . . . 198 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . 215 iv Table 10. 11. 13. LIST OF TABLES The Developed Countries, Ranked by Per Capita Gross Domestic Product, 1963 . . . . The Underdeveloped Countries, Ranked by Per Capita Gross Domestic Product, 1963 The Aggregate (Unweighted) Commodity Terms of Trade of the Developed Countries, 1948-1964 The Aggregate (Weighted) Commodity Terms of Trade of the Developed Countries, 1948-1964 Trends in the Commodity Terms of Trade of TWenty Deve10ped Countries as Measured by the Least Squares Technique: 1948-1964 Trends in the Commodity Terms of Trade of Twenty Deve10ped Countries as Measured by the Least Squares Technique: 1954-1964 The Aggregate (Unweighted) Commodity Terms of Trade of the Underdeveloped Countries, 1948-1964 . The Aggregate (weighted) Commodity Terms of Trade of the Underdeveloped Countries, 1948-1964 . Trends in the Commodity Terms of Trade of Thirty- one Underdeveloped Countries as Measured by the Least Squares Technique: 1948-1964 Trends in the Commodity Terms of Trade of Forty- two Underdeveloped Countries as Measured by the Least Squares Technique: 1954-1964 Percentage Change in the Commodity Terms of Trade of Fifteen Developed Countries: 1937 to 1963-1964 . Percentage Change in the Commodity Terms of Trade of Eleven Underdeveloped Countries: 1937 to 1963-1964 The Aggregate (Unweighted) Income Terms of Trade of the Developed Countries, 1948-1964 V Page 44 45 49 50 53 57 6O 61 68 78 79 82 Table Page 14. The Aggregate (Unweighted) Income Terms of Trade of the Underdeveloped Countries, 1948-1964 . . . . . . . 83 15. The Aggregate (Weighted) Income Terms of Trade of the Developed and Underdeveloped Countries, 1948-1964 . . . . . . . . . . . . . . . . . 86 16. Increases in Capacity to Import of Forty-eight’ Countries as measured by Changes in the Income Terms of Trade, 1948—1964 . . . . . . . . . . . . . . . 88 17. Instability of Export Prices, Import Prices, and the Commodity Terms of Trade in a Sample of Forty-five Countries, 1948-1964 . . . . . . . . . . . . 102 18. TwO Measures of Instability of the Income Terms of Trade in a Sample of Forty-five Countries, 1948-1964 . . . . . . . . . . . . . . . . . . . . . . . 108 19. Three Measures of Instability of Export Earnings in a Sample of Seventy-eight Countries, 1948-1964 . . . 124 20. Instability of Export Prices and Volume in a Sample of Forty-five Countries, 1948-1964 . . . . . . . 133 21. A Summary of Instability in International Trade, 1948-1964 . . . . . . . . . . . . . . . . . . . . 139 22. Three Measures of Export Concentration by Commodity in a Sample of Thirty-three Countries, 1963 . . . . . . 151 23. A Comparison of Export Concentration by Commodity in 1954, 1959, and 1963 . . . . . . . . . . . . . . . . 153 24. A Comparison of Export Concentration by Commodity in 1954-1959, and 1963: Summary . . . . . . . . . . . . 155 25. The Relationship Between Export Instability and Export Commodity Concentration . . . . . . . . . . . . . 158 26. A Comparison of Geographic Concentration of Exports in 1954, 1959, and 1963 . . . . . . . . . . . . 169 27. A Comparison of Geographic Concentration of EXports in 1954, 1959, and 1963: Summary . . . . . . . 171 28. The Relationship Between EXport Instability and EXport Concentration: Commodity and Geographic . . . . 174 29. The Ratio of Primary-Product Exports to Total Exports. 1959 and 1963 . . . . . . . . . . . . . . 177 Table 30. 31. 32. 33. The Relationship Between Export Instability and Export Concentration Including the Ratio of Primary to Total Exports . . . . Percentage Change in the Commodity Terms of Trade of Twenty Developed Countries: 1948-1949 to 1963-1964 . . . . . . . . . . . . Percentage Change in the Commodity Terms of Trade of Thirty-one Underdeveloped Countries: 1948-1949 to 1963-1964 . TWO Measures of Instability of Import Volume in a Sample of Forty-one Countries, 1948-1964 . vii Page . 179 . 203 . 205 . 212 Figure LIST OF ILLUSTRATIONS The Commodity Terms of Trade of Twenty Developed Countries, 1948-1964: Aggregate Indices . The Commodity Terms of Trade of Forty-two Underdeveloped Countries, 1948- 1964. Aggregate Indices . . . . . . The Income Terms of Trade of TWenty Developed and Thirty-nine Underdeveloped Countries, 1948-1964: Aggregate (Mean) Indices . The Income Terms of Trade of Twenty Deve10ped and Thirty-nine Underdeveloped Countries, 1948-1964: Aggregate (Median) Indices . . . The Income Terms of Trade of Twenty Developed and Thirty-nine Underdeveloped Countries, 1948-1964: Aggregate (Weighted) Indices . viii 62 84 85 87 LIST OF APPENDICES Appendix I. The Commodity Terms of Trade of Sixty-two Countries During the Post-war Period . II. Estimated Levels of "Real" Consumption Per Head in 1960 . . III. A Comparison of the Commodity Terms of Trade of Twenty Developed Countries in 1948-1949 and 1963-1964 . . . . . . . . . . . IV. A Comparison of the Commodity Terms of Trade of Thirty-one Underdeveloped Countries in 1948-1949 and 1963-1964 V. The Income Terms of Trade of Fifty-nine Countries During the Post-war Period . VI. Instability of Import Volume . ix Page . 198 . 202 . 203 . 205 . 207 . 211 INTRODUCTION In recent years there has been considerable discussion of the relationship between international trade and economic development. The feeling is widespread among government officials in under- developed countries that traditional exports, mainly primary prod- ucts, should not be expanded but rather investment should be channelled to other areas, usually in some form of industrialization, either to create new experts or to reduce imports. This belief is popular for at least two reasons. First, it is widely held that there is a tendency toward secular deterioration of the terms of trade of underdeveloped countries. Second, the concentration on a narrow range of products for export is said to lead to greater fluctuations in export earnings. Since the margin for error in development planning is small, it is important to determine, first, whether there has been a secular deterioration of the terms of trade of the underdeveloped countries and, second, whether diversification of exports is likely to provide a greatly increased measure of stability in export earnings. This study will attempt to test these two hypotheses empirically. The first hypothesis--that the terms of trade of underdeveloped countries inevitably decline--is tested in Chapters I and II. After' discussing the various terms of trade concepts and the relationship between economic welfare and the commodity terms of trade, the secu- lar deterioration argument is outlined in Chapter I and its 1 2 theoretical and statistical underpinnings critically evaluated. The empirical evidence for the post-war period is presented in Chapter II.1 The conclusion is that there is little evidence to suggest that the commodity terms of trade of the underdeveloped countries have declined. Post-war trends in the income terms of trade are also examined in Chapter II. Fluctuations in the commodity and income terms of trade are examined in Chapter III. The second hypothesis--that eXport earning instability is related to export concentration--is examined in Chapters IV and V. Chapter IV is devoted to a discussion of instability in international trade and contains empirical evidence relating to fluctuations in export earnings in the post-war period. The Chapter also lays the groundwork for the tests found in Chapter V. The conclusion is that instability of export earnings is related to export concentration; hence, export diversification should lead to greater stability of export earnings. The results and conclusions are summarized in Chapter VI. 1The study is restricted to the post-war years as the pre-war period has been studied rather exhaustively by others. CHAPTER I THE TERMS OF TRADE There are at least seven different concepts of the terms of trade, some of which do not lend themselves to empirical measurement. The most familiar is the commodity or net barter terms of trade which is the ratio of a country's export prices to her import prices relative to some base period. Symbolically, the commodity terms of trade can be defined as eP 0 'v I O i—' 0 p. 'U H 1.:- Po where e_represents exports, i,represents imports, P the price index number, 9 the initial year and 1,the given year.1 A rise in the commodity terms of trade indicates that a larger volume of imports could be obtained, on the basis of price relations only, in exchange for a given volume of exports; therefore, a rise in a country's com- modity terms of trade index is usually considered "favorable" to that country. Export and import price indices necessary to compute the commodity terms of trade are available for sixty-two countries for at least part of the post-war period. 1The notation is Jacob Viner's. See his Studies in the Theory 9f International Trade (New York: Harper and Brothers, 1937), p. 538. 4 Since countries, particularly the less-developed, are concerned not only with their export-import price ratios but also their "capac- ity to import," it may be desirable to correct the commodity terms of trade for changes in export volume.2 A second terms of trade concept, the income terms of trade, is designed for this purpose and can be expressed as .31. eP eQ eQ To q = O __l_. = T , __l_ where Q represents a volume ’ 1P eQ eQ __1 o o 1P index.3 A rise in Tc,q indicates that a country can obtain a larger volume of imports from the sale of its exports; hence, such a change is deemed "favorable." This "capacity to import" is based on export earnings and should not be confused with total capacity to import which depends not only on exports but also capital flows and other invisible exchange receipts and payments. In some cases, however, the income terms of trade may be mis- leading, particularly as a guide to changes in economic welfare. Consider the following examples.4 First, assume that import prices and quantities are unchanged and the value of exports and imports 2G. S. Dorrance, "The Income Terms of Trade,” Review of Economic Studies, XVI (1948-49), 50-56. 3Albert H. Imlah uses a ”total gain from trade" index in which the quantity of exports is replaced by the quantity of total trade (exports and imports). Imlah, ”The Terms of Trade of the United Kingdom, 1798-1913," Journal of Economic Histogy, X (November, 1950), 175-83. “The first example is Gottfried Haberler's. See his A Survey of nternatiggal Trade Theory (lst ed. rev.; Princeton, New Jersey: Princeton University Press, 1961), p. 27. 5 remain equal. If export prices fall by ten percent and export quan- tities increase so as to keep the value of exports constant, the income terms of trade will show no change. Second, assume that export prices remain constant while the volume of exports increases. If import prices increase proportionately , the income terms of trade will not change. In each case, however, the country is worse off than before because for the same volume of imports, it must export more. The commodity terms of trade indicate, correctly, a deteriora- tion in economic welfare while the income terms of trade show no change. Therefore, the income terms of trade concept appears more useful as a measure of "capacity to import“ than as a measure of the gain from trade.5 Data to compute the income terms of trade for at least part of the post-war period are available for fifty-nine countries. A third concept, the gross barter terms of trade, has been sug- gested by Frank W. Taussig as being more useful than the commodity terms of trade if there are unilateral transfers in the country's balance of payments.6 The gross barter terms of trade, a measure of the rate of exchange between the whole of a country's physical ' imports as compared with the whole of its physical exports, can be 5E1y Devons, "Statistics of United Kingdom Terms of Trade," The Manchester §chool of Economic and Social Studies, XXII (September, 1954), 268-69. United Nations,.Department of Economic Affairs, Economic Commission for Latin America, Economic Survey of Latin America, 1242 (New York, 1951), p. 15. 6Frank‘W. Taussig, International Trade (New York: The MacMillan Company, 1927), pp. 113-14. denoted as An increase in the gross barter terms of trade indicates that more imports are received for a given volume of exports; hence, an ' increase in Té is usually declared "favorable." If there are no uni- lateral transfers, exports which are surrendered without compensation, or imports which are received without a corresponding counterpayment, Tc = Tq; otherwise, the gross barter and commodity terms of trade will diverge.7 As Jacob Vinera and Gottfried Haberler9 have pointed out, the gross barter terms of trade can be misleading because the concept treats as equivalent cases situations which have to be judged sepa- rately. For example, a country's gross barter terms of trade may decline because the country is paying reparations or because it exports capital. While the gross barter terms of trade concepts treats both cases equivalently, the impact on the economy is not the same and it is clear that they should be judged separately.lo For this reason, the gross barter terms of trade concept is seldom used although data in the form of export and import quantity indices are 7Gerald M. Meier, Internatgonal Trade and Development (New York: Harper andRow, 1963), p. 42. 8Viner, 29. 213., pp. 562-63. 9Haberler, _p.'git., p. 27. loErich Schiff, ”Direct Investments, Terms of Trade, and Balance of Payments,” anrterly Jggrnal of Economics, LVI (February, 1952), 310-16 . g 7 available to calculate TA for fifty to sixty countries in the post- war period. Another concept, the single factoral terms of trade, is designed to take into consideration changes in productivity in the export sector and consists of the commodity terms of trade multiplied by an export productivity index. If E represents a productivity index, the single factoral terms of trade can be written as B I: O 6 6 T Po F1 T F1 11 c,f = -I--" ;- = c ;- ' I; I“0 F0 1 Po A rise in the single factoral terms of trade indicates that a greater quantity of imports can be obtained per unit of factor input used in the production of exports; hence, a rise in Tc,f is considered "favorable.” Clearly, the commodity and single factoral terms of trade will diverge if there is technological change in the export sector. Since productivity changes are likely to be important over time, it would be desirable to compute the single factoral terms of trade: unfortunately, productivity statistics are difficult to obtain. Hence, little use has been made of this concept.12 11The notation differs slightly from Viner's in respect to the single and double factoral terms of trade. Viner, 9p. git" pp. 559-61. 12E1y Devons has compiled the single factoral terms of trade of the United Kingdom for 1935 and 1946 through 1953. Devons, pp. 213., pp. 265-68, 273. Robert E. Lipsey has calculated the single factoral terms of trade for United States' agricultural and manufactured prod- ucts for 1879 and 1889 through 1957. Lipsey, Price and Qgentity Igends in the Forei rade of the nited tates (Princeton, New Jersey: Princeton university Press, 19 3 , pp. 25-30, 465-68. 8 The double factoral terms of trade is similar to the previous concept except that the double factoral terms of trade takes into consideration changes in productivity in the export sector of foreign countries. Symbolically, the double factoral terms of trade can be represented as e e 3.1. .31 e e T __ 1’0 F0 , c,ff - 1P 1F __l. ._;L i i Po F‘o An increase in the double factoral terms of trade indicates that one unit of domestic factors embodied in exports now exchanges for more units of the foreign factors embodied in imports; therefore, such a change is said to be "favorable.“ The single and double factoral terms of trade will diverge when there is a change in the factor cost of producing imports. However, the double factoral terms of trade have little relevance to the welfare of the importing country since it is concerned with whether it receives more goods per unit of resources engaged in export production, not whether these imports contain more or less foreign inputs than before. Statistically, it would be extremely difficult to compute an import productivity index since most countries import commodities from many countries. Viner lists two other concepts of the terms of trade--the real cost and utility terms of trade.13 Both are designed to measure changes in economic welfare by introducing utility functions into the expression. However desirable the real cost and utility terms of trade, they are impossible to calculate since they require knowledge l3V’iner, 9p.‘git., pp. 559-60. of the various utility functions. In summary, we have noted seven different terms of trade con- cepts: however, the last two are clearly impossible to calculate. 0f the rest, the gross barter and double factoral terms of trade are not particularly useful. The three remaining concepts, the commodity, income, and single factoral terms of trade, appear useful, but the latter is impossible to calculate for a large number of countries because of the lack of productivity data. Since the income terms of trade is mainly a measure of "capacity to import" we are left with only the commodity terms of trade as both a possibly useful theoreti- cal concept for measuring the gain from trade and one that is readily available for a large number of countries. In the next section, the relationship between the commodity terms of trade and the gain from trade is discussed. A11 further references to the terms of trade are to the commod- ity terms of trade unless otherwise stated. 'We shall return to the income terms of trade briefly at the end of the next chapter since the concept does provide a measure of the country's export-based capacity to import. The Terms of Trade andEeonomic'Welfare A change in a country's terms of trade can affect its economic welfare in a number of ways.14 If a country's terms of trade improve as its domestic output expands, its real income rises faster than output since each unit of exports now exchanges for more imports. An 1”It is assumed that changes in economic welfare are reflected in changes in real national income. 10 improvement in the terms of trade may also release resources from the export for domestic expansion since the same amount of imports can be obtained for fewer exports. Moreover. the government may be able to capture all or part of the gain through taxation or increased profits of governmental marketing boards, thereby making more resources available for economic development. However, should a country's terms of trade deteriorate as its domestic economy eXpands, a part of the benefit from the expansion is transferred to other countries.15 Theoretically, it is possible that the deterioration of the terms of trade may be so great that the gain from the growth in out- put is more than offset by the loss from the adverse terms of trade. This phenomenon, called immiserizing growth by Jagdish Bhagwati,l6 would arise only if either the growing country faces an inelastic demand schedule for its exports or, at constant relative commodity prices, growth actually reduces the domestic production of impor- tables.l7 Such a situation is unlikely to occur if the country has some flexibility in its structure of output so that resources can move from one sector to another. If the necessary conditions do l51f growth occurs only in the country in question, the presump- tion is that the country's terms of trade will decline provided that the foreign offer curve is less than infinitely elastic. However, if the country's expansion is ultra-import biased, the country's terms of trade will improve. Meier, _p. git., pp. 46-47. If growth is also occurring in other countries, there is no such presumption. 16Jagdish Bhagwati, "Immiserizing Growth: A Geometrical Note," Review of Economic §tudies, XXVI (June, 1958), 201-05. l7Neither condition is sufficient. For a mathematical formu- lation of the necessary conditions, see Jagdish Bhagwati, "Inter- national Trade and Economic Expansion," American Economic Review, XLVIII (December, 1958), 949-50. ll exist, the country could impose taxes to offset the deterioration of its terms of trade.18 _So far, it appears that an improvement (a deterioration) in the terms of trade marks a gain (loss) in welfare; however, such is not necessarily the case. It depends on the cause of the change in the terms of trade. An improvement in a country's terms of trade caused by a shift in the foreign offer curve, with the country's own offer curve unchanged, is always favorable provided it doesn't lead to widespread unemployment in the country's export industries.l9 Simi- larly, a deterioration of the terms of trade caused by a shift in the foreign country's offer curve is clearly unfavorable given the level of employment. However, if the domestic offer curve also changes, it is necessary to consider the cause of this change to determine its impact on economic welfare. First, as has long been recognized, it is possible for a country to improve its terms of trade by restricting the volume of its trade, assuming that the foreign offer curve is less than infinitely elas- tic and the improvement is not offset by retaliation of other countries. Up to a certain point,20 an improvement in the terms of trade will increase welfare; however, after that point, further 18R. A. Mundell, ”The Pure Theory of International Trade,“ American Economic Review, L (March, 1960), 85. 19The shift could lead to unemployment if the home countny's offer curve were inelastic in the relevant range. Gottfried Haberlen "Terms of Trade and Economic Development," Economic Development for Latin America, ed. Howard Ellis (London: macMillan and Company, ‘ ' 1963). p- 277. 20The point where the foreign offer curve is tangent to the home country's highest indifference curve. James E. Meade, A Geometr of International Trade (London: George Allen and Unwin, Ltd., 1952), p. 76. 12 increases in welfare will not be forthcoming as the rise in the terms of trade is offset, or more than offset, by a fall in the volume of trade.21 Thus, an increase in the terms of trade does not necessar- ily mean an increase in welfare. Second, a country's offer curve may change because of increased productivity in the export sector. While this may lead to a deteri- oration in the commodity terms of trade, it does not necessarily ' imply a reduction in welfare. So long as productivity in the export sector is rising faster than the prices of its exports are falling, the country's real income is increasing despite the deterioration in its terms of trade.22 Clearly, the single factoral terms of trade concept is more relevant here. Furthermore, even if productivity is constant in the export sec- tor, a deterioration in the terms of trade is not sufficient evidence of a loss in welfare. If factors are employed in the import- competing sector with lower productivity than in the export sector and export industries expand to absorb more of these factors, the real income of the economy will increase despite the deterioration in the terms of trade.23 21Haberler, "Terms of Trade and Economic Development," op, git,, p. 278. For a discussion of the optimum tariff, see Harry G. Johnson, ”Optimum Tariffs and Retaliation," International Trade and Economic Growth: Studies in Pure Theory (London: George Allen and Unwin, 1958), pp. 31-61. 22The country's gains would have been greater, of course, had the terms of trade not deteriorated. 23Robert E. Baldwin, "Secular Movements in the Terms of Trade," American Economic Review, Papers and Proceedings, XLV (May, 1955), 263-64. Theodore Morgan, "The Long-run Terms of Trade Between Agri- culture and Manufacturing," Economic Development and Cultural Change, VIII (October, 1959), 17-19. 13 Finally, it is always possible that the commodity and income terms of trade move in opposite directions. If the foreign offer curve is elastic, or shifts out sufficiently as the country's offer curve shifts, the volume of exports may increase enough to improve the income terms of trade despite a deterioration in the commodity terms of trade. In a country concerned with financing economic deve10pment, the income terms of trade may be more relevant than the commodity terms of trade. From these examples, it is evident that knowledge of the change in the commodity terms of trade is not very useful in itself in drawing conclusions as to the change in economic welfare.2u It is essential to go beyond the terms of trade to analyze the cause of the change in order to examine the welfare implications. Trends in the CommoditygTerms of Trade: Theoretical Foundation Despite the uncertain relationship between the commodity terms of trade and economic welfare, various economists have argued that there has been a very uneven distribution of the gains from inter- national trade as evidenced by a secular movement in the terms of trade of certain groups of countries.25 From the ensuing controversy, 2“Hans Staehle, "Some Notes on the Terms of Trade," Inter- nationalg§ocial Science Bulletin, III (Spring, 1951), 33-37. 25United Nations, Department of Economic Affairs, Economic Commission for Latin America, The Economic Development of Latin Amer- ica and its Primcipal Problems (Lake Success, 1950), pp. 8-14. Hans Singer, "The Distribution of Gains between Investing and Borrowing Countries," American Economic Review, Papers and Proceedings, XL (May, 1950), 473-85. Reprinted in Studies in Economic_Development, eds. Bernard Okun and Richard W) Richardson (New York: Holt, ' Rinehart and Winston, 1961), pp. 170-83. Page references are to the latter. Raul Prebisch, ”Commercial Policy in the Underdeveloped 14 two questions have emerged. First, have the long-run terms of trade moved in favor of the countries that export manufactured goods at the expense of those that export primary products? Second, have the long-run trends in the terms of trade of the underdeveloped countries been unfavorable vis-aivis the developed countries? Although much of the evidence is applicable to both questions, the problems are not identical since some developed countries, for example, Australia and New Zealand, are net exporters of primary products.26 The first hypothesis--that the terms of trade of primary producing countries inevitably deteriorate vis-aivis the industrial countries--is gener- ally referred to as the Singer-Prebisch Thesis. For convenience, the second hypothesis--that the terms of trade of underdeveloped (devel- oped) countries inevitably decline (improve)--will be called the Kindleberger Thesis. Ime Singer-Prebisch Thesis There are two main theoretical explanations of the alleged secu- lar deterioration of the terms of trade of primary producing coun- ' tries.27 First, it is asserted that industrial countries are more Countries," American Economic Review| Papers and Proceedings, XLIX (May, 1959), 251-73. United Nations, roceedin s of the United flatigng Conference on Trade and Develomment (New York, 1964). United Nations, Report by the Secretary-General of the United Nations Con- ference on Trade and Development, Towards a New Trade Policy for We: (New York. 1964)- 26Other developed countries export some primary products and underdeveloped countries export some manufactured goods. 27Theodore Morgan provides an excellent summary of other expla- nations. MOrgan, "Trends in Terms of Trade and Their Repercussions on Primary Producers," Imternational Trade Theomy in a Develoming §§z%§, edited by Ro Harrod assisted b Dou las Ha s New Yerk: St a r ins' Press, 1963), pp. 55-57, 68- . I shoul be noted that this view is contrary to that commonly held by the Classical-z- ‘ 15 monopolistic than primary producing countries. Second, it is argued that there is a disparity in the rates of increase in demand for imports between the industrial and primary producing countries. Each is discussed below. According to the first argument, restrictive business practices are more common to industrial than primary producing countries; hence, the terms of trade of the industrial (primary producing) coun- tries improve (decline) because of the monopolistic tendencies of the former. However, it is difficult to understand how this explanation can suffice to explain movements in the terms of trade 9193;};mg unless it is argued that the industrial countries are becoming mgmg monopolistic. Monopolistic practices in the industrial countries could cause the terms of trade index to be more favorable (for the industrial countries) at both the beginning and end of a period than it otherwise would be. But monopoly power would influence the trend in the terms of trade only if the degree of monopoly power changed during that period.28 In Opposition to the hypothesis, it has been argued that compe- tition in world markets for manufactures is now greater than in the past since more firms and countries are now exporting manufactured economists who believed that the operation of diminishing returns in primary production would cause the prices of primary products to rise relative to prices of manufactures. See John M. Keynes restatement of the Classical position in his "Reply to Sir William Beveridge," Economic Journal, (December, 1923), pp. 476-88.. 28This argument is based on the measure of monopoly power pro- vided by Abba Lerner. Symbolically, the measure is (P - C)/P where 3 represents price and Q represents marginal cost. Lerner, "The Con- cept of Monopoly and the Measurement of Monopoly Power,” Review of Economic Studigs, (June, 1934), pp. 157-75. 16 goods.29 Moreover, "in minerals production, there may, in fact, be more monopoly than in manufacturing. ‘World production of minerals is typically dominated by a few large firms that attempt to maintain prices. Agriculture is, however, generally competitive at least until governments step in to protect or assist it."30 However, empirical evidence on monopoly power is not clear.31 Even if industrial countries are more monopolistic than pri- mary producing countries, it does not necessarily follow that the terms of trade will turn against the latter. If the rate of techno- logical progress tends to be higher under monopolistic conditions than under pure competition, long run output will expand more rapidly under monopoly than under competition.32 The faster output grows, all other things equal, the more the terms of trade will deteriorate. Hence, it could be argued that the more monopolistic the country, the more likely its terms of trade will deteriorate. 4L 29Haberler, ”Terms of Trade . . "_p. cit. , p. 284. Charles P. Kindleberger, ”Terms of Trade for Primary Products," National Resources and International Development, ed. Marion Clawson ' (Baltimore: Jehn Hopkins Press, 1964), pp. 345-46. 30M. o. Clement, Richard L. Pfister, and Kenneth J. Rothwell, eoretical Issues in nternational Economics (Boston: Houghton Mifflin Company, 19 7 , p. l 5. Government intervention in agricul- ture may be the rule rather than the exception. 31Charles P. Kindleberger, The Terms of Trade: A EurOpean Case Study (New York: John Wiley and Sons, 1956), pp. 243-45. Moreover. the United Nations has concluded that it ”is not possible to set ' forth even within broad ranges of approximation the prevalence of restrictive business practices in either national or international trade." United Nations, Economic and Social Council, Restpigtive fimsiness Practices (New York, 1953), p. 9 3ZIt has been argued that only those firms which operate on a large scale can afford the research and development necessary to systematically cut costs and introduce new products. Bo Sodersten, A§tudy of Economic Growth and International Trade (Stockholm: Almqvist and'Wiksell, 1964), pp. 164- 65. 17 It is also argued that industrial countries have relied more heavily on the imposition of tariffs than have primary producing countries. In particular, it is argued that the industrial countries have imposed tariffs on a number of foodstuffs and raw materials, primarily food, while the primary producing countries have imposed tariffs on both foodstuffs and manufactured goods. Nevertheless, there is no guarantee that the terms of trade will be affected by the imposition of a tariff; it depends on the elasticity of the offer curve facing the individual country. Moreover, when both parties impose tariffs, the end result depends on the net weight of their commercial policies. Unfortunately, "there seems to be no chance of testing empirically the view that the underdeveloped countries lose "33 on balance from commercial policy. However, as in the previous case, it is difficult to accept this as an explanation of the move- ment of the terms of trade gym; Emmg unless it is argued that the industrial countries are becoming increasingly protectionistic. One final aspect of the first argument relates to technological progress and the distribution of the gains from such progress. Hans Singer and others have argued that, although increases in produc- tivity have been greater in manufacturing than in primary production, prices of the former have increased relative to prices of the latter because of difference in the way in which the benefits of reduced costs have been shared.3# Over time, the gains from technological progress can be 33Kindleberger, The Terms of Trade: A European Case Study, pp. p13,, p. 246. 34 Singer, pp. Q;L., pp. 174-76. 18 distributed in either of two ways (or some combination)--to factors of production in the form of higher earnings or to consumers in the form of lower prices. In a closed economy, there will be an increase in real income in either case. But when international trade is con- sidered, the producers and consumers are no longer necessarily in the same country. In fact, it is argued that the gains from increased productivity in industrial countries have generally been reaped in the form of higher wages and profits while such gains in primary pro- ducing countries have resulted in lower prices. Thus, the industrial (primary producing) countries obtain, on the basis of price relations alone, increasing (decreasing) amounts of imports from the primary producing (industrial) countries for a given quantity of exports. While Singer offers no explicit explanation as to why produc- tivity gains are distributed in this fashion, Prebisch believes believes that "the characteristic lack of organization among the workers employed in primary production prevents them from obtaining wage increases comparable to those of the industrial countries (during the upswing of the business cycle) and from maintaining the increases to the same extent (during the downswing)."35 In criticism of the argument, it is not clear that trade unions and firms actually exercise sufficient monopoly power to distribute productivity gains through rising money wages and profits rather than falling prices. Moreover, even if labor unions and firms do possess power to maintain or increase prices domestically, world prices need not reflect domestic conditions since a country may find 35United Nations, The Economic Development of Latin America. 22. 923,-. p- 13- l9 itself priced out of the international market. The second main argument is based on differences in the income elasticity of demand for manufactured and primary products. It is asserted that as income increases, the demand for manufactured prod- ucts grows relatively faster than the demand for primary products; hence, there will be a decline in the price of primary products vis- a-vis manufactured products.36 The differences are attributed to the operation of Engel's law in the case of food and, in the case of raw materials, to technological progress which reduces the amount of raw materials used per unit of output and the development of synthetics and other substitutes.37 In regard to foodstuffs, however, it should be noted that while the income elasticity of demand may be quite low in the industrial countries, it is undoubtedly higher in many of the poorer countries?8 Moreover, in assessing the impact of synthetics on raw material prices from 1953 to 1962, the United Nations Food and Agricultural Organization (FAO) has tended to minimize their effect. They state: The extent to which synthetic competition has contrib- uted to this deterioration in agricultural raw material prices is indeterminate. If, at any time, the output of all synthetic materials had ceased, prices of their natural counterparts would certainly have risen in the short-term. 36Prebisch,‘pp. g;£., pp. 251-54. Singer,‘gp. g;§., p. 175. 37Engel's Law states that the percentage of expenditures on food‘ is a decreasing function of income. Singer, ibid. It is also argued that the primary producing countries face increasing output of pri- mary products in the industrial countries "which has been the result both of domestic policies, in many cases reinforced by protective barriers, as well as a general increase in productivity stemming from technological progress.” United Nations, Ezoceedings . . ., pp. g;§., p. 6. 38Kindleberger, Th! Iegms 9f Tgade: A Egrgpgam Qgse §tugy, gn.Ig;1., p. 268. 20 However, taking into account long-term supply responses, there is no similar assurance that prices would have been higher than they are today if man-made materials had never come into existence. Prices of some major agricultural products not subject to competition from man-made materials (e.g., coffee and cocoa) have declingg more rapidly than those of agricultural raw materials. The rapid rate of growth in the over-all elastomer and fibre markets has been reflected in simultaneous increases in world consumption of natural, as well as synthetic, raw materials. The progress of the former has been relatively slow, with the result that the share of natural products in total consumption has fallen sharply. This is attributable partly to competition from synthetics, which has tended to place a ceiling on prices and clouded future prospects with sufficient uncertainty to exert some retarding influence on investment. However, competition from synthetics has been only one, and probably not the most important, of the many factors limitingQBroduction in developing countries in the post-war period. Moreover, as far as the over-all level of demand for raw materials which compete with synthetics is concerned, the future appears favor- able, with the possible exception of soap-making materials]+1 Even if it is granted that the over-all elasticity of demand for primary products is low, it is improbable that all primary pro- ducing countries experience the same trend in their terms of trade. This is because most underdeveloped countries export only a narrow range of products and the income elasticity of demand varies from 39United Nations, Emoceedings . . ., _p. g;§., p. 354. "During 1959-1961, world exports of goods competing with synthetics amounted to 24% of the total value of world agricultural trade. More than half (55$) of the total originated in developing countries. The world output of synthetics is heavily concentrated in developed countries," United Nations, Emoceedings . . ., pp. g;§., p. 349. ”mg” pp. 354-55. “$11.4” p- 355. 21 commodity to commodity.“2 Under such circumstances, what is signifi- cant for a specific primary producing country is not the over-all elasticity of demand for primary products but rather the expansion in demand for its own exports.)+3 Even if the income elasticities of demand for industrial and primary products differ, it is theoretically possible that the terms of trade may turn in favor of the primary producing countries. This is because demand alone does not determine the terms of trade--supply also plays a role. Shifts in the supply curves could offset, or more than offset, the effect of the different income elasticities. As originally expounded, the two main arguments discussed above are weak. This is not to say, however, that models with realistic assumptions cannot be developed to show that the terms of trade will turn against the primary producing countries. Harry G. Johnson, for example, has developed a model applicable to the problem at hand.nn Johnson uses a two country model with one country producing mainly manufactured goods and the other producing mainly agricul- tural products. The income-elasticity of demand for manufactured goods is assumed greater than for the agricultural goods. Johnson then considers two possible patterns of technological change--equal improvement in both sectors and improvement in the manufacturing 42The fact of the matter is that coal, and timber and timber products behave very differently from, say, cotton,.fats'and oils, and petroleum products." Kindleberger, The Terms of Trade: A Europggn ggse Study, pp. 213., pp. 265-66. “BMeier, pp. g;§., p. 62. Harry G. Johnson, "Economic Expansion and International Trade:' Manchestep:§chool of Economic and Social Stmdies, XXIII (May, 1955). 95-112. For a more elegant model, see Sodersten, 22.5mm" ,pp. 177-81. 22 sector only. When technological change affects both sectors, the manufacturing country's terms of trade will improve. But when tech- nological change affects only the manufacturing sector, they will probably'worsen.h5 Thus, in theory, either result may prevail. Therefore, whether the terms of trade turn against primary producing countries and in favor of industrial countries is, ultimately, an empirical question. The Kindleberger Thesig To the extent that most develOped countries export mainly manu- factured products and most or all underdeveloped countries export chiefly primary products, the arguments discussed above also apply to the Kindleberger hypothesis. However, a theoretical argument cast entirely in terms of developed and underdeveloped countries has been suggested.”6 As a result of his study of European terms of trade, Charles P. Kindleberger concluded that movements in the terms of trade are related to the stage of a country's development.47 Hence, movements in the terms of trade should favor the develOped countries at the expense of the underdeveloped with the most highly developed coun- tries showing the most favorable terms of trade. According to Kindleberger, the basic reason for the terms of trade to turn against 45Richard E. Caves, Trade and Economic Structure: Medels and Methods (Cambridge: Harvard University Press, 1960), p. 160. uéKindleberger, The Terms of Tmade: A European Case Study, pp. 2A§., pp. 253-57. Also, "The Terms of Trade and Economic Devel-- opment," Review of Economics and Statistics, XL (February, 1958), 81-85 and Foreigp Trade and the National Economy (New Haven: Yale University Press, 1962), pp. 99-115. “7Kindleberger, The Terms of Trade: A European Case Study, ibid. 23 the underdeveloped countries is their immobility (or inadaptability) of supply in response to changes in supply conditions abroad or in demand.“ If, for example, an underdeveloped country experiences a decline in demand for its products, it will be unable to readily reallocate resources away from these sectors; hence, prices fall.49 Similarly, if world demand for products not produced by the country increases, the country will be unable to shift resources to these sectors in order to take advantage of the increase in price. Moreover, if demand for the country's products increases, the country will find itself faced with increased competition from abroad which limits the possible improvement in its terms of trade. In certain circumstances, the underdeveloped country may be "lucky"50 and find itself producing a highly profitable commodity with competition limited by the coun- try's natural advantage. In such cases, its terms of trade may improve. In contrast, resources in developed countries are more likely to move in response to price changes. According to Kindleberger, countries differ in their "capacity to transform” primarily for social reasons. The traditional society is engaged in endless repeti- tion. Consumption and production are carried on in the same way from generation to generation. Much production and consumption proceeds on a subsistence basis outside the uaMore specifically, it is assumed that supply is more inelastic in underdeveloped than developed countries. ugHowever, to the extent that the commodities are produced else- where, firms in other countries may reduce output or actually cease production. Under these circumstances, the tendency for prices to fall would be checked. Kindleberger apparently neglects this possibility. ‘ 50The term is Kindleberger's. 24 market. Change is not absence, but it occurs slowly, and is resisted. Social values dominate economic. The appe- tite for income is kinked; when social needs are filled, leisure is valued above more goods. Succeeding generations follow in the same occupations, which are determined by social status or caste. In such a society capacity to transform is limited. . ..51 In a transforming society, much has altered. Consumers are interested in increasing real income. Producers spec- ialize and work for the market, exchanging goods against money and money against goods. A higher price leads to more labor, land, and capital being attracted to a given product, and more output. A lower price results in reduced production.52 From a theoretical standpoint, Kindleberger's argument is open to several objections. First, it is not clear that inadaptability (adaptability) of supply can be strictly associated with underdevel- oped (developed) countries. Some developed countries--for example, Great Britain in recent years--may have less flexibility in adapting 53 to changing conditions than some underdeveloped countries. More- over, the "luck" element renders Kindleberger's hypothesis indeter- minate. With "luck" an underdeveloped country may have improving terms of trade; hence, prediction is impossible. Moreover, as Bo Sodersten has shown, it is not supply elasticities alone which deter- mine the outcome of the terms of trade but rather ”the growth rates in the different sectors and the demand developments induced by the economic growth."5# SlKindleberger, orei ads and the National Eponomy, pp. 212- p. 100. 521mg” p. 101. 53Kindleberger has recognized this point. Ibid., pp. 102, 109-110. SuSodersten, _p.,gi§., p. 37. 25 However, even if Kindleberger's hypothesis is open to criticism, it is possible to formulate models which would rationalize the alleged tendency. For example, suppose, in a two-factor model, capital is relatively abundant in developed countries and labor is relatively abundant in underdeveloped countries. Assume, as a conse- quence, developed countries export capital intensive goods and under- developed countries export labor intensive products. If the supply of labor grows more rapidly in underdeveloped countries than the capital stock of developed countries, the terms of trade of the underdeveloped countries will, ceteris pgribus, deteriorate. Thus, whether there is any systematic tendency for the terms of trade to turn against the underdevelOped countries is, in the end, an empiri- cal question. Before proceeding, however, it is desirable to review the evidence produced by earlier empirical work. Trends in the Commpdity Terms of Trade: Statistical Eoundation As previously noted, there are really two issues at stake, the terms of trade of industrial and primary producing countries and those of developed and underdeveloped countries. The statistical evidence relating to each is discussed below. Tme Singer-Prebisch Thesis It is claimed that there has been a secular deterioration in the prices of primary products relative to manufactures from the late 26 1800's to the eve of the Second World wer.55 As proof, the United Nations has offered three different indices all based, in varying 56 The first of these indices, Series A, 57 degree, on British data. relies, for the years 1876 to 1929, on League of Nations data which, in turn, is based on the Sauerback British wholesale price index for primary products, and Werner Schlote's price data from British trade statistics.58 The rest of Series A, covering the 1930- 1938 period, is based on world trade data from the Review of'Worlg ‘Tmmgg. Series B, also covering the 1876-1938 period, is based on Schlote's data alone. Series C, covering the years 1913-1948, is the official BritiSh Board of Trade Index converted to a 1938 base. All three indices seem to show that the United Kingdom's terms of trade, despite wide fluctuations, improved considerably over the period. From this, it is inferred that the trend was unfavorable to primary producers and the U. K.'s trading partners. Since there are no overall indices of prices paid and received by the primary producing countries over this period, the British data are appealing. They are available over a long period of time during which the United Kingdom accounted for a substantial proportion of 55United Nations, Department of Economic Affairs, Relative Prices of Ex orts and Im arts of Underdeveloped Countries (Lake Success, 1949), p. 7. 5§I§AQ., pp. 21-25. See Table V, p. 22 for the three indices. 57The main author is Folks Hilgerdt. League of Nations, Indus- trialization and Foreigm Trade (Geneva, 1945), pp. 154-57. 58Werner Schlote, British Overseas Trade from 1700 to the 1 0'3, trans.‘W. 0. Henderson and W} H. Chaloner (Oxford: Blackwell, 1952). The price indices for primary products and manufactures constructed by”W. A. Lewis also rely heavily on Schlote's data. Lewis, "World Production, Prices, and Trade, 1870-1960,” manchester School of Eco- nomic and Social Studies, (May, 1952), pp. 117-18. 27 world trade, especially during the earlier years. The U. K.'s imports were also largely primary products and her exports predomi- nantly manufactured goods. Despite the attractiveness of the data, many economists have pointed to the weak statistical base underlying the Singer-Prebisch Thesis.59 First, British data which are avail- able from 1801 to 1953 show the period 1876 to 1938 to be atypical. Concentration on the years from 1801 to 1953 shows no overall trend, but rather marked short and long-term instability.60 Second, the British data cited above cannot be used to measure the terms of trade actually experienced by the primary producing countries. In the trade statistics, import prices are reported c.i.f. (inclusive of transportation charges) at British ports of entry while export prices are reported f.o.b. (exclusive of transpor- tation charges) at British ports of exit. Therefore, a change in British import prices need not reflect a change in the prices received by foreign exporters since transportation charges may vary. The price of wheat may fall in Liverpool and rise in Argentina if ocean freight rates fall sufficiently. Similarly, a change in British export prices need not reflect a change in the prices foreign importers pay. Hence, improvement in the British terms of trade doesn't necessarily mean a deterioration of her trading partner's terms of trade; it is possible that both could be improving with ngoreover, indices constructed by Imlah and Kindleberger do not show as much improvement as Schlote's. Imlah, "The Terms of Trade of the United Kingdom, 1796-1913," Economic lements in the Pax Britan- nica (Cambridge: Harvard University Press, 1958), p. 87. Kindleberger, The Terms of Trade: A European Case Study, 2p. 9T£., PP- 53ff- 0Morgan, "The Long-run Terms of Trade Between Agriculture and Manufacturing,” pp.,gT§., pp. 2-4. 28 falling transportation costs. Data on ocean freight rates compiled by Douglas C. North and others as well as numerous examples furnished by C. Mk'Wright show that transportation costs fell dramatically over most of the 1876- 1938 time span.61 As a result, P. T. Ellsworth concludes "that a large proportion, and perhaps all, of the decline in the British price of primary products in the period between 1876 and 1905 can be attributed to the great decline in inward freight rates. . . . Since the price of British manufactured exports fell in this period by 15 percent, the terms of trade of primary countries, [if] f.o.b. prices [were] used for their exports as well as for their imports, may well have moved in their favor."62 For the 1913 to 1933 period, Ellsworth believes that falling freight rates accounted for some, but not all, 63 Kindleberger has of the improvement of the U. K.'s terms of trade. constructed a rough index of the "Current-Account Terms of Trade" 'which includes services that seems to confirm Ellsworth's findings. 61Douglas C. North, "Ocean Freight Rates and Economic Develop- ment," Journal of Economic History, XVIII (December, 1958), 537-55. C. M.‘Wright, ”Convertibility and Triangular Trade as Safeguards Against Economic Depression," Economic JournaT, LXV (September, 1955), 545ff. A. X. Cairncross, Hpme and Foreigm Investment, 1820-1913, (Cambridge, England: Cambridge University Press, 1953), pp. 170-79. Kindleberger, The Terms of Trade: A European Case Study,‘pp.‘pr., pp. 20-21. 62?. T. Ellsworth, "The Terms of Trade Between Primary Producing and Industrial Countries," Tnter-American Economic Affairs, X (Summer. 1956). 55-57- 63;p;g,. pp. 62-63. The terms of trade of primary producers 'were relatively stable from 1906 to 1912 and improved from 1933 to 1938. 6“Kindleberger, The Terms of Trade: A_§uropean Case Stud , pp. £fiL§., Chapter 11. It has been suggested that the omission of the Prices of services in computing the terms of trade indices has made tJie terms of trade of underdeveloped countries appear more 29 Moreover, in regard to the United Kingdom's trading partners, there is no reason to believe that all the primary producing coun- ' tries experienced the same trend for the period since the economic structure of countries may vary considerably.65 As demand and supply conditions are unlikely to be identical over time for all primary producing countries, we would expect the prices of some primary products to fall, some to rise, and others to remain constant, even given a general trend in primary product prices. As a consequence, not all primary producing countries need experience declining terms of trade. The same is true for industrial countries. Indeed, Theodore Morgan presents data for six countries (in. addition to the United Kingdom) and concludes, on the basis of the data, that there has been a wide variety of experiences in other countries.66 Accordingly, he emphasizes the importance of not gener- alizing from the experience of a single country. Finally, the terms of trade indices fail to make adequate provi- sion for qualitative improvements in manufactured products and for ‘the introduction of new products.67 Normally, the quality and effi- ciency of manufactured products would appear to improve more rapidly ‘than those of primary products. As a consequence, studies of changes ¥ lanfavorable, or less favorable, than they really are. Price data are, «of course, much easier to obtain for merchandise than for services. Kindleberger, ”Terms of Trade for Primary Products," pp. mi_t_., p. 342. 65Ha.berler, "Terms of Trade and Economic Development," _p. 3.1;. , p» 280. 66Morgan, "The Long-run Terms of Trade Between Agriculture and Manufacturing," pp. ,c_i_t_., pp. 2-4. 67To be sure, these problems are not limited to the indices discussed above . 30 in the terms of trade between primary producing and manufacturing countries are affected by a systematic bias which makes the change appear less favorable or more unfavorable to the primary producing countries than they really are.68 Also, one would expect hundreds of new manufactured products to appear over the years while the range of most primary products would remain relatively unchanged. This intro- duces another systematic bias since the omission of new commodities or their inclusion at relatively small beginning year weights tends to bias a price index upward as new commodities usually fall in price soon after they are introduced.69 In other studies, Kindleberger found no clear trend in the terms of trade of primary products vis-a-vis manufactured commodities.70 ”From a review of Kindleberger's data, combined with U. S. price indexes for the period since 1913, Sarah S. Montgomery found signs of improvement rather than deterioration in world terms of trade for primary products. This was especially the case when they were meas- ured in terms of prices within primary producing countries."71 RObert E. Lipsey concluded in his study of the United States that there seems to be a predominance of improving terms of trade of 68United Nations, Relative Epices of Exports and Imports of Underdevelogd Countries, _p. 933., pp. 133-34. A. N. McLeod, "Trade and Investment in Underdeveloped Areas: A Comment," American Economic Review, XLI (June, 1951), p. 414. 69Baldwin, _p. my. pp. 267-68. 7oKindleberger, Tme Terms of Trade: A Emropean Case Study, pp. 2; ., p. 263, and ”The Terms of Trade and Economic Development," 2.2. sit... pp. 72-85- 71Sarah 8. Montgomery, "The Terms of Trade of Primary Products and Manufactured Goods in International Trade, 1870-1952," (Unpub- ILished Ph.D. dissertation, University of'Wisconsin, 1960), quoted in Lipsey, pp. 213., p. 19. 31 primary products.72 More recently, the United Nations has emphasized the deteriora- tion of the terms of trade of primary products from 1950 to 1962.73 However, the agricultural terms of trade were unusually high during the early fifties; hence, this trend may not reflect a secular move- 74 This is confirmed by examining the ment in the terms of trade. terms of trade index for agricultural products compiled by the United Nations for the 1920-1938 and 1947-1962 periods.75 While this series shows wide movements in the terms of trade, there appears to be no secular change. Because of the availability of the United Nations index, the Singer-Prebisch Thesis is not tested in this study. Theggmpdleberger Thesis Since 1956, there has been a shift in emphasis from the terms of trade of countries which produce primary products and those which produce manufactured goods to the terms of trade of the underdevel- oped and developed countries. No doubt one of the reasons for the shift in emphasis has been the lack of statistical evidence to sup- port the argument in terms of manufacturing and primary producing countries. However, there seems to be little statistical evidence 72Lipsey, ibid.. pp. 20-23. 73United Nations, Epoceedings . . ., o cit., I, 120. United Nations, Towards a New Trade Eolicy for Development ,pp,. cit. , p.18. 741t has also been suggested that the trend may have reversed :itself. United Nations, Epoceedings . . ., ibid., III, 256. United Imations, Towards A New Trade PoTicy forpgeveTppment, ibid., p. 16. 75United Nations, Epoceedings . . ., ibid., III. 257. 32 to support the thesis that the terms of trade of the underdeveloped (developed) countries deteriorate (improve) over time either. As in the previous section, the United Kingdom data are offered as evidence of the secular deterioration of the terms of trade of the underdeveloped countries. As such, the data are inconclusive for the reasons discussed above. In one of the first empirical studies, K. Martin and F. G. Thackeray present data over the years from 1879 to 1913 for three industrial countries--Germany, the United Kingdom and the United 76 States. Of the three, Germany experienced a decline in her terms of trade and the United States and United Kingdom an increase. How- ever, the U. S. figures were derived from the data of Theodore J. Kreps and if National Bureau of Economic Research (NBER) data were substituted for those of Kreps, the U. 8. would show no change.77 JHence, there is no clear trend for these countries for the 1879-1913 ‘period. For the interwar period, Martin and Thackeray present data for 1?our countries--Germany, Japan, the United Kingdom, and the United Eitates. Of the four, only Japan showed a deterioration in her terms <31? trade; the rest showed an improvement. However, "the final year 76Martin and F. G. Thackeray, "The Terms of Trade of Selected <3<>untries, 1879-1938," Bulletin of the Oxford Institmte of Statistics, )C (November, 1948), 373-98, quoted in Lipsey, pp.,pr., pp. 12-13. 77Theodore J. Kreps, "Import and Export Prices in the United s"Lates and the Terms of International Trade, 1880-1914," Qparterly sléaggrnal of Econgmics, XL (August, 1926), 708-20. The NBER data are 1§<5ttnd in Lipsey, ibid. For a comparison of the two sets of indexes and some explanations of the discrepancies between them, see Cha. pter 6 of Lipsey's book. 33 of their study was 1938, almost the peak for terms of trade of indus- trialized countries. Extension of these data to 1960 would wipe out all the gains since 1920 for the U. S. and the U. K. and all since 1925 (the first year shown) for Germany. The U. K. terms of trade would remain, however, considerably above the 1913 level.78 While the study by Martin and Thackeray is inconclusive, Kindle- berger believes he has found some evidence supporting a relationship between the terms of trade and the stage of a country's economic development.79 His study showed that among the European countries the terms of trade of the more developed countries improved the most. Also, by computing the terms of trade of industrial Europe80 and in- verting them, he found that the U. S. had the most favorable terms of trade, and a group consisting of all other countries, mainly under- developed, the worst.81 Statistically, Lipsey has taken issue with Kindleberger on the relationship between the terms of trade and a country's stage of economic development. According to Lipsey, the U. 3. terms of trade (did not change substantially from the 1880's to the 1950's. Also, While the terms of trade of industrial Europe improved, almost all of ‘tJue increase disappears if the data are adjusted so as to exclude the ‘ 78Lipsey, ibid., p. 13. 79Kindleberger, The Terms of Trade: AAEuropean Case Stpdy, pp. .522;E., p. 239. Also, Kindleberger, "The Terms of Trade and Economic Development," 220 £33." p. fl. 80Industrial Europe is defined as the United Kingdom, Germany, FIE'Elnce, Italy, the Netherlands, the Belgian-Luxembourg Economic Urlion, Sweden, and Switzerland. glKindleberger, "The Terms of Trade and Economic Development," Q- «Eli-Lt." pa 810 34 United Kingdom.82 Moreover, if the U. K. is excluded, the U. S. terms of trade show a decline relative to this group where, accord- ing to Kindleberger, the terms of trade of the U. S. should have improved.83 Finally, we have already noted that an improvement in ‘ the terms of trade of one country or group of countries is insuffi- cient evidence to conclude with certainty that the terms of trade of the other countries actually deteriorated. In a more recent study, Morgan found that, from 1953 to 1960, the underdeveloped countries show some tendency toward a deteriora- tion and the developed countries little or no change in their terms of trade.8’+ However, if the 1937-1959 period is considered, the underdeveloped countries show marked improvement in their terms of Morgan uses 1937 as a base, probably because of a lack of trade. data for any other year, to compare the terms of trade with the post- 'war period; unfortunately, 1937 was a year in which the United States ,and industrial EurOpe showed extremely favorable terms of trade.85 11y inference, the underdeveloped countries experienced unfavorable terms of trade in that year. Therefore, it's not surprising that he liound that the underdeveloped countries have shown improvement-— (fleespite a deterioration from 1953 to 1960--in their terms of trade ‘vlien compared to 1937. __ 82Lipsey, pp. pi_t., p. 15. 83Tbid., p. 17. 8“Morgan, "Trends in Terms of Trade and Their Repercussions on Pl":‘Lmary Producers," pp. pip” p. 61. 858ee Chart 2. Lipsey, pp.lpr., p. 14. 35 In conclusion, there seems to be little statistical evidence in the earlier studies to indicate the terms of trade of underdevelOped (developed) countries inevitably deteriorate (improve). CHAPTER II POST-WAR TRENDS IN THE COMMODITY AND INCOME TERMS OF TRADE Despite the weak statistical foundation, the controversy over the terms of trade of the developed and underdeveloped countries per- sists as shown in recent proposals to the United Nations to compen- sate underdeveloped countries for declines in their terms of trade. Therefore, it is of interest to examine the terms of trade of the developed and underdevelOped countrieslduring the 1948 to 1964 period. The year 1948 was selected as a base because it appeared to be the first "normal" year following World War II. Obviously, some countries had recovered from the War prior to 1948 and others did not recover until after 1948.2 After inspecting post-war data for ex- loorts as well as export and import prices, it was concluded that 21948 was the first "normal" year for most countries. This was based Ixartly on the fact that until 1948 the exports of many of the coun- txries experiencing war damage increased rapidly, indicating recovery from the war. After 1948, the increase was much less rapid. More- over, prior to 1948, most of the countries experienced wide lUnited Nations, Proceedings . . ., _p. _c_:_ip., Vol. I. 2Germany and Japan presumably suffered extensive damage during World War II and may have been slow in recovering. Since no terms of ‘LIraade data are available for these countries prior to 1950, that year is considered the base for the two countries. 36 37 fluctuations in their export and import prices. After 1948, the fluctuations were less, again presumably indicating return to more normal international economic relations.3 Moreover, if 1949 were selected as the base, the time trends of the aggregate terms of trade indices would not be materially affected}+ At the time the study was undertaken, 1964 was the last year in which data were available for a large number of countries. For some countries, data are not available for the entire period; the exceptions are noted in the tables below. The procedure essentially involves calculating the terms of trade of each country and aggregating them into single indices for 5 the developed and underdeveloped countries. Fluctuations in the commodity terms of trade will be discussed in the next chapter. \\ . The Data (There are many theoretical and statistical problems involved in 3Other studies have used 1948 as a base. See, for example, Mor- gan, "Trends in Terms of Trade, and Their Repercussions on Primary Producers," _p. pip. Moreover, the International Monetary Fund (IMF), in presenting statistics for the post-war period, often uses 1948 as a base. This, of course, is merely suggestive. See, for example, International Monetary Fund, International Financial Statistics, Supplement to 1965166 Issues (Washington, 1965), pp. xvi-xix. “Based on evidence provided by Michael Michaely, the terms of trade of most developed countries appear to have been more favorable in 1946 and 1947 than in 1948 and 1949. Hence, exclusion of 1946 and 1947 is more likely to result in the developed countries showing an improvement in their terms of trade for the period as a whole. The evidence is too fragmentary to arrive at any conclusion in regard to the underdeveloped countries. Michael Michaely, Concentration in International Trade (Amsterdam: North-Holland, 1962), pp. 144-67. 5This approach brings Morgan's study up to date but the indices are not strictly comparable because of the additional data now available. 38 deriving export and import price indices; for this reason, the in- dices are subject to errors which may lead to biased results. The main theoretical issue is, of course, the "index number problem." Since most economists are familiar with the problem, it is not discussed here in detail.6 It may be recalled, however, that different weighting procedures give different results. Price indices based on fixed weights (Laspeyres indices) have an upward bias while price indices based on current weights (Paasche indices) are biased downward. This is generally the case but it is not inevitably so. Broadly speaking, the relation holds whenever demand or consumption tends to fall off for commodities which have risen most in price over the period concerned. Some countries use Laspeyres indices, others use Paasche indices, and still others use the so-called Fisher ideal index which is the geometric average of the Laspeyres and Paasche 7 indices. The same procedures are also used to calculate chained indices.8 Because of different weighting procedures, the terms of trade indices of various countries will not be strictly comparable and the aggregate indices may be biased. Moreover, some of the in- dices are based on local currencies while others are given in United Discussions of the ”index number problem" appear in many eco- nomic theory textbooks. See, for example, George J. Stigler, Tmp Theory of Price (New York: The MacMillan Company, 1952), pp. 87-91. For a discussion of index numbers in'international trade, see R. G. D. Allen, ”Index Numbers of Volume and Price," International Trade Statistics, eds. R. G. D. Allen and J. Edward Ely (New York: John ‘Wiley and Sons, 1953), pp. 186-211. 7Because the terms of trade are calculated by dividing the ex- port price index by the import price index some of the bias may cancel. 8For a discussion of chained indices, see Allen, pp. p13,, pp- 193-94. 39 States dollars. Besides the "index number problem," the reliability of the recorded data on which to base the indices is in itself open to ques- tion. Undoubtedly it varies considerably in quality from country to country. Moreover, in almost every case, not all commodities enter directly into the computation of the index. For example, Venezuela's export price and volume indices are based solely on petroleum exports. Fortunately, petroleum accounted for 93.5 percent of total Venezuelan exports during the period under consideration. However, there are many cases in which coverage is not as complete; it may go as low as fifty percent.9 More complete coverage could easily result in changes in the indices.10 Also, as noted previously, changes in the quality of existing products and the introduction of new products at relatively small beginning year weights tends to impart a systematic bias to the price indices which makes the terms of trade of the underdeveloped coun- ' tries appear less favorable, or more unfavorable, vis-a-vis the developed countries than they really are. No attempt has been made here to improve the data although a number of countries were excluded from the sample for the reasons discussed above. Construction of an index for even one country 9A discussion of trade coverage for many countries is found in United Nations, Statistical Office, Supplement to the Monthly Bulle- tin of Statistics, Definitions and Explanatopy Notes (New York, 1964), pp. 115-18. The United Nations has made adjustments for incomplete coverage in some instances. For a discussion of other countries, consult the country pages in International Monetary Fund, Interna- tional Finappgal_§tatistipp (washington, various issues). loHans Staehle, "Some Notes on the Terms of Trade,” cited by Baldwin, _p..pr., p. 267. 40 would be a major undertaking, and we are striving for greater gener- ality than in the case of one or a few countries. Therefore, little significance should be attached to minor movements in a country's indices or minor differences between countries: After the empirical evidence has been presented, we shall return to the shortcomings of the data and the possibility of biased results to see how they would affect the conclusions. The Classification of Countries as Developed or Underdeveloped The commodity terms of trade of sixty-two countries, based on export and import price indices found in various issues of the United Nations' Yearbook of International Trade Statistics and Monthly Bul- . letin of Statistics, are presented in Appendix I.11 In 1958, these countries accounted for approximately eighty-eight percent of world exports and eighty-seven percent of world imports with the world totals excluding Cuba and the Soviet bloc. To test the hypothesis, the countries are divided into two groups--developed and underdevel- oped--according to per capita income. In this context, the terms "developed” and "underdeveloped" are restricted to economic status and imply nothing about a country's cultural heritage; they simply refer to "rich" and "poor”. Similarly, they take no account of a country's "capacity" to develop. The division is arbitrary for a number of reasons. First, per 11The terms of trade for each country were derived by dividing the country's export price index by its import price index and mul- tiplying by 100. In each case, 1958 is used as the base for compari- son (1958 = 100). 41 capita income may be an unsatisfactory index of economic develop- ment.12 Mereover, some economists insist that the distribution of- income should be considered since the vast majority of a country's population may be living in abject poverty even if average per capita income is relatively high.13 Even if per capita income is accepted as an index of economic development, the data may be so inadequate as to render international comparisons meaningless},+ Moreover, the data consist, in most cases, of estimates of per capita gross domes- tic product rather than per capita income and thus include allowances for depreciation. It could be argued that the latter is the more relevant concept; unfortunately, the only estimates available for a large number of countries relate to the former. Admittedly, the per capita income concept is an imperfect cri- terion for dividing countries into developed and underdeveloped groups. However, Irma Adelman and Cynthia Taft Morris have shown 12"National income statistics do not include all of the flows of goods and services in a community. They exclude barter transactions and much of the economic activity represented by home-produced, home- consumed output, and they do not take into account the domestic serv- ices of housewives, the services of consumer durables, or the serv- ices of social overhead capital. In addition, national income comparisons of this nature cannot reflect adequately any of the non- material contributions of the society to the welfare of its people." Also, the data may conceal large differences in the composition of output. Irma Adelman, Tpeories of Economic Growth and Development (Stanford, California: Stanford University Press, 1961 , p. 2. See also, Simon Kuznets, Six Lectures on Economic Growth (New York: The Free Press of Glencoe, Inc., 1959), pp. 13-19 and Harvey Leibenstein, Economic Backwardness and Economic Growth (New York: John Wiley and Sons, 1957 , Ch. 2. 13Jacob Viner, Tnternatippal Trade and Economic Develppment (Oxford: The Clarendon Press, 1953), pp. 125ff. lnThis is due essentially to inadequacy of the raw data, inter- nal inconsistencies in the national income accounts, international differences in national income accounting concepts and procedures, and distortions created by the use of exchange rates. 42 that various social and political variables generally associated with different stages of the deve10pment process (i.e., extent of literacy) are highly correlated with per capita gross national product.15 Hence, the variable may serve as a substitute for a more comprehen- sive measure. While per capita income data are, at best, rough approximations, the reader should keep in mind that we are interested in dividing the countries into two groups--not estimating small differences in income between countries--and, for this purpose, the data seem adeouate. Moreover, as discussed below, the same results are obtained if the countries are classified by real per capita consumption in 1960. Since per capita income changes over time, it is necessary to stipulate the year upon which the groupings are based. A country might be considered underdeveloped (developed) relative to other countries in one year of the period under consideration but developed (underdevelOped) during another year. For purposes of this study, 1963 was selected as the base. Acceptance of some other base would affect only a few countries and change the results only slightly, if at all.17 lsIrma Adelman and Cynthia Taft Morris, "A Factor Analysis of the Interrelationship between Social and Political Variables and Per Capita Gross National Product," Quarterly Journal_pf Economics, LXXIX (November, 1965), 555-78. 16Both Adelman and Simon Kuznets conclude that the data are useful. Adelman, Theories of Economic Growth . . ., o . pip., p. 3, and Kuznets, pp. pip., pp. 18-19. 17For example, if an earlier base were accepted, Israel might be considered underdeveloped. However, trial manipulations of the data suggest that the results would not be altered significantly if Israel were classified as underdevelOped rather than developed. 43 The developed and underdeveloped countries are listed in Tables 1 and 2, ranked by per capita gross domestic product. With one exception (Venezuela), countries with per capita gross domestic prod- ‘uct equal. to or greater than $589 are considered developed; those with less than $589, underdeveloped. If Japan ($589) is excluded from the sample, a considerable gap exists in per capita gross domes- tic product between Ireland ($675) and Cyprus ($547). Ireland is customarily classified as a developed country while Cyprus and Argen— tina ($544) are traditionally regarded as underdeveloped. ‘While Japan's per capita gross domestic product is closer to that of Cyprus, she is classified as a developed country, partly because of 19 customl8 and partly for other reasons. However, aggregate (unweighted) terms of trade indices were calculated on the assump- tion that Japan was an underdeveloped country and comparison of the indices show no significant differences. Similarly, as will be men- tioned below, other changes in classification do not appreciably alter the results. Despite Venezuela's high per capita gross domestic product ($848), she is classified as an underdeveloped country primarily because of her uneven distribution of income. The decision can also be explained by a decline in her per capita gross domestic product of some thirteen per cent from 1958 to 1963 and the fact that she has 18See, for example, the International Monetary Fund's classifi- cation. International Monetary Fund, International Financial Statis- ticsI Supplement to l965/66Tpsues,.pp..pmp., p. xiv. 19Irma Adelman has suggested that countries be classified as developed or underdeveloped on the basis of growth rates. On this basis, Japan--with a relatively high growth rate--would be considered developed. Adelman, Theories of Economic Growth . . ., _p.'pr., p. 3. TABLE 1.-—The developed countries, ranked by per capita gross domestic product, 1963a Per Capita Gross Domestic Country Product at Factor Cost, 1963, in United States Dollars United States $2,790 Canada . 1,871 SWitzerland ~ 1,839 Sweden . 1,802 New Zealand 1,617 Australia - 1,533 Luxembourgb. 1,498 Denmark 1,486 Iceland 1,473 Germany, West 1,416 France 1,406 Norway 1,398 United Kingdom , 1,361 Belgiumb 1,318 Finland 1,153 Netherlands 1,080 Israel 961 Austria 928 Italy 776 Ireland 675 Japan 589 aCalculated from: United Nations, Department of Eco- nomics and Social Affairs, Statistical Office, Yearbook of National AccountspEtatistics, 1964 (New York, 1965), pp. 383-86. bThe statistics for Belgium and Luxembourg are given separately here; however, for the rest of the study Belgium and Luxembourg are treated as one country. 45 TABLE 2.--The underdeveloped countries, ranked by per capita gross domestic product, 1963a Per Capita Gross Domestic Country Product at Factor Cost, 1963, in United States Dollars Venezuela $848 Cyprus 547 Argentina 544b South Africa 477 Chile 457b Greece 440 Jamaica 431° Panama 410° Spain 401° Costa Rica 339° Portugal 304 Colombia 298° Mauritius 281 Guatemala 268 Nicaragua 257d Peru 247 El Salvador 245 Malaya 243° Turkey 230 Ghana 209 Honduras 201° Ecuador 182 Iran 169d Brazil 156b Morocco 150° China: Taiwan 146 Rhodesia and Malawi 139 Ceylon 131 Philippines 127b Thailand 101 Sudan 92° Nigeria @9- Cameroon 85d 46 TABLE 2--gontinued W ‘Per Capita Gross Domestic Country Product at Factor Cost, 1963, win United States Dollars Togo $ 84d Kenya 82 Pakistan 77c India . 76° Indonesia 69d Uganda 69 Tanzania 66 Ethiopia 40° Yugoslavia not available 8'Calculated from: United Nations, Department of Eco-’ nomic and Social Affairs, Statistical Office, Yearbook of National Accounts Statistics, 1964 (New York, 1965), pp- 383-92. bDerived by the use of calculated parity rates of exchange rather than official exchange rates. Ibid., pp- 387-92. cEstimate for 1962. dEstimate for 1958. 47 20 been traditionally regarded as underdeveloped. After completing the categorizations, the results were compared to those obtained by the ranking of countries according to estimated levels of real per capita consumption in 1960.21 While the new rank— ing contains no obvious breaks, it should be noted that the twenty countries classified as developed above rank as the first twenty countries in terms of new estimates.22 Venezuela, although ranked 24th, falls considerably below the developed countries on the scale. Yugoslavia, categorized as underdeveloped above despite the lack of per capita gross domestic product data for that country, ranks 27th. Hence, comparison of the two rankings appears to confirm the earlier results. The Commodity Terms of Trade of the Developed Countries The commodity terms of trade of the twenty nations classified as developed can be aggregated into a single, unweighted index using 23 either the mean (arithmetic average) or the median. Since the two 20If Irma Adelman's classification system were used, Venezuela would presumably be categorized as underdeveloped. It should be noted that the division of the countries into two groups is inher- ently arbitrary because the distribution of countries according to per capita gross domestic product is a continuum rather than two distinct groupings. Nevertheless, the current decision seems to conform with custom. 21The data are found in Appendix II. They were compiled by use of various non-monetary indicators. Wilfred Beckerman and Robert Bacon, "International Comparisons of Income Levels: A Suggested New Measure," Economic Journal, LXXVI (September, 1966), 519-36. 22The order, however, differs but this may be due to the differ- ence in years under consideration. The Spearman rank correlation co— efficient for the forty-eight countries common to both series is .97. 23The shares of world eXports and imports of the countries as a group are found in Tables 3 and 4. 48 methods give somewhat different results, both have been used. It is also possible to aggregate each country's terms of trade index into a single index using the mode; however, it is inappropriate because of the distribution. As an additional test, a weighted index has been compiled.24 The aggregate indices of the commodity terms of trade of the developed countries are presented in Tables 3 and 4 and Figure l. Movements in all three indices are closely related. However, the unweighted indices reach a low in 1957 while the weighted index reaches a low in 1951. The difference is due to the fact that move- ments in the terms of trade of the United Kingdom and the United States largely dominate movements in the weighted index and both countries reach a low for the 1948-1964 period in 1951. The indices show sharply declining terms of trade following 1948 and subsequent recovery. However, the terms of trade in each case failed to return to the 1948 level by 1964. From the Figure, it is clear that little or no trend in the terms of trade exists for the period as a whole regardless of the measure employed.25 This was confirmed by fitting regression lines to the data. In each case, the regression coefficient was not 24 The export and import price indexes were aggregated into sin-' gle indices using each country's share of exports and imports, respectively, as weights. The aggregate export price index was then divided by the aggregate import price index and multiplied by 100 to arrive at the aggregate commodity terms of trade index. 25Moreover, various changes in the classification of the coun- tries as developed do not seem to change the results appreciably. For example, approximately the same results emerge if Iceland, Israel and Japan are omitted from the sample individually, in pairs, or all together. 49 .oo-n uow>om emu pom mono mops-exp HmuOU pHuo3 ech o.moH H.moH m.m~ m.em ma eoaa H.eoH m.moH ¢.~n H.¢n om moaa m.moH H.moH m.~m a.mn om Noma w.~oH o.~oH e.Hn H.¢N om Hom- o.NoH w.HoH m.Hm N.m~ ON coma o.~oH «.moH m.mo m.Hm om mmma o.ooH o.ooH m.wo m.am om wmma m.mm o.wo w.mo H.~n om smo- ¢.nm o.om m.mo w.o~ ON omma m.mm ~.HoH H.mo ~.mo om mmma o.ooH m.HoH H.wo m.wo om emo- m.wo ¢.HOH H.wo m.mo om mmoa m.wm o.¢m m.oo ~.ao 0N mne- H.mm H.HOH «.50 ¢.po om ama- m.~oH w.~oH m.mo ¢.¢o ON omaa ~.moH w.oo~ N.wm e.¢o pH meaa H.moH n.0HH p.90 o.¢© 0H meaa AcmemZV opmuH mo AomoZV opmue mo Auooo uoav Auooo uoav mowuucooo mEuoH mumpoEEoo mauoe zuwpoESoo mmuHOQEH pauoz mmuHOme pauo3 mo meow pounwwo3so pounwwo3c3 mo oumnm mo mumpm wonEDZ comanwema .mofluucsoo poaoHo>op ecu mo opmuu we mauou zuwpoesoo Apounwwo3osv eumwouwwm m:H--.m mqmom emu new menu mopoauxo amuOu pHuo3 onHm 50 m.moH m.¢oH o.mm o.m~ ON oom- m.¢oH m.NoH m.~m H.qm om moma m.eoH m.HoH m.- @.Mm om Nom- H.moH w.HoH ¢.Hn H.eN oN Head m.NoH w.HoH m.- ~.mm om coma m.ooH m.mm w.m© m.Hm om mmma o.ooH o.ooH n.mo N.Hm ON wmma H.o¢ m.HoH w.wo H.Nm ON mmma m.mm m.wa n.mo w.o~ om omma 0.0m p.mm H.oo N.mo ON mmma m.o¢ m.m¢ H.wo n.wp om emma w.~¢ m.nm H.wo m.mo om mmma a.mm m.ooH m.oo N.mo 0N mama o.~m w.NoH «.mp ¢.op om Hmma o.mm m.mw w.no ¢.do ON ommH H.¢OH H.mm N.wm ¢.¢o 0H meaa «.moH w.¢m H.mm a.nm ma wema oomph mo mEHeH xoch mowum xopoH mowum Aucoo ummv Aucoo uoav mowuuosoo zufipoeaou uuomEH uuomxm mmuuoaBH mayo: emuuoaxm pauoz mo meow pounmwoz oumwoumm< oumwouww< we wanna mo oumnw umnaoz eomaquaa .mowuuosoo poaoHe>op emu mo opmuu mo mauou muwpoaaoo Amounwfio3v oumwouwwm onHuu.¢ mqmom ecu pew menu mopSHoxm Hmuou pauozm 60 o.ma ¢.¢m H.~H o.HH «N coma n.wm o.wm ~.mH m.¢H mm moma w.oa o.mm n.0H m.mH He mama m.mm N.om o.~H a.ma we Homa o.mm ¢.wm m.oH m.oH we coda m.wm m.mm m.oH m.mH Ne emma o.ooH o.ooH n.wH H.NH Ne wmma m.¢oH n.¢oH m.mH ¢.~H Ne nmma m.ooH o.¢oH H.ma $.NH Ne cmma n.0HH m.HHH m.wH m.wa Nd mnefl m.¢HH m.¢HH m.mH m.mH N¢ qmda H.00H H.moH m.wH n.wH an mmwa ¢.¢oH m.moH H.¢H c.5H mm Nmma w.moH w.HHH mesa o.wH Hm Hmaa ¢.¢0H ¢.on m.¢H n.5H mm oan n.0m o.mw ¢.oH H.m Hm memfi o.wm H.ww m.oH m.a ma weefl Acmfipmzv memufi mo AcwoZV mpeuh mo Aucmo “adv Aucoo umav moauucsoo mauoH huwpoafiou deuce huwpoaaoo mmuuomaH pHuo3 mmuuoaxm pauoz no new» emunwwmsaa pounwwozcb mo mumnm mo oumnm Honasz i' ill 41" ewaaamqma .mmfiuucsoo pwaoao>oeumpcs ecu mo oemuu mo magma hufipoasoo Ammunwwm3csv oumwouwwm «59:-.m mAmom onu paw menu mopsaoxm Heuou vauo3m m.ma 0 «OH w.~oH m.NH o.~a mm eoea n.5m w.NoH N.ooH “.ma m.¢H mm mama H.0m m.ooH m.om m.oH m.mH Ce mood ¢.o¢ m.NoH w.mm 0.5H $.mH He Hood H.HoH o.ooH o.HoH a.oa m.oa oe coma o.HoH ~.wm w.am ¢.oH m.mH oe emma o.ooH o.ooH o.ooH ~.mH H.5H He mmma N.~cH m.NoH n.¢oH m.wH «.ma He mmoa m.ooH H.¢m m.moH H.ma ¢.NH He onoa H.0HH n.5a H.moH m.wH ¢.wH He mmea w.oHH m.mm o.moH m.m~ m.aH He «mad «.moH o.~oH N.¢oH ¢.wH m.w~ mm mmma n.ma ~.NHH m.~HH o.wH H.5g mm Nnma «.mHH ~.oHH w.n~H m.na o.wH om Hmoa o.ooH H.Hm ¢.na m.¢H n.5H wN omma ¢.ow ~.ooH m.Hw ¢.oH H.m ON mead m.mn o.moH o.¢m ~.oH ~.a ma mead mpmuH mo magma xmch moaum xmch mowum Auaoo uoav Auaoo umav mmwuucnoo %quoaaou uuoaeH uuoaxm mmuuanH pauoz mmuuoaxu mayo: mo meow pounwwm3 oumwouww< mummmuwwd wo eumnm mo oumnm nonasz «omauweaa .mmwuucsoo pomon>mpumpca may mo mvmuu mo menu» huwpoaaoo Amounwwmzv oumweuwwm maaaa.m mqm<fi 124 120 116 112 108 104 . 100 96 ~ 92 . 88 80 76 - 62 T? V V i I \ _ ’ v i a / cg ’ . '1 * ",J AA Unweighted Mean ‘ c BB Unweighted Median * CC 'Weighted Mean F 1998 I g A 19152 I L L 19§6 1 A ‘ lgéo I A 13916.; Fig. 2.--The commodity terms of trade of forty-two underdeveloped countries, 1948-1964: aggregate indices. 63 squares line fitted to the data. If the regression coefficient were significantly different from zero, the terms of trade of the country were said to have improved or declined depending on whether the sign of the coefficient was positive or negative. The results, based on a two—tailed test, are found in Table 9. Of the thirty-one countries for which there are data for most of the 1948-1964 period, seven experienced a decline in their terms of trade with the level of significance equal to .05.33 They are: Argentina, Ecuador, Pakistan, Peru, Philippines, South Africa, and Venezuela. It should be noted that there are no data for the first few years of the period for Argentina, Ecuador, Pakistan, Peru, and Venezuela. To the extent that the terms of trade of each is repre- sented by the aggregate indices, omission of the first few years leads to a downward bias. Four countries (Chile, Greece, Iran, and Mauritius) exhibited improvement while the vast majority showed no trend.3u Moreover, if the average terms of trade of the first two years in each country's terms of trade index are compared with the average of the last two years, only fifteen of the thirty-one coun- tries experienced a decrease.35 Thus, the 1948-1964 period offers little support to those who argue that the commodity terms of trade inevitably decline. If 1954 is considered as the base, twenty-one of the forty-two 33The (negative) regression coefficients of Indonesia, Jamaica, Panama, and Spain become significant at the .10 level. Bqu Iceland and Japan were classified as underdevelOped, the list of underdeve10ped countries showing improved terms of trade would increase to six. 35366 Appendix IV. 64 TABLE 9.--Trends in the commodity terms of trade of thirty-one under- developed countries as measured by the least squares technique: 1948-1964 Regression Trend in Period Coefficient the Com- Country Under (Standard Level of modity Consider- Error in Significancea Terms ation Parenthesis) of Trade Argentina 1951-1964 -1.571 significant declined (.644) Brazil 1948-1964 1.504 not none (.898) significant Cameroon 1949—1962 .364 not none (.723) significant Ceylon 1948—1964 - .504 not none (.446) significant Chile 1948-1964 1.837 significant improved {-645} Colombia 1948-1964 - .798 not none (.880) significant Costa Rica 1948-1964 -1.421 not none (.931) significant Cyprus 1950-1964 .615 not none (.575) significant Ecuador 1950-1963 -4.011 highly declined (.479) significant Ghana 1948-1963 .025 not none (.942) significant Greece 1951—1964 1.328 highly improved (.346) significant Guatemala 1948—1962 - .955 not none (1.486) significant Honduras 1948-1964 - .237 not none (.577) significant India 1950-1964 .153 not none (.459) significant TABLE 9--M¢i_raled 65 —Z— 1 Regression Trend in Period Coefficient the Com- Country Under (Standard Level of modity Consider- Error in Significancea Terms ation Parenthesis) of Trade Indonesia 1950-1961 -2.246 not none (1.153) significant Iran 1948-1964 1.460 significant improved (5314) Jamaica 1948-1963 - .656 not none (.345) significant Mauritius 1948-1963 3.553 highly improved (~353) Significant Nicaragua 1948-1964 - .832 not none (1.012) significant Nigeria 1948-1963 .327 not none (.441) significant Pakistan 1950-1963 -7.961 highly declined (1.908) significant Panama 1948-1964 -1.099 not none (.557) significant Peru 1950-1963 -3.317 highly declined (.652) significant Philippines 1948-1964 -2.009 highly declined (.372) significant Portugal 1948-1964 - .002 not none (.305) significant South Africa 1948-1964 -1.318 highly declined (.355) significant Spain 1948-1964 - .745 not none (.356) significant Togo 1949-1963 - .801 not none (1.051) significant TABLE 9--Continued 66 RegresSion Trend in Period Coefficient the Com- Country Under (Standard Level of modity Consider- Error in Significancea Terms ation Parenthesis) of Trade Turkey 1950-1964 - .355 not none (.705) significant Venezuela 1950—1964 -1.170 significant declined (#89) Yugoslavia 1948—1963 - .192 not none (.360) significant aBased on a two-tailed test. The convention followed is to state the result significant if the null hypothesis is rejected at the .05 level and highly significant if it is rejected at the .01 Dixon and Massey, 2p, git., p. 91. level. 67 countries show a decline in their terms of trade while seventeen 5 no trend.36 The remaining four exhibited improving terms of trade.57fl Hence, the results obtained by the least squares method for this period appear to support the Kindleberger hypothesis. However, the results may be misleading. (féuring the first years of the Korean war, the terms of trade of many of the underdeveloped countries improved sharply; the rest of the period was dominated by a decline from the peaks reached during the war. Thus, the indicated trends may reflect short-run adjust- ments rather than long-run trends. This is suggested by the results obtained for the 1948-1964 period and the apparent stabilization of the terms of trade of the underdevelOped countries in recent years. The indices compiled here show improvement in 1963 and 1964 and the aggregate index compiled by the United Nations for the underdeveloped countries show no change from 1963 through 1966.38 The Commodity Terms of Trade and Stgges of Economic Development According to Kindleberger, movements in the commodity terms of 36See Table 10. The countries which exhibited declining commod- ity terms of trade are: Brazil, Cameroon, Ceylon, Colombia, Costa Rica, Ecuador, El Salvador, Ethiopia, Ghana, Guatemala, Jamaica, Kenya, Nicaragua, Nigeria, Panama, Peru, Philippines, Rhodesia, South Africa, Uganda, and Venezuela. The regression coefficients of the Sudan and India become significant at the .10 level. As a conse- quence, the Sudan should be added to the list of countries with de- clining commodity terms of trade and India to the list with improving terms of trade. 37 They are: Greece, Mauritius, Thailand, and Yugoslavia. 38As discussed above, the indices shown here are not strictly comparable to those of the United Nations but the latter do provide some evidence on recent changes in the commodity terms of trade. United Nations, Monthly Bulletin of_§tatistics, 22. git. 68 TABLE lO.--Trends in the commodity terms of trade of forty-two under- developed countries as measured by the least squares technique: 1954-1964 Regression Trend in Period Coefficient the Com- Country Under (Standard Level of modity Consider- Error in Significancea Terms ation Parenthesis) of Trade Argentina 1954—1964 - .519 not none (.774) significant Brazil 1954—1964 —3.151 highly declined (.674) significant Cameroon 1954-1962 -2.727 significant declined (1.003) Ceylon 1954-1964 -1.720 significant declined (.43) Chile 1954-1964 .239 not none (1.234) significant China: Taiwan 1954—1964 .331 not none (.893) significant Colombia 1954-1964 -5.216 highly declined (.959) significant Costa Rica 1954-1964 —6.087 highly declined ~ (.931) significant Cyprus 1954-1964 — .917 not none (.848) significant Ecuador 1954-1963 -5.710 highly declined (~544) significant El Salvador 1954-1964 -7.048 highly declined (.805) significant Ethiopia 1954-1962 -4.648 highly declined (1.069) significant Ghana 1954-1963 -4.957 highly declined (1.364) significant 69 TABLE 10--Continued fi—g- . hflw- .fl. _-..—_ .__ ..- - — -«W... _..._._. u-..“ u-HV.—._..—.-----_——.-..—«_.- ----._.-- Regression Trend in Period Coefficient the Com- Country Under (Standard Level of modity Consider- Error in Significancea Terms ation Parenthesis) of Trade Greece 1954-1964 1.411 significant improved (~51?) Guatemala 1954-1962 —9.379 highly declined (1.470) significant Honduras 1954-1964 -l.291 not none (1.216) significant India 1954-1964 .933 not none (.457) significant Indonesia 1954—1961 -1.212 not none (1.600) significant Iran 1954-1964 .416 not none (.920) significant Jamaica 1954-1963 -2.085 significant declined (-629) Kenya 1954-1963 -2.974 highly declined (.652) significant Malaya 1954-1964 - .290 not none . (1.398) significant Mauritius 1954-1963 3.359 highly improved (.818) ,significant Morocco 1954-1964 - .234 not none (.389) significant Nicaragua 1954-1964 -5.367 highly declined (1.111) significant Nigeria 1954-1963 -l.780 significant declined C645) TABLE 10--Continued 70 m- —.————. _________,___.._____. Regression Trend in Period Coefficient the Com- Country Under (Standard Level of modity Consider- Error in Significancea Terms ation Parenthesis) of Trade Pakistan 1954—1963 -2.315 not none (2.164) significant Panama 1954-1964 -3.314 highly declined (.840) significant Peru 1954-1963 -2.684 significant declined (.826) Philippines 1954-1964 -l.598 highly declined (.396) significant Portugal 1954-1964 - .172 not none (.452) significant Rhodesia 1954-1963 -5.138 significant declined (1.613) South Africa 1954-1964 —l.l75 highly declined (.284) significant Spain 1954-1964 .314 not none (.417) significant Sudan 1954—1963 -l.500 not none (.724) significant Tanzania 1954-1963 - .271 not none (.736) significant Thailand 1954—1964 2.317 highly improved (.439) significant Togo 1954—1963 -3.219 not none (1.786) significant Turkey 1954-1964 - .060 not none (1.274) significant TABLE 10--Continued 71 Regression Trend in Period Coefficient the Com- Country Under (Standard Level of modity Consider- Error in Significancea Terms ation Parenthesis) of Trade Uganda 1954—1963 -5.323 highly declined (.634) significant Venezuela 1954-1964 -2.965 highly declined (.446) significant Yugoslavia 1954-1963 .721 highly improved (.201) significant 8‘Based on a two-tailed test. The convention followed is to state the result significant if the null hypothesis is rejected at the .05 level and highly significant if it is rejected at the .01 Dixon and Massey, gp,lgit., p. 91. level. 72 trade are directly related to the stage of a country's development. Under the hypothesis, the higher the stage of a country's develop- ment, the more favorable its terms of trade should be over time. As noted above, if the countries for which data are available for most of the 1948-1964 period are divided into two groups--developed and underdeveloped--there is little or no evidence to support the hypoth— esis. This test, however, allows for only two stages of development; hence, a more general test is desirable. To test the hypothesis, the countries which displayed statis- tically significant trends in their terms of trade are ranked by their regression coefficients (in descending order). If the coun- tries are also ranked by per capita gross domestic product (again in descending order), it is possible to compute the rank correlation coefficient between the two series. If the Kindleberger hypothesis is true, there should be a significant, positive rank ordering.39 If the eighteen countries which displayed statistically signifi- cant trends in their terms of trade for the 1948-1964 period are ranked by the size of their regression coefficients and by per capita gross domestic product, the Spearman rank correlation is .260 which is not significant at the .05 level.40 Thus, the result fails to 39For a discussion of the test, see Sidney Siegel, Nonparametric Statistics (New Yerk: MbGrawsHill, 1956), pp. 202-13. 401m this section, the tests are all one-tailed tests unless otherwise stated. The regression coefficients of the countries were significant at the .05 level using a two-tailed test. However, if the level of significance is .10, four additional countries (Indo- nesia, Jamaica, Panama, and Spain) are added to the sample. In this case, the Spearman rank correlation coefficient is .374 which is significant at the .05 level. This result is discussed below. 73 confirm Kindleberger's hypothesis.41 A similar result is obtained by ranking (in descending order) the countries by percentage change in_ their terms of trade from 1948-1949 to 1963—1964 and by per capita gross domestic product.42 For the fifty countries for which data are available, the Spearman rank correlation coefficient between the two series is .103 which is not significant at the .05 level. In both cases, however, the result is based on the per capita gross domestic product data presented above and the data consist of estimates not only for 1963 but also for 1958 and 1962.43 Because of possible inconsistencies in the rankings, alternative data must be considered which refer to a common time period. If 1958 estimates of per capita gross domestic product are used, the Spearman rank correlation coefficient between this series and a ranking of eighteen countries by their statistically significant regression coefficients for the 1948-1964 period is .193 which is not f 41If only the 1954-1964 period is considered, the results con- firm the Kindleberger hypothesis. If the countries are ranked by the regression coefficients which are significant at the .05 level using a two-tailed test, the Spearman rank correlation coefficient, based on a sample of thirty-five countries, is .506 which is significant at the .01 level. If the countries are ranked by the regression coeffi- cients which are significant at the .10 level, the Spearman rank cor- relation coefficient is .430 which is also significant at the .01 level. The latter result is based on a sample of thirty—seven countries. qutrictly speaking, the percentage change is from the initial two years of each country's terms of trade index to the last two years since data are not available for the entire period for a number of countries. See Appendices III and IV. uBMoreover, some of the estimates are calculated by the use of parity rates of exchange rather than official exchange rates. United Nations, learbook of National Accounts:§tatistics,1964, gp.‘git., pp. 387—92. The data were used to divide the countries into two groups and, because of.the gap between per capita gross domestic products of the groups, seemed adequate for that purpose. 74 significant at the .05 1eve1.4u Moreover, if the rankings are by 1958 per capita gross domestic product and by percentage changes in terms of trade from 1948-1949 to 1963-1964, the rank correlation coefficient is only .067 which is not significant at the .05 level. The latter is based on a sample of fifty countries. In both cases, the result fails to confirm the hypothesized relationship between the terms of trade and the stage of a country's development. As discussed earlier, Wilfred Beckerman and Robert Bacon pro- vide estimates of per capita consumption in 1960 for eighty coun- tries.“5 These estimates provide an alternative method of ranking the countries in the sample and also have the advantage that the data refer to a common period.”6 If the countries which displayed statistically significant trends for the 1948-1964 period are ranked by their regression coef— ficients and by per capita consumption in 1960, the Spearman rank correlation coefficient, based on a sample of seventeen countries, is .125 which is not significant at the .05 level.“7 Hence, the test 44 This is based on a sample of countries whose coefficients were significant at the .05 level using a two-tailed test. If the coun- tries whose coefficients were significant at the .10 level are con- sidered, the resulting rank correlation coefficient for the twenty- two countries is .282 which is also insignificant at the .05 level. asBeckerman and Bacon, gp.,git., p. 533. uéHowever, use of the per capita consumption data reduces the size of the sample. For example, there are no per capita consumption data for Jamaica, Panama, or the Philippines. “7This is based on the countries whose regression coefficients were significant at the .05 level using a two-tailed test. If the .10 level is considered, the sample is increased to nineteen and the resulting rank correlation coefficient is .281. Despite the increase, the new rank correlation coefficient is not significant at the .05 level. . 75 fails to confirm the Kindleberger hypothesis“,8 The same result is also obtained if the countries are ranked by the percentage change in their terms of trade from 1948-1949 to 1963-1964 and per capita consumption in 1960 .“9 Regardless of the test, the results for the 1948-1964 period all show positive--but non-significant--rank correlations.50 On this basis, the Kindleberger hypothesis is not confirmed by the evidence presented here.51 As discussed above, the shortcomings of the export and import price indices used to compile the terms of trade indices in this study may lead to biased results. For example, it would appear that the quality of manufactured goods improves more rapidly than the quality of primary products and that this difference is not fully n8If, however, the countries which diSplayed statistically sig- nificant (at the .05 level with a two-tailed test) trends in their terms of trade for the 1954-1964 period are ranked by the size of their regression coefficients and by per capita consumption in 1960, the Spearman rank correlation coefficient is .421. This coefficient is significant at the .05 level based on a sample of twenty—six countries. If the same procedure is followed with the twenty-eight countries which displayed statistically significant regression coef- ficients at the .10 level, the resulting rank correlation coefficient is .373 which is also significant at the .05 level. ugThe Spearman rank correlation coefficient, based on a sample of forty-one countries, is .373 which is also insignificant at the .05 level. 50Except for the case noted in footnote 40 which is based on per capita gross domestic product data which include estimates for sev- eral different years. If the countries are ranked by data which refer to a common time period--per capita gross domestic product in 1958 or per capita consumption in l960--the rank correlations are clearly non-significant. The latter appears to be the only reasonable approach. 510n the other hand, if only the 1954-1964 period is considered, the coefficients are all positive and significant. Hence, the evi- dence for the 1954-1964 period supports the Kindleberger hypothesis. As discussed earlier, selection of the 1954-1964 period may lead to 76 reflected in the price indices. Hence, to the extent that under- developed countries export primary products and developed countries export manufactured goods, the terms of trade of the underdeveloped countries appear more unfavorable vis4a-vis the developed countries than they really are. Similarly, one would expect hundreds of new manufactured products to appear over the years while the number of primary products remains about the same. If this is the case, the terms of trade of the underdeveloped countries would, once again, appear more unfavorable visJa-vis the developed countries than they really are since the omission of new commodities or their inclusion at relatively small beginning year weights tends to bias a price index upward as new commodities usually fall in price soon after they are introduced. Because these biases make the terms of trade of the underdeveloped countries appear more unfavorable vis-a-vis the developed country than they really are, they strengthen the argument presented here. In regard to the other shortcomings, the evidence is not clear. For example, it is uncertain whether incomplete coverage makes the terms of trade of the underdeveloped countries appear more favorable vis-a-vis the developed countries or less favorable than they really are. As a consequence, the possibility remains that there is a sys- tematic bias which makes the terms of trade of the underdevelOped countries appear more favorable vis-a-vis the developed countries than they really are. However, the burden of proof is clearly with those who argue that there has been a secular deterioration of the terms of trade of the underdeveloped countries. biased results; hence, the hypothesis is taken as unproven. 77 Comparison with the Pre4War Period Even if the 1948—1964 period is considered, it must be empha- sized that the results cover only a seventeen year span. All earlier studies have been characterized by wide fluctuations in the terms of trade; hence, the results for the post-war period may be atypical when considering secular trends in the commodity terms of trade. Therefore, it is desirable to compare the commodity terms of trade of the post—war period with those of the pre—war period. By using Morgan's data, it is possible to compare the post-war commodity terms of trade of twenty-six countries with their terms of trade in 1937.52 The percentage changes in each country's terms of trade from 1937 to 1963—1964 were compiled and are presented in Tables 11 and 12. It was found that in ten of the fifteen countries classified as develOped the terms of trade improved from 1937 to 1936-1964. Similarly, when the group mean and median for the devel- oped countries are considered, both show an increase. Thus, it would appear that the results lend support to the Kindleberger hypothesis. However, this is not the case when the underdeveloped countries are considered. 0f the eleven underdeveloped countries, seven experi- enced an increase which is approximately the same percentage as for the developed countries. Moreover, when the underdeveloped coun- tries' group mean and median are considered, both show a greater in— crease than those of the developed countries. Hence, on this basis, 52Morgan, "Trends in Terms of Trade, . . .," 22. git., pp. 74—75. It must be recalled, however, that the terms of trade of the devel- oped countries were relatively favorable vis—a-vis the underdevel— oped countries in 1937. This imparts a bias to the comparison. 78 TABLE ll.--Percentage change in the commodity terms of trade of fifteen developed countries: 1937 to 1963-1964 Commodity Terms of Trade Percentage Country 1937 1963-1964 Change Ireland 81.3 105.0 29.2 Japan 87.78 99.5 13.5 west Germany 99.0b 109.0 10.1 Switzerland 99.0° 107.0 8.1 New Zealand 116.3 125.5 7.9 Belgium . 91.7 98.5 7.4 Norway 94.3 101.0 7.1 Sweden 94.3° 98.5 4.5 France 101.0c 104.5 3.5 United Kingdom 103.1 104.0 .9 Canada 102.0 96.0 — 5.9 Denmark 117.6c 104.5 -11.1 Australia 137.0 - 116.5 -15.0 Italy 122 .0 103 .0 -15 .6 United States 123.5 104.0 -15.8 Mean 104.7 105.1 .4 Median 101.0 104.0 3.0 a1934-1936 b1936 c1938 79 TABLE 12.--Percentage change in the commodity terms of trade of eleven underdeveloped countries: 1937 to 1963—1964 Commodity Terms of Trade Percentage Country 1937 1963—1964 Change El Salvador 43.3 75.0 73-2 Colombia 50.0 86.5 73.0 Sudan 68.5a 101.5b 48.2 Costa Rica 75.8 84.0‘ 10.8 Malaya 88.5a 98.0 10.7 Ceylon 85.58 88.0 2.9 India 105.3 108.0 2.6 Honduras 117.6 114.5 - 2.6 Venezuela 74.6 72.0 - 3.5 Philippines 104.2 90.0 -13.6 Turkey 147.1 124.5 -15.4 Mean 87.3 94.7 8.5 Median 85.5 90.0 5.3 F1938 b1962.1963 80 the terms of trade of the underdeveloped countries appear to improve more than those of the developed countries. Moreover, of the devel- oped countries, the terms of trade of the United States--the most highly developed country in the world--declined the most, contrary to the Kindleberger hypothesis. As before, it is desirable to allow for more than two stages of development.53 To this end, the countries were ranked (in descending order) by percentage increase in their terms of trade and by per cap- ita gross domestic product in 1963. If the Kindleberger hypothesis is correct, there should be a positive, significant rank ordering. However, the Spearman rank correlation coefficient between the two series is -.211, which is clearly non-significant.54 Thus, compari- son with 1937 fails to confirm the Kindleberger hypothesis. If any- thing, the results show that the terms of trade of the underdeveloped countries in the sample improved more than those of the developed countries. The_;ncome Terms of Trade As a measure of a country's export-based capacity to import in the post-war period, the income terms of trade of fifty-nine coun- tries have been calculated for the 1948-1964 period (See Appendix V 53Especially since there are only eleven countries classified as underdeveloped. 54If the countries are ranked by per capita consumption in 1960, the Spearman rank correlation coefficient, based on a sample of twenty-two countries, is -.352. The result is not significant at the .10 level using a two-tailed test. It is, however, significant at the .20 level. 81 for the results).55 In 1958, these countries accounted for approxi- 56 mately eighty-eight per cent of total world exports. To examine the recent trends, the countries are classified as "developed" or "underdeveloped" using the same scheme discussed earlier in the chapter.57 As in the case of the commodity terms of trade, the developed and underdeveloped countries' individual indices can be aggregated into a single unweighted index using the mean and median. The results of such an aggregation are shown in Tables 13 and 14 and Figures 3 and 4. Similarly, a weighted index has been compiled and is presented in Table 15 and Figure 5.58 The aggregate indices show, unlike those of the commodity terms of trade, clear upward trends for both the developed and underdevel— Oped countries. This is not unexpected because of the overall expan- sion of world trade volume, where the volume index appears only in the numerator of the income terms of trade formula. On an individual basis, the average annual growth rates of each country's capacity to import, as measured by the income terms of trade, are shown in Table 16. In fortyeseven of the forty-eight 55The income terms of trade were compiled by multiplying the commodity terms of trade by the export volume index and dividing by 100. The base is 1958 (1958 = 100). 56The world total excludes Cuba and the Soviet bloc. See Table 15 for shares on a year by year basis. 57Australia, Austria, Belgium-Luxembourg, Canada, Denmark, Fin- land, France, West Germany, Iceland, Ireland, Israel, Italy, Japan, Netherlands, New Zealand, Norway, Sweden, Switzerland, United Kingdom and the United States are classified as developed; the rest are considered underdeveloped. 58To compile the index, the individual country's income terms of trade were weighted by the country's share of group exports. This is the same procedure used by Morgan in "Trends in Terms of Trade,. ..," 212- .93- ' 82 TABLE 13.--The aggregate (unweighted) income terms of trade of the developed countries, 1948-1964 Number of Year Countries Mean Median 1948 16 54.9 50.0 1949 16 61.9 55.4 1950 20 63.4 63.9 1951 20 68.8 68.3 1952 20 66.1 67.4 1953 20 73.7 70.3 1954 20 80.7 78.7 1955 20 85. 85.5 1956 20 92.2 90.7 1957 20 97.7 96-3 1958 20 100.0 100.0 1959 20 113.8 113.4 1960 20 126.6 123.7 1961 20 136.8 134.1 1962 20 147.9 143.8 1963 20 163.2 155.1 1964 . 19 175.8 171.4 83 TABLE 14.—-The aggregate (unweighted) income terms of trade of the underdeveloped countries, 1948-1964 Number of Year Countries Mean Median 1948 18 61.0 58.3 1949 20 60.7 56.1 1950 26 75.5 74.8 1951 28 82.0 81.6 1952 32 77.0 76.9 1953 36 86.2 83.5 1954 39 96 . 6 95 - 6 1955 39 101.6 97.7 1956 39 102.1 99.0 1957 39 101.8 103.8 1958 39 100.0 100.0 1959 39 109.4 106.8 1960 39 111.4 110.0 1961 39 113.3 107.0 1962 38 121.4 114.8 1963 35 134.8 136.6 1964 22 144.8 138.4 170 160 150 140 130 120 110 . 100 90 80 70. 60 50 - 84 AA Developed Countries B BB Underdeveloped Countries 3948‘ .41922- - A1956. 1419.60. ‘41964 L Fig. 3.--The income terms of trade of twenty develOped and thirty-nine underdeveloped countries, 1948-1964: aggregate (mean) indices. 85 170 t A 160 . 150 . 140 . .B 130 L 120 t 110 . 100 . 90 t 80 t 70 . 60 t B AA Developed Countries F BB Underdeveloped Countries 50 L 40 » h 19138;;‘19452‘ - .1956; - _1960‘ _ 11964 Fig. 4.--The income terms of trade of twenty developed and thirty-nine underdeveloped countries, 1948-1964: aggregate (median) indices. 86 TABLE 15.--The aggregate (weighted) income terms of trade of the developed and underdeveloped countries, 1948—1964 Developed Countries Underdeveloped Countries Number Share of ‘Weighted Number Share of Weighted of world Ex- Income of ‘World Ex- Income Coun- portsa Terms of Coun- portsa Terms of Year tries (Per Cent) Trade tries (Per Cent) Trade 1948 15 57.9 68.4 18 8.6 60.9 1949 16 64.4 72.3 19 8.1 62.3 1950 20 64.4 68.0 25 14.7 84.5 1951 20 66.4 66.1 27 15.9 94.9 1952 20 69.2 70.9 30 15.8 78.9 1953 20 69.3 78.8 34 17.0 87.2 1954 20 68.7 79.9 38 18.2 97.5 1955 20 69.2 84.9 38 17.6 102.8 1956 20 70.8 93.8 38 16.7 104.5 1957 20 72.1 99.8 38 15.9 103.8 1958 20 71.7 100.0 38 15.8 100 0 1959 20 71.9 109.5 38 16.0 111.9 1960 20 73.2 123.9 38 15.2 116.9 1961 20 74.1 132.7 38 14.6 117.9 1962 20 73.9 141.8 37 13.9 123.6 1963 20 74.1 153.4 34 13.6 130.8 1964 20 75.0 170.6 23 10.9 137.7 aWorld total excludes Cuba and the Soviet bloc. 170 . 160 L 150 l 140 , B 130 L 120 F 110 r 100 . 90 . 80 . 70 , AA Developed Countries 60 I 3 BB Underdeveloped Countries 1948 1952 1956 1960 1964 A 4; L A L Fig. 5.--The income terms of trade of twenty developed and thirty-nine underdeveloped countries, 1948-1964: aggregate (weighted) indices. 88 TABLE l6.--Increases in capacity to import of forty-eight countries as measured by changes in the income terms of tradea, 1948-1964 Country Per Cent Country Per Cent Israel 16.09 Togo 5.48 Japan 15.68 Spain 5.44 Germany 14.66 Nigeria 5.42 Italy 11.89 Mauritius 5.31 Austria 11.78 Philippines 4.43 Netherlands 9.97 Argentina 4.19 Yugoslavia 8.87 Canada 4.18 Greece 8.85 Venezuela 4.15 France 8.39 Chile 4.10 Denmark 8.19 United Kingdom 3.99 Jamaica 8.19 United States 3.97 Iceland 8.02 Ghana 3.94 Cameroon 7.33 Ecuador 3.67 Switzerland 7.30 Brazil 3.31 Norway 7.22 Guatemala 2.81 Nicaragua 7.13 New Zealand 2.59 Belgium-Luxembourg 6.91 Cyprus 2.46 Finland 6.76 Costa Rica 2.27 Sweden 6.74 India 2.27 Peru 6.31 Australia 1.88 Ireland 6.23 Ceylon 1.55 Portugal 6.10 Colombia .92 South Africa 6.08 Honduras .32 Panama 5.99 Indonesia — .81 8‘The average annual growth rate is calculated by the use of the logarithmic least-square method whereby an exponential function: Z = aebt, t = l, 2, . . . n is fitted to the income terms of trade over time by simple regression analysis applied to the logarithmic form of the equation. In the equation, g represents the income terms of trade, 3 and b constants, 3 time, and §_the total number of years in the series. The average annual rate of change is given by 2*, the regression coefficient. 89 countries for which there are data for almost the entire 1948-1964 period, the growth rate was positive; the only exception was Indo- 59 nesia. The average annual increase for all countries in the sample is 6.01 per cent with average annual increases of 8.12 and 4.50 per cent, respectively, for the developed and underdeve10ped countries. Thus, it appears that, with some exceptions, the "capacity to importfl as measured by the income terms of trade, of the developed countries has increased more rapidly than most of the underdeve10ped countries. Based on a one-tailed test, it was found that the average annual terms of the developed countries was significantly greater than that of the underdeveloped countries. To further test the relationship between percentage changes in export-based capacity to import and stages of economic deve10pment, the countries were ranked by percentage changes in their income terms of trade and by per capita gross domestic product in 1963. In this case, the Spearman rank correlation coefficient between the two series is .381. Based on a sample of forty-seven countries, the coefficient is significant at the .01 level using a one-tailed testffl' 59Indonesia may be a special case for two reasons. First, the period for which data are available, 1950 to 1961, is shorter than that of the other countries. Second, Indonesia's economy has been subject to a number of shocks. 60This was at the .05 level. This may be due to a more rapid expansion of world trade in manufactured products. 61Approximately the same results are obtained if the countries are ranked by percentage changes in their income terms of trade and by per capita consumption in 1960. The rank correlation coefficient between the two series is .348 which is significant at the .05 level using a one-tailed test. This result is based on a sample of thirty- nine countries. 90 Hence, disaggregation supports the result obtained above. In conclusion, it appears the higher the stage of a country's deve10pment, the more favorable its income terms of trade over time. Thus, it might be argued that aid to underdeveloped countries be 63 increased in order to assist in their develOpment programs. Never- theless, such an argument may be misleading for a variety of reasons. First, there is the question of causality. It may be that developed countries are better able to expand their exports (because they are developed) than underdeveloped countries but it is also undoubtedly true that some countries are developed because they have been able to expand their exports. Second, the income terms of trade measure only a country's export-based capacity to import and neglect invisible items and capital flows. Clearly, it is a country's total capacity to import that is relevant; hence, movements in the income terms of trade may be misleading.64 Third, the concept takes no account of a country's need for foreign exchange. For example, consider a country which deliberately undertakes a program of import substitution. If its resources are devoted to this purpose, the country's income terms of trade may increase less rapidly than other countries. However, because of import substitution, the country may have less need for 62The results are consistent with those obtained by Morgan in "Trends in Terms of Trade and their Repercussions on Primary Pro- ducers," _p, git., pp. 65-66. 63There are, of course, many other arguments for increasing aid to underdeveloped countries. 64For evidence on total capacity to import, see J. Marcus Flem- ing and Gertrud Lovasy, ”Fund Policies and Procedures in Relation to the Compensatory Financing of Commodity Fluctuations," International Monetary Fundggtaff Papers (November, 1960), pp. 1-76. 91 foreign aid than countries whose export-based capacity to import have increased more rapidly. Summary Of the seven terms of trade concepts discussed above, only two—- the commodity and income--are both useful and easy to calculate with existing data. Hence, we are left with only the commodity and income terms of trade as possible guides to changes in economic welfare and the income terms of trade as a measure of a country's "capacity to import." While welfare implications are often drawn from movements in the commodity terms of trade, it was noted that the commodity terms of trade are a poor guide to changes in economic welfare. In most cases, it is necessary to examine the causes of the change in the terms of trade in order to assess the impact on economic welfare. In regard to the income terms of trade, this concept is also misleading as a guide to changes in economic welfare and, in certain cases, is even more misleading than the commodity terms of trade. However, because of the present concern with financing economic development, changes in export volume as well as changes in export and import prices are considered. Despite the uncertain relationship between changes in the com- modity terms of trade and changes in economic welfare, economists and politicians have insisted that primary producers and, more recently, underdeveloped countries have suffered a secular deterioration in their commodity terms of trade and some have suggested that the coun- tries should be compensated for the decline. Although the 92 statistical and theoretical foundations of the argument are extremely weak, the controversy has persisted. The main contribution of this chapter is to summarize the evi- dence relating to the argument for the post-war period. On the basis of both the aggregate commodity terms of trade indices of the devel- oped and underdeveloped countries and a consideration of each coun- try's terms of trade, the 1948 to 1964 period offers little evidence to support the hypothesis that the commodity terms of trade of the underdevelOped (developed) countries deteriorate (improve) over time. Similarly, when the develOped-underdeveloped dichotomy is abandoned, there appears to be no significant relationship between the stage of a country's development and movements in its terms of trade. If only the 1954—1964 period is considered, both the aggregate indices and consideration of individual countries appear to support the hypothesis. For example, if regression lines are fitted to the commodity terms of trade as a function of time, many of the developed (underdeveloped) countries show an improvement (a deterioration) in their commodity terms of trade. However, the result may be spurious because the terms of trade of most developed (underdeveloped) coun— tries reached a low (high) during the early 1950's as a result of increased export prices of primary products brought on by the Korean war. Hence, it is uncertain whether the least squares method meas- ured the inevitable deterioration (improvement) of the commodity terms of trade of the underdeveloped (developed) countries or the gradual movement from a period of abnormal highs (lows) brought on by the Korean'War. For this reason, it is argued that the relevant time period is from 1948 (or 1949) to 1964. Since the aggregate 93 indices provided by the United Nations show no change from 1963 through 1966, the argument could presumably be extended through 1966. Even if the 1948-1964 period is considered, it must be empha- sized that the results cover only a seventeen year span. However, when the post—war terms of trade of twenty-six countries were com- pared to their terms of trade in 1937, it was found that the terms of trade of the underdeveloped countries improved more than those of the developed countries. Thus, there appears to be little evidence to indicate that the terms of trade of underdeveloped countries inevi- tably decline vis—a—vis the developed countries. Indeed, the diver— sity of results suggests that each country must be studied individually.65 ‘While there is little evidence to support the hypothesis that the commodity terms of trade of the underdeveloped countries deterio— rate over time, the income terms of trade of the underdeveloped countries appear to have improved less rapidly than those of the developed countries. The average annual percentage increase for the twenty developed countries is 8.12 while the corresponding percentage for the twenty-eight underdeveloped countries is 4.50. It should be emphasized that the income terms of trade are not a measure of "total capacity to import" which would include foreign aid and other items. 5It is possible to divide the countries into sub-groups—-such as the mineral exporting countries--and examine the commodity terms of trade of these groups. However, it is easier to proceed to a consideration of each country. CHAPTER III FLUCTUATIONS IN THE COMMODITY AND INCOME TERMS OF TRADE In the previous chapter, we found that there was little evidence to indicate that the commodity terms of trade of the developed (underdeveloped)countries inevitably improve (decline). On the other hand, we did find that the growth rates of the developed countries' capacity to import, as measured by the income terms of trade, were significantly greater than those of the underdeveloped countries. However, nothing was said of fluctuations in the commodity and income terms of trade. As fluctuations in the commodity and income terms of trade may have adverse effects on the domestic economy, we shall examine the degree of their instability in this chapter and test to see if under— developed countries experience more instability than developed coun- tries.1 If instability is harmful and underdevelOped countries experience more instability, the problem is presumably worse for underdeveloped countries than for developed countries. Measures of Instability Before discussing the various methods of measuring instability, 1The relationship between international economic instability and the domestic economy is discussed in the next chapter. However, it might be noted at this point that instability may lead to fluctua— tions in national income, a misallocation of resources, and balance of payments problems. 94 95 it should be noted that the study is concerned only with year-to-year fluctuations. This is not to deny that fluctuations within-years or over the cycle are not important; indeed, they may be substantial.2 However, it may be argued that within-year fluctuations are not important provided year-to-year fluctuations are mild. Also, it may be that within-year fluctuations, like seasonal variations, are more easily predicted (and provided for) because of the shorter time horizon. Moreover, there is insufficient data on a monthly or quar- terly basis. As for cyclical movements, the post—war period is too short to study fluctuations of this type.3 There are a number of methods available to calculate instability The most common procedure is simply to compute the average year—to- year percentage f1uctuation.n If Il denotes the instability index (average year-to-year fluctuation, per cent), g the variable in ques- tion, 3 the year, and g the total number of years covered in the series, the index can be represented as n z — z 100 z tz t'1 I _ t=2 t-l 1 _ n - 1 Although easy to calculate, this method is deficient in that it makes no allowance for trend factors. If the variable in question 2United Nations, Department of Economic Affairs, Instability in Egport Markets of Under-Developed Countries (New York, 1952), pp. 15-18, 21-23, 32-35, 44-46. 3The United Nations study contains data on cyclical movements in international trade for the 1901-1950 period. Ibid. “The method was used, for example, by Michael Michaely to meas- ure fluctuations in export and import prices and the commodity terms of trade. Michaely, 22. git., pp. 68—70. 96 has risen (or fallen) continuously over time, the index will show strong "fluctuations" even if there have been no fluctuations at all. For example, consider a country whose income terms of trade were 10, 20, 30, and 40 in successive years. Using this method, the index would show an average annual fluctuation of sixty—one per cent while graphically there is only a linear trend with no fluctuations at all. Because of the strong upward trend in the income terms of trade (and other variables) of many countries, use of such an index would seri— ously overstate the degree of instability. However, countries with strong upward trends in their income terms of trade (or other vari- ables) would show greater "instability" than other countries; hence, this method of computing instability is not used.5 The United Nations has suggested another index which is quite similar to the first.6 In it, the year-to-year differences in the variable in question are divided by the larger of the two values, Zt and zt-l rather than always dividing by Zt-l as in the previous index. Symbolically, the instability index (average year-to-year fluctuation, per cent), denoted as I2, can be represented as: n lzt ‘ Zt-li 100 2 ~ t=2 max (Zt’ Zt-l) n - 1 This procedure has the advantage of making a crude correction for trend factors. Using the second approach, the measured average 5Michaely justified the use of this method on the grounds that there is little or no trend in export and import prices and the com- modity terms of trade. Ibid., p. 69. 6For a discussion of the United Nations method, see United Nations, Instability in Export Markets of Underdeveloped Countries, 22. 313,. PP- 77-79- 97 year-to-year "fluctuation" of the income terms of trade in our previ- ous example would be 34.3 per cent.7 Since the method has been widely used, makes some allowance for trend factors, and is easily interpreted, I have used this procedure as well as two others to calculate instability indices for the countries in question. To correct for trend, I have used methods suggested by Benton F. Massell.8 To calculate the instability index, 13, a linear regres- sion line is fitted to the variable in question, 2, eXpressed as a function of time, t, and the residuals, 3, obtained by subtracting the estimated values of the variable in each year from the actual value. In each case, the absolute value of the differences between u and ut 1 are obtained and divided by the larger of 2t and Z t t—l t0 obtain the trend-corrected year-to-year fluctuation, wt, of export earnings. The year-to—year fluctuations are then summed, divided by the n-l observations and multiplied by 100 to obtain the trend— corrected average year-to—year fluctuation of export earnings in per- centage terms. Symbolically, the instability index, I3, can be expressed as: i 100 Z 'w t: |ut ' ut-l‘ I = where w = 3 n - 1 t max(Zt, zt-l) 7For the same set of data, I will always be greater than, or equal to, I . The I index ranges from zero to plus infinity while the I2 index ranges from zero to 100. 8Benton F. Massell, "Export Concentration and Fluctuations in Export Earnings: A Cross-Section Analysis," American Economic Review, LIV (March, 1964), 47-63. 98 This measure assumes, of course, that the trend can most appropri- ately be approximately by a linear function of time.9 Another equally satisfactory measure of instability--called the normalized standard error of the estimate by Massell--is obtained by dividing the standard error of the estimate by the mean of the obser- vations. This measure, like the previous one, is a pure number and is independent of the overall level and rate of growth of the vari- able. It can be written as n 2 u t-_-1t n - 2 '- . I4 - 2’ where Z is the mean of the observations. 'While both measures, 13 and I4’ are obtained by fitting a linear regression line to the variable as a function of time, they are con- ceptually different. The instability index 13 is more a measure of the year-to-year change in a country's export earnings while I4 is largely a measure of the variation of the series as a whole around the trend line. Although both measures are affected by the appro- priateness of fitting a linear trend line to a particular country's export earnings, I4 would be more affected by a poor fit than 13. Because the two measures give somewhat different results, both are subsequently used. However, only I2 and I3 are considered in this chapter since it deals solely with year-to-year fluctuations. Numerous other indices of instability have been suggested. In his study, International Economic Instability, Joseph D. Coppock 9An exponential trend was also fitted to the data, but, in general, the linear trend provided a better fit. 99 10 uses what he calls the log variance method. If V denotes the loga— rithmic variance of the series and m_the arithmetic mean of the dif- ference between the logs of zt-l and 2t, 2t and zt+l’ and so forth, the measure can be written as: n 2 2 2 log t -m t=2 2t 1 Vlog — n _ 1 where the instability index, I5, is equal to the antilog of the square root of V108. As a measure of instability, COppock chose this method over I3 because it "was less laborious and lent itself to machine methods."ll Because of ready availability of computer time, I have used the other methods instead of Coppock's log variance method. Coppock discusses still another index which is obtained by fit- ting a linear regression line to the variable in question as a func- tion of time and then dividing the absolute value of the residual (i.e., the absolute value of the difference between the estimated value and the actual value) by the estimated value for each year. The resulting percentages are then summed and the result divided by the number of years in the period under consideration to yield what Coppock calls the average-percentage-deviation-from-trend index. Symbolically, the instability index, I6, can be represented as 100 2 lzt - (B0 +—Bl32l B+t I6 = t-l n0 B; where BO + Blt represents the 10Joseph D. COppock, International Economic Instability (New York: McGraw-Hill, 1962), pp. 23-24. 11Ibid., p. 23. 12Ibid., p. 24. 100 estimated value of export earnings in each year. However, this index is a measure of instability about the trend, not a measure of average year-to-year fluctuations and possesses no real advantages over 13 and In as instability indices; hence, I have not used this method. While the methods of measurement differ substantially, Coppock has shown that three of the methods give similar results.13 In his study, Coppock computed instability indices of export earnings of eighty-three countries during the 1947-1958 period using the United Nations method (I2), the log variance method (I5)’ and the average- percentage-deviation-from-trend method (I6)' He found that the sim- ple correlation coefficient between the instability indices calculated by the United Nations and log variance methods was .89 which is the same as the simple correlation coefficient between the instability indices calculated by the log variance and average— percentage—deviation-from—trend method. Hence, we should eXpect the different measures to give somewhat similar results, at least in the absence of trends.14 In calculating the instability indices, the period under consid- eration is from 1948 to 1964. As in Chapter I, the year 1948 was selected as a base because it appeared to be the first "normal" year following world War II. At the time the study was undertaken, 1964 was the last year for which data were available for a large number of countries. Therefore, in calculating the average year-to-year fluc— tuations, the first and last changes are from 1948 to 1949 and from 13Ibid., p. 25. 14This was generally the case in this study. However, in cer- tain instances, the result hinged on the choice of the instability index. 101 1963 to 1964, respectively. The data consist of the indices consid- ered in the previous chapter. As noted earlier, data are not avail- able for the entire period for some countries; the exceptions are noted in the tables. The Commodity Terms of Trade After calculating the average year-to—year fluctuation of the commodity terms of trade using both the 12 and 13 measures, it was found that they were almost identical.15 Hence, only the I2 series is reported in Table 17. The series ranges from 15.1 per cent (Ghana) to 1.8 per cent (Netherlands) with an average of 5.8 per cent for the forty-five countries.16 'With a number of exceptions, the commodity terms of trade appear to fluctuate more in underdevelOped than in develOped countries. To test whether the commodity terms of trade of underdeveloped countries fluctuate more than those of the deve1Oped countries, the countries are divided into two groups along the same lines discussed in Chapter 1.17 Using this classification scheme, it is found that the mean of the average year—to-year fluctuation, 12, of the commod- ity terms of trade for the twenty developed countries is 4.1 per cent while the mean for the twenty-five countries classified as 15The simple correlation coefficient between 12 and I3 is .99 for the commodity terms of trade. 16 The standard deviation is 2.6 per cent. 17The developed group, as in Chapter II, includes Australia, Aus- tria, Belgium-Luxembourg, Canada, Denmark, Finland, France, West Germany, Iceland, Ireland, Israel, Italy, Japan, Netherlands, New Zealand, Norway, Sweden, Switzerland, United Kingdom, and the United States. The rest are considered underdeveloped. 102 TABLE 17.--Instability of export prices, import prices, and the com- modity terms of trade in a sample of forty-five countries, 1948-1964 Average Year-to- Average Year-to- Average Year-to- Year Fluctuation Year Fluctuation Year Fluctuation Country of Export Prices, of Import Prices, of the Commodity Per Cent Per Cent Terms of Trade, Per Cent I2 12 I2 Argentinaa 6.2 5.4 7.1 Australia 9.4 3.2 7.9 Austria 6.9 7.9 6.9 Belgium- 5.2 3.9 2.9 Luxembourg Brazil 9.2 8.2 10.9 Canada 2.6 3.6 2.5 Ceylon 7.2 5.6 8.6 Chile 8.2 4.5 8.6 Colombia 8.2 2.2 8.1 Costa Rica 7.8 2.3 7.8 Cyprusb 8.0 4.7 6.1 Denmark 2.9 4.7 4.5 Ecuadorc 5.3 2.9 6.1 Finland 8.8 7.9 6.6 France 5.8 6.7 4.0 'West Germanyb 2.9 4.8 3.0 chanad 15.7 4.1 15.1 Greecea 7.5 5.4 4.9 Guatemalae 10.2 2.1 9.7 Honduras 6.0 2.5 7.1 Icelandc 1.6 5.5 5.5 Indiab 6.1 5.6 5.7 Ireland 3.3 3.9 4.0 Israelb 3.9 4.9 3.4 Italy 4.7 5.6 3.3 Jamaicad 5.1 5.0 5.8 Japanb 5.3 6.5 4.0 Mauritiusd 6.2 4.3 6.5 Netherlands 3.3 4.3 1.8 New Zealand 8.6 4.0 8.8 Nicaragua 9.2 2.3 9.3 Nigeria 7.3 4.6 5.5 Norway 4.6 4.7 2.9 Panama 5.8 2.3 7.5 Peruc 7.3 3.1 7.1 Philippines 7.6 3.4 6.8 Portugal 4.7 4.3 3.9 South Africa 5.7 3.9 4.5 Spain 5.0 5.8 5.6 103 TABLE 17--Continued _. . -o—uu Average Year-to- Average Year-to- Average Year—to- Year Fluctuation Year Fluctuation Year Fluctuation Country of Export Prices, of Import Prices, of the Commodity Per Cent Per Cent Terms of Trade, Per Cent 12 I2 12 Sweden 5.4 4.4 3.? Switzerland 2.4 4.0 2.9 United Kingdom 3.0 4.7 3.3 United States 2.3 3.9 3.2 Venezuela 1.9 3.6 4.6 Yugoslaviad 7.4 4.8 5.1 a 1951-1964. b1950.1964. c1950.1963. d1948.1963. e1948.1962. 104 underdeveloped is 7.1 per cent, a difference of three percentage points. This difference is significant at the .01 level using a one- tailed test. Hence, the commodity terms of trade of the underdevel— Oped countries appear to fluctuate more than the terms of trade of the develOped countries. This result is confirmed by ranking (in descending order) the countries by per capita gross domestic product and by fluctuations in the commodity terms of trade and calculating the Spearman rank corre— lation coefficient. If 1963 per capita gross domestic product data are used, the Spearman rank correlation coefficient is —.50 for a sample of thirty-eight countries. This is significant at the .01 level using a one-tailed test. The same results are obtained if the countries are ranked by per capita gross domestic product in 1958 or by per capita consumption in 1960.18 As the commodity terms of trade consists of the ratio of export to import prices, fluctuations in the terms of trade must be the result of fluctuations in export and/or import prices. Hence, it is of interest to examine fluctuations in the two series in order to determine the main source of instability in the commodity terms of trade. After calculating both 12 and I3 for the two series, it was found that the measures are almost identical; hence, only 12 is 18The coefficients are -.47 and -.45, respectively, which are both significant at the .01 level using a one-tailed test. The for- mer is based on a sample of forty-four countries and the latter on a sample of thirty-eight countries. 105 presented in Table 17.19 This lends support to Michaely's claim that "where a trend may be discerned it seems, from casual observation, to be mild, in comparison with the genuine price fluctuations."20 The results indicate that export prices fluctuate more than import prices. The mean of the average year-to-year fluctuation, I2, of export prices is 6.0 per cent while the mean of the average annual fluctuation, I2, of import prices is 4.5 per cent. Fluctuations in the export price series range from 15.7 per cent (Ghana) to 1.6 per cent (Iceland) while fluctuations in the import price series range from 8.2 per cent (Brazil) to 2.1 per cent (Guatemala).21 On an individual basis, fluctuations in export prices exceeded those of import prices in twenty-eight of the forty—five countries. This may be because countries tend to specialize in exports and generalize in imports. As discussed above, the mean of average annual fluctuations, I2, of the commodity terms of trade is 5.8 per cent. If export prices fluctuate 6.0 per cent per year and import prices 4.5 per cent, this means that some-~but not all--of the instability in export and import prices is offsetting to provide for greater stability in the commode ity terms of trade. If the countries are divided into developed and underdeveloped groups, it is found that the terms of the average year-to-year 19The simple correlation coefficient between I2 and I3 is .99 for export prices and .98 for import prices. ZOMichaely,'gp. git., p. 69. 21The standard deviations for the export and import price indices are 2.7 and 1.5 per cent, respectively. Thus, it appears there is less variability from country to country in import prices than in export prices. 106 fluctuation, 12, of export prices for the twenty developed countries is 4.6 per cent while the corresponding mean for the underdeveloped countries is 7.2 per cent, a difference of 2.6 percentage points. Using a one—tailed test, this difference is significant at the .01 level. This result is confirmed by ranking the countries by per capita gross domestic product in 1963 and by fluctuations in export prices and calculating the Spearman rank correlation coefficient. In this case, the coefficient is -.43 (for a sample of thirty-eight countries) which is significant at the .01 level using a one-tailed test. The result is also confirmed if the countries are ranked by per capita gross domestic product in 1958 (-.62) or by per capita consumption in 1960 (-.47). The former is based on a sample of forty-four countries and the latter on a sample of thirty-eight coun— tries. Thus, it appears that export prices tend to fluctuate more in underdeveloped countries than in developed countries. This could be due to differences in the degree of specialization in exports between countries or to specialization in different types of products. The hypotheses are discussed at length in the next chapter. The mean of the average year-to-year fluctuation, I2, of import prices for the twenty develOped countries is 5.0 per cent while the average for the underdevelOped countries is 4.1 per cent, a differ- ence of nine-tenths of one per cent. Using a one-tailed test, this difference is significant at the .05 level. Hence, it appears that import prices fluctuate more in develOped countries than in under- developed countries.22 However, rank correlations between per capita 22This could be explained if fluctuations in the demand for im- ports by developed countries have a greater impact on world prices than do fluctuations in the demand for imports by underdeveloped 107 gross domestic product and fluctuations in import prices fail to reveal such a relationship. If 1963 per capita gross domestic prod- uct is used, the coefficient is .04 for a sample of thirty-eight countries. This is insignificant at the .10 level using a one-tailed test. Similarly, if 1958 per capita gross domestic product and 1960 per capita consumption are considered, the rank correlation coeffi- cients are .10 and -.06, respectively, which are both insignificant. The former is based on a sample of forty—four countries and the latter on a sample of thirty-eight countries. Hence, we conclude that fluctuations in import prices do not vary systematically with stages of a country's development. Thus, the observed differences in stability in the commodity terms of trade of the develOped and underdeveloped countries are apparently due largely to differences in the stability of export rather than import prices. The Income Terms of Trade Because the income terms of trade of most countries under con- sideration have exhibited strong upward trends, the two measures of instability, I2 and I3, diverge considerably. Therefore, both have been included in Table 18. The results indicate that the mean of the average year-to—year fluctuation, 12, of the income terms of trade for the forty-five countries is 10.5 per cent while the mean of the countries. In this connection, it should be noted that import prices generally fluctuate more than export prices in develooed countries. On an individual basis, this is the case in fifteen of the twenty countries. However, it may simply be that developed countries import commodities which fluctuate more than commodities imported by underdevelOped countries. 108 TABLE l8.--Two measures of instability of the income terms of trade in a sample of forty-five countries, 1948-1964 Trend—Corrected Average Year—to- Average Year-to- Country Year Fluctuation, Year Fluctuation, Per Cent Per Cent 12 I3 Argentinaa 14.5 14.8 Australia 15.5 15.1 Austria 11.0 5.9 Belgium- 8.2 7.2 Luxembourg Brazil 10.1 10.0 Canada 6.5 6.3 Ceylon 11.1 11.0 Chile 10.7 10.5 Colombia 9.5 9.5 Costa Rica 9.5 9.1 Cyprusb 7.8 7.4 Denmark 8.8 5.2 Ecuadorc 10.6 10.6 Finland 11.5 10.1 France 10.9 6.5 West Germanyb 14.0 3.7 Ghanad 9.1 11.0 Greecea 10.7 6.3 Guatemalae 7.7 8.1 Honduras 9.0 9.0 Icelandc 11.8 9.0 Indiab 7 . 8 7. 6 Ireland 10.1 8.4 Israelb 15.0 14.7 Italy 12.8 9.4 Jamaicad 11.3 10.9 Japanb 14.2 10. 0 Mauritiusd 15.8 13.8 Netherlands 10.7 4.5 New Zealand 8.2 8.0 Nicaragua 17.2 16.6 Nigeriad 8.6 6.9 Norway 9.3 7.1 Panama 13.4 13.3 Peruc 9.3 8.5 Philippines 9.5 8.3 Portugal 10.6 9.4 South Africa 8.9 6.8 Spain 10.7 9.8 TABLE l8--Continued 109 a. eye. ”a... _-- Average Year-to- .-_-—.-——.e..___ .—_—_..- Trend-Corrected Average Year-to- Country Year Fluctuation, Year Fluctuation, Per Cent Per Cent I2 I3 Sweden 8.8 5.1 Switzerland 7.7 3.8 United Kingdom 5.3 3.2 United States 8.1 7.1 Venezuela 5.2 4.9 Yugoslaviad 16.0 16.0 a 1951-1964. b1950.1964. c1950.1963. d1948.1963. e1948.1962. 110 trend-corrected average annual fluctuation, I 3, is 8.9 per cent.23 The I2 series ranges from a high of 17.2 per cent (Nicaragua) to a low of 5.2 per cent (Venezuela) while the I3 series ranges from 16.6 per cent (Nicaragua) to 3.2 per cent (United Kingdom). To test whether the income terms of trade fluctuate more vio- lently in underdeveloped than in developed countries, the countries are classified into two groups using the same classification scheme as in the previous section. After dividing the forty-five countries into the two groups, it is found that the mean of the average year- to-year fluctuation, 12, of the income terms of trade of the twenty developed countries is 10.4 per cent and the mean of the underdevel— oped countries is 10.6 per cent, a difference of only two-tenths of one per cent. Using a one-tailed test, the difference is not signif— icant at the .10 level. The same result is obtained by compiling Spearman rank correlation coefficients between rankings by per capita gross domestic product (consumption) and fluctuations in the income terms of trade. If per capita gross domestic product in 1963 is considered, the resulting rank correlation coefficient is .06. Based on a sample of thirty-eight countries, the coefficient is insignifi- cant at the .10 level using a one-tailed test. If per capita gross domestic product in 1958 and per capita consumption in 1960 are con- sidered, the coefficients are -.08 and -.l3, respectively. The former is based on a sample of forty-four countries and the latter on a sample of thirty-eight countries. Both are insignificant at the .10 level using a one-tailed test. 23The simple correlation coefficient between 12 and I The standard deviations of the two series are 2. 8 and 3. 3 BerS cent, respectively, for I2 and 13. 111 If the trend-corrected series, 13, is used, the means are 7.5 and 10.0 per cent, respectively, for the developed and underdeveloped countries; this is a difference of some 2.5 percentage points and is significant at the .01 level using a one-tailed test. The relation- ship is confirmed by compiling Spearman trank correlation coeffi— cients between per capita gross domestic product (consumption) and fluctuations in the income terms of trade. The Spearman rank corre— 1ation coefficient between per capita gross domestic product in 1963 and fluctuations in the income terms of trade is —.37. Based on a sample of thirty-eight countries, the coefficient is significant at the .01 level using a one-tailed test. If per capita gross domes— tic product in 1958 and per capita consumption in 1960 are considered, the correlation coefficients are -.46 and —.48, respectively. ‘With samples of size forty-four (for the former) and thirty-eight (for the latter), both are significant at the .01 level using a one—tailed test. The main reason for the greater differential in the I3 series is, of course, the strong upward trend in the income terms of trade in the developed countries. Since the developed countries can more or less count on the continued upward trend in their exports, the rele- vant comparison involves the trend-corrected series; hence, we con- clude that the income terms of trade of the underdeveloped countries fluctuate more violently than the income terms of trade of the developed countries. As shown earlier, the income terms of trade are calculated by multiplying the commodity terms of trade of each country by the coun- try's export volume index. Hence, fluctuations in export volume may offset fluctuations in the commodity terms of trade to provide for 112 greater stability in the income terms of trade. Since the income terms of trade (10.5 and 8.9 per cent, respectively, for 12 and I3) fluctuate more widely than the commodity terms of trade (5.8 per cent) regardless of the method used to compute the indices, it appears that fluctuations in export volume intensify rather than offset fluctuations in the commodity terms of trade.24 Conclgsion In this chapter, it was found that the commodity and income terms of trade fluctuate more for underdeveloped countries than for developed countries. If, as generally assumed, instability is harm— ful, this means the instability problem is worse for underdeveloped countries. However, Alasdair I. MacBean has recently questioned the assumption that the present degree of international instability is harmful.25 If he is correct, the results shown here are not particu— larly meaningful. Because of certain methodological problems, much of MacBean's statistical evidence is questionable.26 Nevertheless, it is no longer clear that international instability is harmful.27 24This is true for both the developed and underdeveloped coun- tries. Fluctuations in export volume will be discussed in the next chapter. 25Alasdair I. MacBean, Export Instability and Economic Develop- ment (Cambridge: Harvard University Press, 1966). 26On this point, see A. Maizels, Review of Export Instability and Economic Develppment, by Alasdair I. MacBean, American Economic Review, LVIII (June, 1968), 575-80. 27The author plans to investigate this question in the future. CHAPTER IV INSTABILITY OF EXPORT EARNINGS The second hypothesis to be tested in this thesis concerns the instability of export earnings and one prOposed solution--export diversification. Specifically, is diversification of exports (in both the product and geographical sense) likely to provide a greatly increased measure of stability in eXport earnings? If so, a country may wish to embark on a purposeful program of export diversification in order to stabilize its export earnings. If not, the country may wish to pursue other approaches to stabilization or concentrate on alleviating the effects of instability. To test the hypothesis empirically, several measures are needed. The first is a measure of instability of export earnings. Using measures developed in the previous chapter, the degree of instability of export earnings is calculated for seventy-eight countries in this chapter. The second measure, that of export diversification, is developed in the following chapter where the two measures are brought together and several hypotheses concerning instability of export earnings and export diversification are tested using regression analysis. Instability and Its Effects Before proceeding to the empirical analysis, however, it is desirable to discuss the relationship between international 113 114 instability and the domestic economy and offer several reasons why underdeveloped countries may be expected to experience more insta— bility than developed countries. Instability of export earnings may affect the domestic economy in a variety of ways. For example, fluctuations in exports may lead to instability in the domestic economy through the foreign trade multiplier.l Variations in export earnings have a direct impact on producers' incomes and, as a consequence, affect their expenditures on consumption and investment. As their eXpenditures change, reper- cussions will spread throughout the economy to produce additional changes in the level of national income (unless offset by fiscal and/ or monetary policy).2 Changes in the total value of exports may also affect the domestic supply of money so as to produce changes in 3 national income. variations in export proceeds may cause fluctua— tions in prices, employment, and real wages for the same reasons. Along with fluctuations in the level of national income, fluctu— ations in export earnings and prices may have undesirable social and political effects through changes in the distribution of income, especially in underdeveloped countries. For example, a sudden in- crease in the export earnings of an important crop may appear to re- distribute income in an arbitrary manner. One group of farmers is enriched while domestic prices may rise to make other groups worse 1For a formal analysis, see Charles P. Kindleberger, Interna- tional Economics (3rd ed.; Homewood, Illinois: Richard D. Irwin, 19635. pp. 177-99. 2The impact will depend, in part, on the size of the export sector relative to the rest of the economy. 3Henry C. Wallich, Menetary Problems of an Export Economy (Cambridge: Harvard University Press, 1950), pp. 206—09. 115 off. Similarly, a sudden drop in earnings can create very serious tensions, particularly in societies where existing tensions are already high. This may be especially true in countries where race or tribal connections determine occupations or people live at or near subsistence levels. Along with the impact on national income and its distribution, variations in export earnings may affect the allocation of resources. If instability creates uncertainty, investors may be deterred from specializing in eXport industries which yield the highest returns because they do not wish or cannot afford to take the risks involved in such enterprise.“ If this occurs, the country will suffer loss of income (even if full employment is maintained) and foreign exchange. Hence, instability may prevent an optimum allocation of resources. In addition to the allocation effect, instability may reduce the rate of investment because it introduces uncertainty into the invest— ment process. First, instability increases the difficulty of esti- mating returns on investment and, second, it introduces the possibil- ity of foreign—exchange problems which would impede the importation of necessary capital goods or raw materials at a given time. "In addition, the existence of these risks may make the suppliers of capital and credit charge higher interest rates and impose more stringent conditions."5 Under these circumstances, the immediate “Moreover, Benjamin Higgins has argued that the "repercussions may be felt through the government budget because of the close tie between revenues" and export earnings of underdeveloped countries. Benjamin Higgins, Economic Development: Principles, Problems, and Policies (New York: WQ'W. Norton and Company, 1959), p. 557. Because of uncertainty, government planning for development will become more difficult. 5MacBean, pp. cit., p. 29. 116 effect would probably to be reduce investment and national income and, in the long run, impede the country's rate of growth. Apart from these effects, severe export instability will tend to produce temporary belance of payments problems in countries which are short of international reserves.6 Such instability would be partic- ularly harmful to underdeveloped countries because of their lack of reserves and the possibility that luxury imports or non-necessities have already been cut out of their imports. Under these circum- stances, any drop in export earnings of foreign exchange is likely to cut into capital good and raw material imports. Moreover, recur- ring balance of payments problems may lower confidence in the mainte- nance of existing exchange rates and lead to flights of capital. Thus, instability seems to be harmful and, in certain instances, it appears to be more harmful to underdeveloped than to developed countries. In this regard, certain other considerations are also important. First, underdeveloped countries may be subject to greater instability than developed countries. Second, the export sector of underdeveloped countries is alleged to be of greater quantitative importance-~relative to the rest of the economy-—than the export 7 sector of develOped countries. Finally, underdeveloped countries 6To the extent that countries borrow to even out their capacity to import over booms and slumps, the problem is mitigated. Instabil— ity in the system may cause balance of payments difficulties which lead to import restrictions and declining world trade. However, the general solution may be to increase the liquidity of the system rather than to reduce instability. Since we are concerned with indi— vidual countries and the possibility of stabilizing their export earnings through export diversification, we need not be concerned with the system as a whole. 7Henry C. wallich, "UnderdevelOped Countries and the Interna- tional Monetary Mechanism," Money, Trade, and Growth (New York: MacMillan, 1951), pp. 15-16. Higgins, pp. git,, pp. 555-57. ll? are said to lack the techniques and facilities necessary for effec- tive countercyclical monetary and fiscal policies.8 In addition to providing measures of export instability, the main purpose of this chapter is to test the validity of the first hypothesis. The other hypotheses deserve consideration but are beyond the scope of this study.9 Causes of Instability Any number of explanations could be offered to account for greater instability in underdeveloped countries. One possibility is, of course, that instability is related to export concentration. From casual observation, it appears that the exports of most under- developed countries are concentrated heavily in one or a few commod— ities while exports of the developed countries are more evenly distributed over a wide range of commodities; hence, fluctuations in a country's export earnings may be directly related to export concentration. On an a priori basis the hypothesis is appealing since if a country's eXports consist primarily of one good, any change in the export earnings of that good constitutes about the same change in the country's export earnings as a whole. The larger the number of goods exported by the country or the more evenly resources are divided among the goods produced for export--or, in short, the lower the degree of commodity concentration of a country's exports--the more 8MacBean, pp. git., p. 26. Higgins, ibid., p. 557. 9MacBean argues the second hypothesis is not supported by the evidence. Ibid., p. 58. 118 likely it is that changes in export earnings of individual exports will tend to offset each other, thereby stabilizing the country's export earnings as a whole. Of course, the ultimate source of inter- national economic instability lies in those factors which affect demand and supply, not export concentration per se. A second hypothesis which purports to explain the greater insta— bility of export earnings in underdeveloped countries relates to the concentration of exports by geographical region. For example, if a country ships the bulk of its exports to one market, it would, of course, be highly sensitive to changes originating in that market whereas countries exporting to a wide variety of markets would not be particularly sensitive to changes in any one market. Under this hypothesis, the exports of underdevelOped countries would presumably be more highly concentrated in the geographical sense of the terms than exports of develOped countries. This seems plausible as "most underdeveloped countries sell the greater part of their exports to the United States, the United Kingdom, or France."10 If the hypoth- esis is correct, countries may wish to diversify their exports geographically. A third hypothesis relates to the type of product which a coun- try exports. Under this hypothesis, fluctuations in export earnings may result from the export of primary products as opposed to indus— trial products, since the exports of primary goods may fluctuate more violently than exports of industrial goods or, alternatively, the earning time paths of primary products may be more intercorrelated than those of industrial goods or industrial goods and primary 10Ibid., p. 25. 119 products. It has been generally accepted that the prices of primary prod- ucts fluctuate more than prices of most manufactured goods.11 A major reason for this lies in the short-run inelasticity of both sup- ply and demand in relation to prices of primary products as compared with prices of manufactured goods.12 On the supply side, there is a lack of short-term response to price changes because of the lengthy period required to bring about a significant increase or decrease in production in such major craps as coffee, cocoa, tea, and rubber. Moreover, even in cases where creps are planted and harvested within the same year, the decision to plant must take place many months before harvest time. In each case, a new level of prices can affect only future plantings. In regard to most metals and minerals, the price-elasticity of output is likely to be low also.13 In such industries, the major costs appear to be fixed and mines are likely to continue in opera— tion so long as prices are covering variable costs. If variable costs are low, prices may have to fall considerably before opera- tions cease. This tendency is reinforced by the fact that closure and subsequent reopening of many mining operations involve substan- tial costs. Conversely, when demand increases, output may increase llIQid., p. 23. Also, Haberler, "Terms of Trade and Economic Development,” gp.‘g;t., p. 289. 12Henry C. Wallich, "Stabilization of Proceeds from Raw Mater- ials Exports," Economichevelopment for Latin America, 0 . git., p- 349. l3MacBean,,gp. git., p. 23. At this point, it is convenient to follow MacBean's convention of distinguishing between elasticity of output and elasticity of supply. 120 but many underground mines operate shift systems which limit their ability to vary labor inputs. Hence, with both agricultural and mining products, it seems likely that the response of output to changes in prices should be low compared with output elasticities of manufactured goods. It is true that the supply elasticity of commodity exports may be greater than the output elasticity, because export- ers' stocks, domestic demand, and even, in some cases, imports for re-export may respond to price incentives, but the additional flexibility from these is generally likely to be small. Few underdeveloped countries hold large stocks of exportable commodities; such stockpiling is often barred by technical difficulties and expense. Domestic consumption of exportable commodities is also, in general, minute. Importing for re—export in quantities great enough to affect elasticities significantly is rarely possible for underdeveIOped countries. On the demand side, there appears to be evidence to indicate that response to changes in the prices of most food and raw materi— als is slight. The price elasticities of demand for food and 15 beverages are commonly accepted as low. Moreover, the demand for many raw materials is derived from the demand for final products and their costs form only a small part of the cost of such products. Under these circumstances, even a substantial change in the prices of raw materials will not be reflected in changes in the prices of the finished products.16 Hence, the price elasticities of demand for raw materials are likely to be low. If supply elasticities are low, fluctuations in demand resulting from changes in tastes, industrial activity, speculation, and the lugbid. l5gpgg. Also, wallich, "Stabilization . . .," pp. 313., p. 349. 16 MacBean, ibid. 121 like, will induce large fluctuations in export prices and earnings. If the price elasticities of demand are low, changes in supply caused by the weather or any other factor will result in sharp move— ments in export prices. Export earnings will also fluctuate but the change in export prices will tend to be offset somewhat by the change in export volume. Given the further factor that both supply and demand for most commodities are characteristically unstable, a high degree of price instability is inevitable. For both food and agricultural raw materials it is evident that output variations are very likely because of the normal hazards of farming.17 In most underdeveloped countries these hazards are aggravated by lack of technique or resources for flood control and lack of pesticides and fertilizers which have helped to reduce output varia- bility of agriculture in richer and more develOped agri- cultural systems. In addition, some current changes in output may be related to price conditions of several years earlier. Tree crops such as coffee and cocoa exem- plify this form of instability. While minerals are less subject to output variability, demand for them tends to be more closely related to cycles in indusgrial activity and speculation and can fluctuate widely.l "In brief, low price elasticities combined with uncontrolled variability in demand, supply, or both provide an entirely credible explanation for sharp instability in both prices and proceeds of primary products."19 Since, on the average, primary products form a greater proportion of exports in underdeveIOped countries than they do in deveIOped countries, this may account for greater instability of export earnings in the former. If this hypothesis is correct, the underdeveloped countries may wish to reduce their dependence on the export of primary products in order to stabilize thetrexport earnings. 17For an elaboration, see Higgins, pp. 213., pp. 5h5-57. l8MacBean, pp. g;§,, p. 25. 19Ibid. 122 The Empirical Evidence As before, the study is concerned with only year-to-year fluctu- ations, and in calculating the instability indices, the period under consideration is from l9b8 to 1964. As in the previous chapters, the years 1938 and 1964 were selected because 1948 appeared to be the first "normal" year following World war II and 1964 was the last year for which data were available for a large number of countries at the time the study was undertaken. Therefore, in calculating the average year-to—year fluctuations, the first and last changes are from l9h8 and from 1963 to 1964, respectively. For some countries, data are not available for the entire period; the exceptions are noted in the tables. The methods used to calculate the instability indices are the same as in the previous chapter. Data to calculate instability indices of export earnings are available for a large number of countries. I have selected seventy- eight which, in 1964, accounted for 94.5 per cent of total world ex- ports excluding Cuba, Indonesia, and the Sino—Soviet area.20 The remaining countries were excluded mainly because of lack of data or because territorial or other changes made the existing data incon— sistent. Also, a few countries, such as South Korea, were excluded because they experienced severe disruptions to their economy and it seemed desirable to omit the most obvious cases. Data, in United States dollars, for these countries are found in the Supplement to 1965166 Issues of International Financial Statistics 20Coppock included eightysthree countries; the main difference in coverage is that he included the Sino-Soviet area, Cuba, and Indo- nesia. Coppock, pp, pip., pp. 158—60. 123 published by the International Monetary Fund.21 The data refer to merchandise trade only because of the scarcity of data on trade in services; however, in most countries, merchandise trade is of over— whelming importance. Dollar figures are used since we are concerned in part with fluctuations in a country's ability to finance imports and this suggests use of a commonly accepted international currency?Z Also, countries which have devalued (or revalued) their currencies show sharp discontinuities in their exports when recorded in domes— tic currencies which are not so apparent when recorded in dollars. Under such circumstances, dollar figures would provide a better meas- ure of actual fluctuations of exports than the domestic currency. Hence, it seems desirable to use dollar figures in cases where ex- change rates have varied. Finally, dollar figures are used because we wish to relate the instability indices to other indices which are based on dollar figures. Even assuming that the basic data are reasonably accurate-—which is not always the case--the use of dollar figures increases the pos-' sibility of errors since the official exchange rate used to convert exports to dollar terms may diverge considerably from the free ex- change rate. Nevertheless, to measure fluctuations in a country's export earnings in relation to that country's economic development and its balance of payments, use of dollar figures seems to be the logical choice. The results, shown in Table 19, indicate that export earnings 1International Monetary Fund, International Financial Statis- tics. Supplement to 1965/66 Issues (Washington, 1965), pp. xvi-xix. 221m other contexts, use of domestic currencies may be more appropriate. 124 TABLE 19.——Three measures of instability of export earnings in a sample of seventy-eight countries, 1948—1964 Trend-Corrected Average Year-to- Average Year-to- Normalized Year Fluctuation Year Fluctuation Standard Error Country of of of the Export Earnings, Export Earnings, Estimate Per Cent Per Cent 12 I3 I4 Iran 23.4 20.3 .387 Liberiaa 21.4 22.6 .204 Sudan 18.3 19.2 .207 Norway 18.1 9.4 .105 Japan 17.8 9.9 .158 Urugua 17.2 17.1 .224 Israel 16.8 19.0 .275 Yugoslavia 16.4 14.1 .236 Nicaragua 16.0 13.2 .201 Iraqc 15.6 10.6 .131 Malaya 15.5 15.2 .210 China: Taiwan 15.4 13.4 .327 Bolivia 15.3 15.6 .204 Mauritius 15.1 14.6 .175 Barbados 14.2 13.8 .154 Germany,W'estc 13.4 4.8 .057 Finland 13.1 12.6 .131 Iceland 13.0 12.3 .176 Hong Kong 12.9 12.3 .216 Argentina 12.9 12.8 .204 Guadeloupe 12.7 12.3 .109 Panama 12.5 11.6 .242 Pakistan 12.5 13.2 .224 Turkey 12.4 11.8 .142 Dominican Republic 12.4 12.2 .104 Ethiopia 12.4 12.7 .134 Tanzania 12.4 11.4 .136 Cameroond 12.3 10.9 .118 Austria 12.0 5.5 .050 Ecuador 12.0 12.5 .118 Thailand 11.9 10.6 .132 Italy 11.8 9.8 .195 Peru 11.8 10.1 .172 Philippines 11.8 10.9 .116 Australia 11.7 11.2 .131 Cyprus 11.5 10.8 .136 Angola 11.5 9.9 .132 Reunion 11.4 11.0 .126 Burma 11.3 11.2 .120 TABLE l9--Continued 125 Trend-Corrected Average Year—to- Average Year—to- Normalized Year Fluctuation Year Fluctuation Standard Error Country of of of the Export Earnings, Export Earnings, Estimate Per Cent Per Cent 12 13 I4 Greece 11.3 7.3 .106 Ugandae 11.2 11.8 .160 Martinique 11.1 9.7 .127 Chile 11.0 11.0 .109 Tunisia 10.9 10.3 .153 Jamaica 10.7 10.9 .106 Kenyae 10.4 7.5 .102 France 10.4 7.5 .092 El Salvador 10.2 7.8 .119 United Arab Republic 10.1 10.1 .139 Netherlands 10.0 4.5 .077 Nigeria 10.0 8.9 .090 Portugal 10.0 9.3 .128 Costa Rica 9.9 8.5 .103 Brazil 9.8 9.5 .121 Sweden 9.8 8.3 .109 Colombia 9.5 9.2 .197 Honduras 9.5 9.7 .109 Ghana 9.5 9.1 .094 Belgium-Luxembourg 9.4 9.9 .116 Mexico 9.1 7.7 .098 India 9.1 8.6 .114 United States 9.0 9.0 .089 Ireland 8.9 7.2 .104 Trinidad 8.4 6.7 .107 Saudi Arabiac 8.3 4.9 .074 Denmark 8.2 4.5 .079 Mozambique 8.1 8.3 .101 New Zealand 8.0 7.6 .083 Ceylon 7.7 7.4 .077 Guatemala 7.6 6.6 .106 Spain 7.5 8.0 .156 N. Antilles 7.5 7.3 .177 South Africa 7.4 6.2 .066 Switzerland 7.3 4.4 .081 Canada6 7.0 5.8 .076 Morocco 6.7 5.8 .080 126 TABLE 19—-Continued Trend-Corrected Average Year-to- Average Year-to- Normalized Year Fluctuation Year Fluctuation Standard Error Country of of of the Export Earnings, Export Earnings, Estimate Per Cent Per Cent 12 13 I4 Venezuela 6.5 5.0 .150 United Kingdom 5.4 4.0 .043 al948—l963. b1948 estimated. °1950-1964. dData refer to East Cameroon only. e1949-l964. 127 fluctuated considerably during the period under consideration.23 However, the results vary somewhat depending on the method used to calculate the indices. If the United Nations method is used, the mean (unweighted arithmetic average) of the average year-to—year fluctuations, 12, of export earnings is 11.6 per cent while if the trend-corrected method is employed, the mean of the average year-to year fluctuations, I3, is 10.2 per cent. As an additional measure of instability, the Table contains the normalized standard error of the estimate, 14’ for each of the countries; however, since this index is largely a measure of the variation of the series as a whole around the trend line rather than year-to-year fluctuations, the discussion in this chapter is limited to I2 and I3.24 Since we have two measures of the average annual fluctuation of export earnings, which measure is relevant when the two diverge? Generally speaking, the two indices will diverge if there is a strong upward trend in the export earnings of the country in question. Typ- ically, the developed countries have experienced strong upward trends in their export earnings; hence, use of the United Nations method would be somewhat misleading since it provides only a crude correc- tion for such trends. Moreover, as the developed countries are 23However, there is considerable diversity of experience. The I series ranges from a high of 23.4 per cent (Iran) to a low of 5.4 per cent (United Kingdom) while the I series ranges from 22.6 per cent (Liberia) to 4.0 per cent (United Kingdom). The standard devi- ations for the two series are 3.4 and 3.7 per cent, respectively. The simple correlation coefficient between the two series is .84. 24The mean of the normalized standard error of the estimates is .139 or, in percentage terms, 13.9 per cent. The standard deviation is .062 or, in percentage terms, 6.2 per cent. The simple correla— tion coefficient between 12 and I4 is .68 while the coefficient between 13 and I4 is .76. 128 assured, more or less, of a continuing upward trend in their export earnings, they are likely to be concerned only with fluctuations in their export earnings about that trend, thus, the relevant measure is the trend-corrected index, I3. Similarly, some underdeveloped countries have experienced strong upward trends in their export earn- ings; therefore, the relevant measure for these countries is the trend-corrected index, I3. However, in either case, the United Nations method may provide a better index of instability if the trend cannot easily be approximately by a straight line.25 For those coun- tries which did not eXperience a trend in their export earnings, either measure would presumably be acceptable. In regard to earlier studies, Coppock found considerably more instability in export earnings than indicated above. Using the log variance method, he found the mean of the average year-to-year fluc— tuation of export earnings to be 21.8 per cent. As noted earlier, the countries under consideration are not the same in both studies. However, the discrepancy remains even if the means are obtained for the countries common to both samples. The log variance mean for the fifty-nine countries is 21.3 per cent while the means for I2 and I3 are 11.7 and 10.2 per cent, respectively, for the same countries. The difference in results could be attributed to a number of 25To compute an index of instability, MacBean uses the average percentage deviation of export earnings from their five-year moving average (centered on the mid-year). He believes this is a better measure of instability when no simple linear trend exists. Since I obtained "good" fits assuming a linear trend, I did not use MacBean's method. The latter has the disadvantage of losing two years from the beginning and end of the time series and is also less convenient for computer calculation. MacBean,Ipp,'p;p,, p. 34. 129 factors, including, of course, differences in the method of computa- tion.26 However, the main reason why Coppock found more instability in export earnings than reported here is because the period he stud- ied was largely dominated by the recovery from World war II and the Korean'war. As noted earlier, Coppock's study covers the 1946—1958 period while the present study spans the 1948-1964 period. Since many countries were recovering from the war in 1946, 1947, and 1948, their export earnings increased substantially during those years, and if 1946 and 1947 are included in the computation of the indices, it is obvious that the indices will show greater instability than if the two years are omitted. The Korean war mav also have unduly influ— enced Coppock's indices because of the short time span under consideration.27 To test whether export earnings fluctuate more in underdeveloned than in develOped countries, the seventy-eight countries are divided into two groups using the same classification scheme as in Chapter 11.28 As such, the scheme is subject to the shortcomings previously listed. 61n Coppock's study, export earnings are often recorded in the domestic currency of the country under consideration while all export earning data in this study are in the U. 8. dollars. This difference probably adds to the discrepancy. 27It may be, of course, that the degree of international eco- nomic instability is declining. Any number of reasons cOuld be offered for such a decline. 28The following countries are considered developed throughout the remainder of the study: Australia, Austria, Belgium-Luxembourg, Canada, Denmark, Finland, France, W. Germany, Iceland, Ireland, Israel, Italy, Japan, Netherlands, New Zealand, Norway, Sweden, Switz- erland, United Kingdom, and the United States. The following coun- tries, not considered in Chapter II, are classified as underdeveloped: Angola, Barbados, Bolivia, Burma, Dominican Republic, Guadeloupe, Iran, Liberia, Libya, Martinique, Mexico, Mozambique, Reunion, Saudi Arabia, Trinidad and Tobago, Tunisia, United Arab Republic, and Uruguay. 130 The means of the average year-to-year fluctuation of export earn- ings of the twenty countries classified as developed are 11.0 and 8.4 per cent, respectively, for 12 and I3 while the corresponding means for the fifty-eight countries categorized as underdevelOped are 11.7 and 10.9. If the I2 means are considered, the difference is only seven-tenths of a per cent which is not significant at the .10 level.29 However, if the trend-corrected means are considered, the difference is 2.5 percentage points which is significant at the .01 level. The difference between the 12 and I3 series is accounted for by the strong upward trend in eXport earnings for the developed coun- tries during the post—war period. Germany, for example, exhibited a strong upward trend in its export earnings. Her average year-to-year fluctuation, 12, of export earnings, is 13.4 per cent; however, when corrected for trend, the average fluctuation, I3, is only 4.8 per cent. As a rule, the countries classified as underdeveloped did not show such a strong upward trend in their export earnings. Since the developed countries, for the most part, can count on a continued upward trend in their export earnings, the instability about this trend, measured by 13, is more important than the average year-to-year fluctuation, 12, which is not corrected for trend. In the case of the underdeve10ped countries, either 12 or I3 is consid— erably greater than 13 for the developed countries. Hence, export earnings tend to fluctuate more in underdeveloped countries than in 29Based on a one-tailed test. The tests which follow are one— tailed tests unless otherwise stated. 131 develOped countries with, of course, some exceptions.30 As an additional test, the countries were ranked (in descending order) by per capita gross domestic product (consumption) and by fluctuations in export earnings. The Spearman rank correlation coef- ficient between the 1958 per capita gross domestic product series and the I2 series is —.l95. Based on a sample of seventy-four countries, the coefficient is barely significant at the .05 level. On the other hand, if the 13 series is considered, the coefficient is -.344 which is highly significant. Approximately the same results follow if 1963 per capita gross domestic product and 1960 per capita consumption are considered. Based on a sample of fifty-seven countries, the coeffi- cient between the per capita gross domestic product series and the 12 series is -.210 which is significant at the .10 level but not at the .05 level. When the 13 series is considered, the coefficient in- creases to —.415 which is significant at the .01 level. The coeffi- cient between the per capita consumption series and the I2 series is -.264. Based on a sample of fifty-six countries, the coefficient is significant at the .05 level. However, if the I3 series is consid- ered, the coefficient is -.460 which is, once again, significant at the .01 level. Sources of Export Earning Instability Fluctuations in export earnings may be due to fluctuations in 30Alasdair I. MacBean reached the same conclusion on the basis of Coppock's data. MacBean, ”Causes of Excessive Fluctuations in Export Proceeds of UnderdevelOped Countries," Oxford University Institute of Economics and Statistics Bulletin, XXVI (November, 1964), 325. See also, MacBean, pp. p23,, p. 36. 132 export prices, export volume, or some combination of the two. There— fore, it is of some interest to countries considering various poli— cies to stabilize export earnings to know the major source of instability. Unfortunately, no conclusive results can be reached in regard to this question partly because of the nature of the data. In order to compute instability indices for export prices and volume, it is necessary to use the export price and volume indices as reported in various publications of the United Nations and the Inter— national Monetary Fund. As such, the indices are subject to all the shortcomings listed above. The most important shortcoming, however, is that the price and volume indices are largely based on domestic currencies while export earnings are measured in U. 8. dollars; hence, the instability indices for export prices, volume, and export earn- ings are not really comparable. For this reason, we can only make some tentative suggestions about the stability of export prices and volume in regard to export earnings. Data in the form of price and volume indices are available to calculate instability indices for export price and volume for forty- 31 In 1964, these countries accounted for 86.2 per five countries. cent of world exports, where the world total excludes Cuba, Indonesia, and the Sine-Soviet area. Export price instability was discussed in the previous chapter; however, the instability index, I is included in Table 20 for 32 2’ convenience. The export volume instability indices, I2 and 13, are 31The sample is limited by the availability of export price and volume data. 32Since the 12 and 13 indices are almost identical, only 12 is presented. 133 TABLE 20.--Instability of export prices and volume in a sample of forty-five countries, 1948—1964 Trend-Corrected Average Year—to- Average Year—to- Average Year—to— Country Year Fluctuation Year Fluctuation Year Fluctuation of EXport Prices of Export Volume, of Export Volume, Per Cent Per Cent Per Cent I2 12 13 Argentina8 6.2 13.1 12.4 Australia 9.4 8.4 7.7 Austria 6.9 12.2 6.7 Belgium-Luxembourg 5.2 8.6 5.0 Brazil 9.2 9.9 10.0 Canada 2.6 5.6 5.0 Ceylon 7.2 4.1 3.5 Chile 8.2 7.0 6.7 Colombia 8.2 6.7 6.9 Costa Rica 7.8 9.1 9.3 Cyprusb 8.0 5.9 5.4 Denmark 2.9 9.4 3.8 Ecuadorc 5.3 10.2 8.5 Finland 8.8 9.5 6.9 France 5.8 12.2 7.8 Germany, Westb 2.9 11.2 6.5 Ghanad 15.7 10.2 9.5 Greecea 7.5 8.3 5.2 Guatemalae 10.2 6.6 6.1 Honduras 6.0 7.5 7.4 Icelandc 1.6 11.0 9.7 Indiab 6.1 6.1 5.6 Ireland 3.3 9.3 7.7 Israelb 3.9 15.0 13.9 Italy 4.7 12.5 9.2 Jamaicad 5.1 8.6 8.0 Japan 5.3 12.2 10.0 Mauritiusd 6.2 13.2 13.0 Netherlands 3.3 11.1 4.3 New Zealand 8.6 5.7 5.3 Nicaragua 9.2 15.1 14.7 Nigeriad 7.3 7.1 6.2 Norway 4.6 8.0 6.4 Panama 5.8 11.2 11.6 Peruc 7.3 8.4 6.7 Philippines 7.6 9.0 6.4 Portugal 4.7 8.6 7.0 134 TABLE 20-—Qontinued Trend-Corrected Average Year-to- Average Year-to- Average Year—to- Country Year Fluctuation Year Fluctuation Year Fluctuation of Export Prices, of Export Volume, of Export Volume, Per Cent Per Cent Per Cent 12 I2 I3 South Africa 5.7 7.1 4.4 Spain 5.0 9.4 10.5 Sweden 5.4 8.5 4,4 Switzerland 2.4 7.5 4.4 United Kingdom 3.0 4.6 3.4 United States 2.3 7.4 3.4 Venezuelab 1.9 6.2 3.7 Yugoslaviad 7.4 13.3 13.u a 1951-1964. b195o-196u. c1950-1963. d1948—1963. 61948—1962. 135 also presented in Table 20.33 The results indicate that, regardless of the measure employed, export volume appears to fluctuate more than export prices.3u The mean of the average year—to—year fluctuation, 12, of export prices is 6.0 per cent while the means of the average annual fluctuation of export volume are 9.2 and 7.5 per cent, respectively, for I2 and I3. This implies that the main source of instability of export earnings is export volume.35 This is the conclusion reached by the United Nations.36 Although export volume appears to fluctuate more than export prices, we cannot conclude with certainty that approaches designed to stabilize export volume will be more successful in stabilizing export 33The simple correlation coefficient between I and I is .80. Fluctuations in export volume range from 15.1 per cent (Nigaragua) to 4.1 per cent (Ceylon) in the I series and 14.7 per cent (Nicaragua) to 3.4 per cent (United Kingdom) in the 13 series. The stardard deviations for the two series are 2.7 and 2.6 per cent, respectively. 3“This is consistent with results Obtained by MacBean and the United Nations. MacBean, Export Instability . . ., _p. 213., p. 46. United Nations, Instability in Expprt Markets of Underdeveloped Coun— tries (New York, 1952). 35While, on the average, fluctuations in export volume tend to exceed fluctuations in export prices, there are a number of excep- tions when each country is considered individually. If the I series of export prices and volume are considered, fluctuations in export prices exceeded those of export volume in nine (or twenty per cent) of the forty-five countries. If the I series of export prices is compared with the I series of export volume, it is apparent that fluctuations in engrt prices exceeded those in export volume in eighteen (or forty per cent) of the forty-five countries. Since the trend-corrected series is the relevant index in most instances, it is evident that there are many exceptions to the rule. 36United Nations, Instability in Export Markets . . ., pp. pip., p. 57. However, Wallich has argued that price fluctuations may be the main source of export earning instability. Wallich, "Stabiliza— tion . . .," pp. p13,, pp. 345—46. ' 136 earnings than those designed to stabilize export prices. It may very well be that fluctuations in eXport volume and export earnings would 37 be minimized by stabilizing export prices. To determine whether a country or group of countries should prusue a policy designed to stabilize export volume or export prices is beyond the scope of this study. 'We are studying the stability of export earnings in regard to one solution—-export diversification. By diversifying exports, it is hoped that fluctuations in export prices (and volume) of individual products will cancel out to provide for greater stability of export earnings. As given above, the means of the average annual fluctuation in export earnings are 11.6 and 10.2 per cent, respectively, for 12 and 13. If export prices fluctuate 6.0 per cent annually and export vol— ume fluctuates 9.2 or 7.5 per cent depending on whether I2 or I3 is used as a measure, it appears that some of the instability in export prices and volume cancel out to provide for greater stability in export earnings. This implies that fluctuations in export earnings are due primarily to fluctuations in supply rather than demand.38 If the forty-five countries are divided into developed and under; develOped groups, export prices fluctuate 4.6 per cent in the former and 7.2 per cent in the latter. As we have seen, the difference, 2.6 percentage points, is significant at the .01 level. The means of the average year-to-year fluctuation of export volume of the developed 37For example, if current changes in volume are related to price changes in previous periods, stabilization of prices may be the best approach. 38Some of the instability may, of course, be due to fluctuations in demand. 137 countries are 9.5 and 6.8 per cent, respectively, for 12 and I3. For the underdeveloped countries, the corresponding means are 8.9 and 8.1 per cent. If the I2 means are considered, the difference is not sig— 39 nificant at the .10 level using a one-tailed test. However, if the I measure is considered, the difference is significant at the same 3 level.“0 As before, the relevant measure involves the I3 index. Hence, we conclude that both export prices and volume fluctuate more in underdeveloped countries than in develOped countries. If fluctuations in export earnings, prices, and volume of the developed and underdeveloped countries are compared, the results show greater instability in the underdeveloped countries but shed little light on the source of instability in export earnings. For the developed countries, the percentages are 8.4 for export earnings, 4.6 for export prices, and 6.8 for volume.”1 For the underdeveloped countries, the corresponding percentages are 10.9, 7.1, and 8.1. 39Spearman rank correlation coefficients compiled between the I2 series and series based on per capita gross domestic product in 1958, per capita gross domestic product in 1963, and per capita consumption in 1960 were all found to be insignificant at the .05 level using a one—tailed test. “01f the countries are ranked by per capita gross domestic prod- uct, the resulting rank correlation coefficients generally support the hypothesized relationship. The Spearman rank correlation coeffi- cient between the 13 series and the series based on per capita gross domestic product in 1958 is —.354. Based on a sample of forty—four countries, the coefficient is significant at the .05 level using a one-tailed test. The same result is obtained if the countries are ranked by per capita consumption in 1960. However, if the countries are ranked by per capita gross domestic product in 1963, the rank correlation coefficient is only —.222 which is not significant. However, the sample size has declined from forty-four to thirty-eight and this may account for the difference. 1Because of the trend factor, the trend-corrected measure, I3, is used in regard to export earnings and volume. 138 Summary Fluctuations in export earnings affect the domestic economy in at least three important ways. Fluctuations in export earnings may generate fluctuations in national income and employment. They may discourage investment in the export sector and result in a mis- allocation of resources. Finally, fluctuations in export earnings may cause balance of payments problems. As a consequence, instabil- ity may be harmful.42 In this regard, it was argued that underdevelOped countries may be subject to greater instability than the develOped countries be- cause their exports are more highly concentrated-—in both the commod— ity and geographical sense of the term--or because their exports consist primarily of agricultural products which are inherently unstable. The main conclusions in regard to international instabila ity are summarized below. In the case of export earnings, the means of the average year- to-year fluctuations (for a sample of seventy-eight countries) are 11.6 and 10.2 per cent, respectively, for 12 and 13.!43 Apparently fluctuations in export volume account for most of the instability since export volume (9.2 and 7.5 per cent, respectively, for 12 and I3) fluctuate more than export prices (6.0 per cent).44 42 As we have seen, MacBean has recently questioned the assump— tion that instability is harmful. In addition, see Sydney Caine, "Instability of Primary Product Prices--a Protest and a Pr0posa1," Economic Journal (September, 1954), pp. 610-14 and wallich, "Stabilization . . .,“ pp. pgp., pp. 347—48. u3For a summary of the results, including the commodity and income terms of trade, see Table 21. These percentages are based on a sample of forty—five countries. 139 upmuH mo o.oH o.oH mm m.a e.oH om a.m m.oH me magma maooaH opens Ho o.a H.a mm N.e H.e ON a.m w.m me mapme aoaeoaaoo e.a a.oH Hm o.w H.0H o~ m.m e.oH He masao> upoaaH H.m a.m mm . a.o m.a om m.a ~.m me masao> “Hosea H.e H.e mm ¢.e o.m oN e.e m.e me mmoaum uuoaaH H.a N.a mm m.e o.e om a.m 0.0 we mmuaum uuoaxm a.oH a.HH mm e.m o.HH om N.Oa e.aa ma aweaapmm “Hoaxm mH NH mofinudoou mH NH mmwuucooo mH NH mo Honaoz mo Hoaasz ucmu pom .coHuoouoon uaoo pom moHuucooo mmHHucsoo mmwuucooo meoHo>ma Hoacc¢ .cowumnuosHh mo pomoHo>opumpdD mwmuo>< Hwoca¢ Honaoz pmuoouuoo owmum>< one Heuoa pomuH onu mo apex NO one: «oma-mema .memau Hecoaumcwmuaa ea auaaaneumca mo apeaasm <--.H~ mamae 140 By dividing the seventy-eight countries into groups it was found that the export earnings of underdeveloped countries (11.7 and 10.9 per cent, respectively, for 12 and I3) fluctuated more than the earn— ings of developed countries (11.0 and 8.4 per cent, respectively, for I2 and I3), especially if the trend-corrected measure, I3, is con- sidered. Differences in the relative stability of export earnings between the developed and underdevelOped countries may be attributa- ble to export concentration or to differences in the composition of exports. After dividing the forty-five countries for which data are available into two groups, it was found that export prices tended to fluctuate more in underdeveloped countries (7.2 per cent) than in developed countries (4.6 per cent). Similarly, export volume appeared to fluctuate more in underdeveloped countries (8.9 and 8.1 per cent, respectively, for 12 and I3) than in developed countries (9.5 and 6.7 per cent, respectively, for I2 and I3), but only if the data are corrected for trend.u5 These results are also consistent with the hypotheses offered above. As we saw above, export volume appeared to fluctuate more than export prices. This is true also of import volume and import prices. The means of the average year-to-year fluctuations of import volume for a sample of forty-one countries are 10.4 and 8.8 per cent, respec- tively, for I2 and 13 as compared to 4.5 per cent for import prices (based on a sample of forty-five countries). As in the case of export volume, import volume evidently 45This case illustrates clearly how the relative degree of instability between the develOped and underdeveloped countries is affected by the choice of measure. 141 fluctuated more in underdeveloped countries (10.7 and 9.6 per cent, respectively, for I and I3) than developed countries (10.1 and 8.0 2 per cent, respectively, for 12 and 13). However, this is not the case in considering fluctuations in import prices; the import prices of the developed countries (5.0 per cent) appeared to fluctuate more than import prices of the underdeveloped countries (4.1 per cent). Because of the small difference between the two means--1ess than nine-tenths of a percentage point--not much importance is attached to the result. However, this result is not necessarily inconsistent with the hypotheses offered above because there is much less varia— tion among countries in regard to the degree of commodity or geo- graphic concentration of imports than in regard to the degree of commodity or geographic concentration of exports.b’6 On this basis, one would expect less variability between countries in regard to import price stability. Similarly, if countries "generalize" in imports, they are likely to import both agricultural and non- agricultural products; hence, there may be less variability between countries in regard to import price fluctuations on this basis.”7 In regard to the commodity terms of trade, part of the fluctu- ation of export (6.0 per cent) and import prices (4.5 per cent) ' apparently cancels out to provide for greater stability in the com— modity terms of trade (5.8 per cent). Since the commodity terms of trade are calculated by dividing the export price index by the import uéMichaely, pp. pip., pp. 13, 21. 47Also, the result can be explained if one assumes that fluctu- ations in the demand for imports on the part of develOped countries affect world prices more than fluctuations in the demand for imports by the underdevelOped countries. 142 price index, data to calculate instability indices of the commodity terms of trade are available for the same forty—five countries dis- cussed under the section on export and import prices. The commodity terms of trade fluctuated more in underdevelOped countries (7.1 per cent) than in develOped countries (4.1 per cent) by three percentage points. The income terms of trade fluctuated more than the commodity terms of trade. The means of the average year-to-year fluctuation of the income terms of trade for the forty-five countries just discussed are 10.5 and 8.9 per cent, respectively, for I2 and I3. Therefore, it can be concluded that fluctuations in export volume tend to accen- tuate fluctuations in the commodity terms of trade, thus providing for greater instability in the income terms of trade. Also, it appears that the income terms of trade fluctuate more in underdevel— Oped countries (10.6 and 10.0 per cent, respectively, for I2 and I3) than in developed countries (10.4 and 7.5 per cent for the corres- ponding percentages), particularly if the trend-corrected series, 13, is used. However, there are a number of important exceptions. A number of conclusions can be drawn from this chapter. First, the recorded degree of international economic instability depends, to some extent, upon the method of measurement. In almost all cases, use of the United Nations method produced results showing greater instability in international trade than if the trend-corrected method were used. Use of Michaely's method would have resulted in measures showing even greater instability; Where the two measures differed considerably, the relevant comparison involved the trend-corrected series. Second, fluctuations in export earnings are considerable for 143 some countries; such instability may be harmful. Third, in all but one case—-import prices-~fluctuations in international trade tended to be greater for underdeveloped countries than for developed countries. If instability is harmful, this means the problem is worse for underdevelOped countries than developed countries. Any number of explanations could be offered to explain the greater instability for underdevelOped countries. However, the re— sults obtained above are consistent with the three hypotheses offered at the beginning of the chapter. Hence, they warrant further inves- tigation. Before proceeding, however, let us note one important point. we have found that the stability of export earnings varied con- siderably from country to country.“8 Since the export earnings of some countries--such as the United Kingdom--appear quite stable, these countries may not be interested in export diversification (or any other scheme) to promote greater export earning stability.49 For countries--such as Iran--which have eXperienced considerable insta— bility in their export earnings, such policies may be quite attractive. However, the degree of export earning instability is only one factor which a country must consider before deciding whether to pursue a program of export diversification or, for that matter, any “awe will discuss the International Textile Agreement and other obstructions to free trade below. 49These countries may be more interested in increasing their eXport earnings. Since their export earnings are already relatively stable, the countries may experience an increase in instability if they eXport commodities which are relatively unstable. 144 program designed to reduce export earning instability. For example, it is possible that stability may be purchased only by a reduction in the present value of future export earnings. In the limiting case a country could reduce its instability to zero by eliminating exports completely. Hence, any domestic policy designed to stabilize export earnings must be examined to determine its impact on the present value of future export earnings as well as the stability of those earnings. It may be that the present value of the future export earning stream for a particular country will be maximized with the present degree of instability. Hence, that country would not wish to embark on a program of export diversification; instead, it may wish to concentrate on alleviating the effects of instability. Therefore, whether any particular country wishes to embark on a program of export diversification depends on that country's individual situation. No doubt some countries find the instability problem to be of little importance; nevertheless, other countries are undoubtedly interested in stabilizing their export earnings in order to prevent recurring balance of payments problems and to facilitate their economic devel- opment. Therefore, it is important that we determine whether diver- sification of exports leads to greater stability in export earnings, and whether there seems to be "some" degree of diversification necessary to achieve a given level of instability. CHAPTER V EXPORT CCNCENTRATICN AND FLUCTUATICNS IN EXPCRT EARNINGS Since we are testing whether instability of export earnings is related to the concentration of exports, we need some measure of export concentration. In regard to commodity concentration, this measure should show a smaller number if a country divides its re- sources more evenly among existing exports, or, given existing exports, produces an additional product for export, or some combin— ation. In this sense, diversification of a country's exports can be considered as a program designed to reduce its commodity corcentra— tion. In regard to geographic cncentration, the measure should show a smaller value if a country divides its exports more evenly among its trading partners, or, given the existing pattern of trade, pro- duces exports which are sent to new markets, or some combination. In this sense, diversification of a country's exports can be considered as a program designed to reduce its dependence on individual markets. One commonly used measure is the so—called Hirschman index, written C = ,’£(Xi/X)é, where g denotes the concentration index or ratio, Xi the value of exports of commodity i’in some specified year, and E the value of all exports in the same year.2 Thus, for example, _A 1The discussion below is in terms of commodity concentration but the same remarks apply to geographic concentration. 2Albert O. Hirschman, "The Paternity of an Index," American Eco— nomic Review, LIV (September, 1964), 761-62. Hirschman's bibliography 145 146 if a country eXports only a single commodity the concentration index, 3 C, is numerically equal to one. If a country eXports additional products, the concentration ratio will be reduced.“ Similarly, if a country producing a given number of products for export divides its resources more evenly among those products, the ratio will be re- duced.5 The concentration ratio will also be reduced if an addition— al product is produced for export and resources are re-allocated among existing exports provided that: 0asm as --- ma ss mm as sea cos ass ass saa ems saa ssa ass oma --- --- mosa-ossa essence sea ass ass ass mes sea sea as sea sea was as ms as ss Nsa sea essa-mssa saaseso ass mes ass ass ss ms sea oos ass saa Naa mos ss as mm --- --- assa-cmss saasso as as as as mm os sea ous «ms saa mma saa sea saa saa om ow essa-mssa seam assoc ms om ss mm as ms ooa saa ass saa ass mas mes ass oos mm ea assa-mssa assaoaoo saa «as as ms as as sea Nas sea was mes cos sea --- --- --- --- essa-Nmsa sesame u NSH £0 st saa ass saa ass «as sea was sea saa mas saa saa sea as as ss essa-sssa ssaao as sw ms as sea sea sea as ass «as sea as as ass «as ss mes essa-msss aoasso as as ss ass mos sea ass ass ass mes aoa ass sea as ms as as esss-wssa sesame --- --- ow aw sm as sea mm mm ass saa as ma ms ow as --- Nssa-sssa assesses sm ma as mm as as ass as ms ass mas as as ss sa me as ssss-ssss assess ss ms as as sea as sea sea ass as as as ass was ms sea sea esss-msss wssosssxsa naswmsom mes ass mos ass ass mes sea as as as as ss ass as sea saa ass esss-mssa saaas=< saa saa sea mes mes sea ass was was saa was mma ssa mma ems ass saa essa-mssa ssasaass< mes sea ass ass aaa saa sea as ss saa saa saa saa ass --- --- --- esss-smsa essaasma< ssss mesa sass sass ossa smsa wssa sass onsa mssa smsa mssa Nssa amsa ossa sass mess . assesses $9758 sea u mssa seems so mamas saasozzoo may aoammm mnonmm m<3uHmom mmH UZHMDQ mmHMHZDOU OSHuNHNHm mo MGmao ecu wcamaadsoo madam: cu saco madam dump oLHo sauce we used umpam .m.a.o pmsam> one use posaowmcmuu mum menu zoan3 oa huucsoo was possess ou poaaowscmuu son» paw o>onm pocoaucms moaaussoo moans ecu mo sum spoow doauommamau .n.o.w .mppdeo ”womamno weapons whom waapnaoxo .mosam> acauomwcmau .m.a.o .mmauucaoo wounu onu coca mnu mo musedaa mm poucsoo cuss pecans Eouw seasonsa .Auoxums copcoa was so pmmmommm mm possm>v mucosmap paw Ammomao>o osmm pawns do pouasmou assess mmoaw um possw>v uo>sam paw psow pom udooxo .mpcmwb no oxaacmwdmy .mmcox Scam uaxo mo used man up mossm> .mukOdaHn .Ammnmma msoaam> ..u .n .c0uwdanmm3v moaumaumum Hmaocmcam HmcoaumduoudH .pcsm assuodoz HmcoaumduoucH paw .Amosmma msoaam> .xaow 3ozv moaumaumum mo cauosaam hanudoz .mGOHumz pouadb .Ammsmma mooaum> .xaow BoZV moaumaumum mummy HmcoaumsuoucH mo xoonummw .mcoaumz pmuada "Eouw poumssoamom --- sea ass sea ass ss sea as sea as as ss ass sea as ss saa assa-sssa sa>sssosss aa aa sa as ss as sea sea as ss ss as as as as --- --- sass-oasa assssasess ass ass ass ass ass ass sea as as as as as os as as ass sea sass-sssa assess amuse: ass ass ass ass ass sea sea as os ss as os as as as as as sasa-sssa sesseas senses --- aa as sa sa as sea ass ass saa saa --- --- --- --- --- --- assa-sasa assess: aaa aaa aaa aaa saa ss sea aaa aaa saa saa saa aaa aaa oaa --- --- sass-oasa sashes --- as as as as ass sea as as saa ass saa ass ass ass aa --- assa-sssa osoa sea saa sea ass sea aoa sea as as as as sa --- --- --- --- --- essa-aasa sceaassa --- saa aaa sea ass ass sea aoa sea saa aaa --- --- --- --- --- --- assa-sasa ssacssasa sea ass ass ass ass aoa sea as as sea ass ass ss as aaa ass ss essa-sssa sessassasss ss ss ss ass ass ooa sea as ss ss ss ss ass aaa os as ass sasa-sssa assess --- as sea sea ass as sea ass aaa sea saa as --- --- --- --- --- assa-aasa asses ass oas ass as sea ass ass ss sea ss sea ass as aaa aaa aaa oaa sass-ssss aasss sass assa aass sass ossa sass sass aasa sass aasa sass aasa aasa sass sass sass sass cos n sass .massa so szssa wasnozzoo saa nmwwmmw smazaou UoDGHuGOUauH xHszmm< 201 .muu0dxo adsosuoco> HouOu «0 undo Hod m.mm sow poucsooom asmsouuod .moma cu omms 80am .muHdeo Enososuom do ksoaaucm women one mmoapca ossso> pom moans uuoaxo onHa .n ouocuOOw mom: .9 ouOCuoow mom .mxakcwwcmH ou haco kaddm mums onHw .moma .sm sonsouoa co po>aomsap mp3 soan3 pomammmmz paw mamoposm mo coaumumpmm umsuomm .poaama momauwema osu pom musedsa cmaddamdmm Hauou mo ucoo nod m.an ou pmusnoam msmcmm ou mupdem mmumum topaz: .saco mamdmm ou museums mouMum topaz: wow ma xopca moaud upOdEHm .poaaod momsuwema was How muaOQEa cmswmumoaz HMUOU mo ucoo Hod m.wm ou poudzoam mommamoaz cu muhdeo meadow vegan: .hsco mommamoaz ou mupdeo moumum wouasb sow ma xopca moaua uuomEHp vosdauGOOuuH anzmmm4 APPENDIX II ESTIMATED LEVELS OF "REAL" CCNSUNPTION PER HEAD IN 1960a (UNITED KINGDOM = 100) Country Index Rank Country Index Rank United States 140 1 Chile 27 26 Sweden 125 2 Yugoslavia 22 27 Canada 108 3 Cyprus 21 28 Australia 106 4 Greece 21 29 United Kingdom 100 5 Brazil 20 30 Switzerland 96 6 Colombia 20 31 New Zealand 95 7 Fed. of Rhodesia 18 32 Denmark 87 8 Malaya 18 33 West Germany 86 9 Mauritius 16 34 Norway 83 10 Turkey 16 35 France 75 11 Peru 3 36 Belgium 74 12 Morocco 13 37 Netherlands 73 13 Taiwan 12 38 Finland 67 14 Iran 12 39 Austria 66 15 Ceylon 9 40 Italy 53 16 Ecuador 8 41 Iceland 48 17 Ghana 8 42 Ireland 47 18 Thailand 6 43 Japan 46 19 India 5 44 Israel 45 20 Nigeria 4 45 South Africa 42 21 Indonesia 4 46 Argentina 39 22 Pakistan 4 47 Spain 34 23 Sudan 2 48 Venezuela 31 24 Ethiopia 1 49 Portugal 28 25 aFrom: 'Wilfred Beckerman and Robert Bacon, "International Com- parisons of Income Levels: Journal, LXXVI (September, 1966), 533. 202 A Suggested New Measure," Economic APPENDIX III A COMPARISON OF THE COMMODITY TERMS OF TRADE OF TWENTY DEVELOPED COUNTRIES IN 1948—1949 AND 1963-1964 Despite the evidence of improving terms of trade in some of the ccnnrtries over the 1948-1964 period, in only eight did the average of the last two years in the terms of trade index exceed that of the initial two years. In three countries there was virtually no change, while in the other nine countries, the average of the last two years was less than the average of the first two years. The results, shown in Table 31 below, are derived by calculating the average for the first and last two years in each country's index and then dividing the difference between the two by the average of the first two years. A plus (minus) sign indicates an increase (a decline) in the country's commodity terms of trade. TABLE 3l.--Percentage change in the commodity terms of trade of twenty developed countries: {1948-1949 to 1963-1964 Country Per Cent Country Per Cent Germany, West +51.4 Sweden - .5 Iceland +25.4 Canada — 1.5 New Zealand +16.7 United States - 2.8 Japan + 9.9 Italy - 3.7 United Kingdom + 8.3 Denmark - 4.1 Norway + 5.8 France - 5.0 Switzerland + 3.9 Netherlands - 5.9 Finland + 2.5 Austria - 9.0 Ireland - .5 Belgium-Luxembourg — 9.6 Israel - 5 Australia ~17.1 203 204 .At first glance, the results obtained by the least squares tech- riityuea and by comparing the average of the first and last two years in each country's index seems inconsistent. However, the indices of nuost; of the developed countries fell off sharply during the first years of the Korean War and the remainder of the period was dominated bar a :recovery from this sharp decline. Consequently, the least sqiurres technique indicates an improvement in the terms of trade of a rnndber'cfl'countries while a comparison of the first and last two ytxrrs in each index shows that a majority of the countries did not :regain.their earlier levels. The results obtained by comparing the average for the first and last two years in each country's terms of trade index are largely dependent on the choice of the period under consideration. For example, if the first two years of each index occurred during the Korean war, more countries would have registered an improvement in their terms of trade when compared to the last two years. Hence, this approach may be inapprOpriate under the circumstances. The results do, however, suggest that the least squares technique may also be inappropriate. APPENDIX IV A COMPARISON OF THE COMMODITY TERMS OF TRADE 0F THIRTY-ONE UNDERDEVELOPED CCUNTRIES' IN 1948—1949 AND 1963-1964 Toe1certain extent, the trends indicated by the least squares methminay be misleading. If the averages for the first and last years:h1each index are compared, only fifteen countries out of thirty-one actually had lower average terms of trade in the last two years of the period under consideration. The change in the commodity terms of trade, obtained in the same manner as for the developed countries in Appendix III, is shown.irl Table 32 ‘below. A minus (plus) sign indicates a decline (an increase) in the country's commodity terms of trade. TABLE 32.--Percentage change in the commodity terms of trade of thirty-one underdeveloped countries: 1948-1949 to 1963-1964 Country Per Cent Country Per Cent Pakistan -50.7 Ghana + .9 Ecuador -37.7 Costa Rica + 1.2 .Perui —28.8 Guatemala + 3.8 Philippines -27.l Cameroon + 4.5 Indonesia -25.9 Togo + 4.6 Venezuela -l9.l Nigeria + 7.6 Argentina —l5.0 Iran + 7.6 Spain -1l.4 Colombia +10.2 Ceylon — 8.3 Portugal +13.8 Jamaica - 8.1 Honduras +15.7 India - 6.9 Cyprus +18.9 South Africa - 6.6 Nicaragua +20.0 205 L5 206 TABLE 32 -- Continued Country Per Cent Country Per Cent 'Tugoslavia - 6.5 Greece +22.3 'Purkey' — 5.3 Chile +46.6 Panama. - 2.1 Mauritius +46.6 Brazil +97.6 The reason for the divergence between the results obtained by the least squares technique and by comparing the average of the first and last two years in each country's terms of trade index is similar to that of the develOped countries. During the first years of the Korean war, the terms of trade of many of the underdeveloped coun- tries improved sharply; the rest of the period was dominated by a decline from the peaks reached during the war. Hence, use of the least squares technique indicates that many of the countries experienced declining terms of trade for the period as a whole. In many cases, however, the terms of trade never reached a point as low as in the initial two years of each index. .. . ... Hld'l '11 v 14.41..“ fluid] ill. at..- a --- sea ass ass saa ssa ass ss as as saa as ea sa sa aa sa assa-ssss mamas ass ssa saa ass saa saa sea ss aa as aa ss sa oa aa --- --- essa-oasa ass: .samsaow ass ssa saa aaa ass saa sea as as ss as ea as ss as aa os essa-sssa mosses ssa sea ass saa aaa aaa ssa ss as ass ss ca as ss sa sa as essa-sssa sesasas --- --- as ass saa as sea ssa ss as aaa ssa --- --- --- --- --- assa-aasa easessss ass saa ass ass ass ass soa ssa as as as sa sa --- --- --- --- essa-aasa aoss>ass as --- saa ass as ssa saa ssa ass as as ass aa as os as --- --- asss-oass possess ass ass saa aaa saa saa sea as as as sa ss ss sa aa ss sa essa-sssa assasss aaa saa ass ass ass ass osa ass aaa aaa sss as as as sa --- --- essa-oasa saasao as as as as as as sea as as sea ssa es ss ea sa aa sa essa-sssa seas eases ass as as as ss ass sea aaa aaa aaa aaa saa ass ss ss aa ss essa-sssa essEans ssa aaa ss as as as sea as ea aa as as as --- --- --- --- essa-aasa ceases uwdwfiu saa ssa ass saa saa saa osa saa sea aaa ssa soa ass ss as sa ss essa-sssa sassu ass ss aaa ass ass sea soa ss sea saa sea as aa as aoa aa as essa-sssa soaasu sea saa saa saa saa ass ssa waOa ass as as as as aa ss ca ea essa-sssa assess --- --- ass as as sea oss ss as sa aa sa ss sa as as --- assa-sssa assasass sss ss as saa ass aaa sss ass sea ssa ass ss aa as as aa oa essa-sssa assess ass saa ass saa saa saa ssa ss sea as aa ca aa aa os as sa essa-sssa sasosssssa senawamm ssa aaa ass saa aaa sea osa soa ss aa as as as as as sa aa essa-sssa eaassa< ass aaa ass ass saa saa ssa sea aaa saa saa ass sea sea saa aaa sea essa-sssa «assesses osa ass saa ass «as aaa sea as as ss ass ass ea ooa --- --- --- essa-aasa seasassas assa assa assa assa sssa sass sass aass sass aasa sass aasa aasa sass sass sass sssa . . assssoo asazsoo ssa . sass amass so asses mzoozs saa mosses mnonmm M<3uHmom mmH QZHMDQ mMHMHZDOU mzHZumeHm mo MQ Nanmmm< 207 208 will -- aaa asa aaa aaa aaa aaa aaa aaa aaa ss aaa --- --- --- --- --- assa-aasa amass soa saa saa aaa saa aaa aaa as as as aaa aaa as saa as as ea assa-sasa casss aaa saa saa aaa saa aaa aaa aaa aaa ss as aa aa sa sa aa sa assa-sasa «sagas asses --- aaa saa aaa aaa aaa aaa saa aaa aaa saa aaa as --- --- --- --- assa-aasa asamsaoss asa aaa aaa aaa aaa aaa sea as as aaa as sa as as aa aa aa assa-sasa assaaaoa saa saa sea as saa aaa aaa os ss as as as ea sa aa aa sa assa-sasa saaassaaaas --- asa saa aaa aaa saa aaa aaa aaa aaa as as as as sa --- --- assa-oasa sums aaa ssa saa ss as aaa aaa aaa ss saa as aa as aa aa aa aa assa-sasa asssaaa asa aaa aaa aaa aaa saa aaa ss aaa as as ea aa as aa aa sa assa-sasa saaaoz --- aaa aaa aaa aaa saa aaa as aaa ss aaa ss as sa aa ss aa assa-sasa sagasaz aaa aaa aaa aa aa aaa aaa aaa aaa aaa aaa sa aa as sa sa sa assa-sasa sassaaaoaz aaa aaa aaa aaa aaa saa aaa aaa aaa aaa saa saa ss aaa aaa as as assa-sasa aaaamsa 3az aaa aaa aaa aaa aaa aaa aaa as ss as sa sa as aa aa sa aa assa-sasa ssasauaaaaz saa aaa aaa ss saa ss aaa as as as as aa --- --- --- --- --- assa-aasa oououoz -- aaa aaa saa sa aaa aaa aaa aaa as as as aa sa aa aa sa assa-sasa ssaaausmz aaa aaa aaa aaa aaa aaa aaa aaa saa saa sa sa as --- --- --- --- assa-aasa oasmamz --- aaa oaa saa aaa aaa ooa as as as aa --- --- --- --- --- --- aasa-aasa asscaa aaa aaa asa aaa aaa aaa aaa as sa as aa aa aa aa aa --- --- assa-oasa assaa --- aaa aaa aaa saa as aaa aaa as as as ea aa aa oa as as assa-sasa moasasa aaa aaa aoa asa aaa aaa saa ss aa sa aa aa sa aa aa aa aa assa-sasa saaaa asa aaa aaa asa aaa oaa aaa as sa sa aa aa aa aa aa --- --- assa-oasa assasa asa aaa aaa aaa aaa aaa aaa as as as as as sa as ea as sa assa-sasa scsasaa --- --- --- saa aaa aaa aaa aaa saa aaa aaa aaa saa aaa aaa --- --- aasa-oasa maaaaosca aaa saa saa saa aaa aaa oaa aaa aaa saa ss ss as aaa aaa --- --- assa-oasa «assa --- saa aaa aaa aaa as aaa as as as sa as aa as as --- --- assa-oasa scsaaoa aaa aaa aaa as as as aaa as sea as as sea asa as as aaa as assa-sasa sausage: --- --- aaa aaa aaa ss aaa saa saa aaa aaa aaa as as ss aa aa assa-sasa masaauasa aaa aaa aaa asa as ss aaa as aa aa as aa aa sa --- --- --- assa-aasa moasaa ¢ssa assa assa assa oasa sasa sasa aasa sasa masa «asa masa amsa amsa omsa assa maoa 00a u sasa .maoo wmazsoo InnuIn1unnuuuuuuuuuunuunuuuuuunuhwmmmww 7 aaaaaaoa..> aaazsass 209 mum umnuocm ou cmuummmcmuu cmnu cam m>onm vwcoaucma mmauucsoo mmunu mnu mo >cm mvoow .mzmamz mo scaumawwmm mnu mm czocx aaumauow wmumum cm>mam mLu wcamaumaoo mamamz Cu aasa haamm mumv mnHu .mmauucsou mmufiu mnu aasa muucm mo uuom umuam .m.a.o Umsam> mum usn wmuummmcmuu mum hmsu nuanB Cu zuucaoo msu mo muuoaaa mm vmucsoo ouca cmounm Scum vmuuoasa .Aumxuma aovcoq mnu co vmmmmmmm mm wmsam>v mucoamav vcm Ammmmum>o mHmm uamnu co vmnaammu uanoEm mmouw um vmsam>v Hm>Ham tam waow How unmoxm .mwcmw: no mxa%cmwcm9 .mhcmx Scum uaxm mo upon msu um wmsam> ceauomwcmuu .n.o.w .muuoaxm mmmmumsu wcawamH uuoa weavsaoxm .mmsam> GOauummamuu .w.w.u .muuoaaHn .Ammsmma maoaum> ..o .n .couwcanmmzv moaumauMum Hmaocmaam HmcoaumcumucH .vcsm kHMumcoz amcoaumcumUCH vfim .Ammsmma macaum> .xuow BmZV muaumaumum mo caumaasm hanucoz .mcoaumz vmuwca .Ammsmma msoaum> .xuo» 3mzv moaumaumum mvmuH amGOHumcumuca mo xoonummw .maoaumz vmuacs "Scum vmumasoamom --- saa saa saa oaa saa sea as aa aa sa as sa as ss ss aa aasa-sssa aa>aaaosss as as as as as as aaa sea as as aa sa oa aa aa --- --- sssa-oasa aaaassaca> asa aaa saa saa aaa ss aaa aaa aaa as as ss as aa as as as sasa-sssa saunas saaaa: saa aaa oaa aaa aaa aaa sea as as as ss sa aa aa as aa sa sasa-sssa aosscas saaaa: --- ss ss as as as sea aaa as as as --- --- --- --- --- --- aasa-sasa assess: --- saa aaa aaa saa asa sea as as asa aaa aaa aa saa aa sa --- aasa-sssa osoa saa oaa aaa aaa saa saa aaa saa ssa ss aa aa --- --- --- --- --- sasa-aasa scaaasaa --- ssa saa saa aaa aaa aaa as sea as as --- --- --- --- --- --- aasa-sasa saaaasama aaa aaa asa aaa aaa aaa sea as as as sa aa as sa as aa aa sasa-sssa aasaamuaass aaa aaa saa aaa aaa ssa sea as as as sa aa sa as sa sa . aa sasa-sssa amasss sasa aasa aasa aasa oasa sasa sasa aasa aasa aasa sasa aasa aasa aasa oasa sssa sssa aaa u sasa .snssa mo szsse mzooza may ammm>oo amezsoo 82mm 1", I), 7‘ aasaaaoa..> saazsssa 210 .muuomxm cmamanmcm> HmuOu mo ucmo uma m.ma Mom vmuasooum Bamaouumm .momH ou omma Eoum .muuoaxw Esmaouumm do hamuaucm woman mum mmoavaa wanao> vcm moana uuomxw wsHa .n muocuoom mmm: .n muocuoom wwm .mxaxcmwcme cu haao hammm wumv mnsw .mooa .Hm uwnEmoma co wm>HommHv me gowns wcmHmmM%z vcm wammwonm mo coaumumumm HmEuomw .VOaumm momauwsma mnu How muuomaa cmacmamcmm Hmuou mo uamo uma m.nm ou vmucdoam mamcmm ou muuomxm mmumum vmuacs .zaao mamamm cu manomxm mmumum vmuaca How ma xmccw moaum uuoaaHm .anumm momauwsma mnu How muuoaaa cmswmumuwz Hauou mo ucmo Hum n.wm Cu vmucsoam mswmumuaz cu muuomxm mmumum vmuacs .zaao mawmumUaz ou muuoaxm mmuMum vmuaca Mom ma xmoca moaua uuomaHv vmsawuaoouu> anzmmm< APPENDIX VI INSTABILITY OF IMPORT VOLUME Data in the form of import volume indices are available for forty-one countries accounting for 85.2 per cent of world imports. The results, shown in Table 33 below, indicate that there is considerable instability in the import volume series at least in com- parison with the import price series. The mean of the average year- to—year fluctuation, 12, of import volume is lO.h per cent while the mean of the trend—corrected average fluctuation, I3, is 8.8 per cent. The 12 series ranges from 16.# per cent (Australia) to 5.6 per cent (United Kingdom) and the 13 series ranges from 15.8 per cent (Argentina) to 3.9 per cent (Costa Rica). As in the case of export prices and volume, it appears that import volume fluctuates more than import prices (4.7 per cent) in the forty-one countries for which comparison is possible regardless of the method of measurement employed. On an individual basis, fluc- tnurtions, 12, in import volume exceed those of import prices in every countajr; if the data are corrected for trend, fluctuations in import volxuma exceed those of import prices in every country except one-- Norway. :the world total excludes Cuba, Indonesia, and the Sino-Soviet area. 2 , . . . . . . ’Tne Simple correlation coefficient between the two series is .875 'The standard deviation is 2.8 per cent for both series. 211 212 TABLE 33.--Two measures of instability of import volume in a sample of forty-one countries, 1948-1964 Average Year-to- Trend-Corrected Average Country Year Fluctuation, Year—toaYear Fluctuation, Per Cent Per Cent I2 I3 Argentinaa 15.9 15.8 Australia 16.4 13.2 Austria 11.0 9.6 Belgium-Luxembourg 7.4 5.5 Brazil 11.5 11.5 Canada 7.1 5.7 Ceylon 8.8 8.2 Chile 12.7 10.6 Colombia 11.8 11.7 Costa Rica 7.1 3.9 Cyprusb 10.2 8.3 Denmark 10.1 7.9 Finland 13.1 11.0 France 7.5 7.2 W. Germanyb 11.6 6.4 Ghanac 11.8 11.2 Greecea 11.0 7.5 Icelandd 11.1 8.0 Indiab 13.8 12.3 Ireland 7.8 7.2 Israelb 7.7 7.3 Italy 11.8 9.6 Jamaicac 7.9 6.7 Japanb 15.9 11.6 Mauritiusc 6.6 6.0 Netherlands 11.1 7.9 New Zealand 12.8 11.6 Nicaragua 12.0 11.0 Nigeria0 10.? 7.3 Norway 7.5 4.4 Panama 6.9 6.3 Peru 14.4 12.0 Philippines 11.8 11.7 Portugal 7.3 6.8 South Africa 10.1 9.8 Spain 11.5 14.0 Sweden 8.6 6.5 Switzerland 11.2 9.1 213 TABLE 33--COntinued Average Year-to— Trend-Corrected Average Country Year Fluctuation, Year-to-Year Fluctuation, Per Cent Per Cent I2 13 United Kingdom 5.6 4.7 United States 6.8 5.9 Yugoslaviac 10.5 9.9 a1951.1964 b1950-1964 01948-1963 d1950.1963 By dividing the forty-one countries into develOped and under- developed groups, it is found that import volume fluctuated more in underdeveloped countries than in developed countries.3 For the developed countries, the mean of the average year-to-year fluctua- tion, I2, of import volume is 10.1 per cent while the corresponding mean for the underdeveloped countries is 10.7, a difference of six— tenths of a per cent. If the trend-corrected measure, I3, is used, the mean for the developed countries is 8.0 per cent while the mean for the underdeveIOped countries is 9.6 per cent, a difference of 1.6 percentage points. In the first case, the difference is not signifi- cant at the .05 level using a one-tailed test. The greater 3If the countries are ranked by per capita gross domestic prod- uct and by fluctuations in import volume, the relationship is, for the most part,substantiated. The Spearman rank correlation coeffi- cient between the 1958 per capita gross domestic product series and the I series is -.293 and between the former and the I series is -.355. Based on a sample of forty countries, the coeff cients are significant at the .05 level using a one-tailed test. The same re- sults are obtained by ranking the countries by per capita consumption in 1960. However, if the countries are ranked by per capita gross domestic product in 1963, the rank correlation coefficients are not significantly different from zero at the .05 level using a one-tailed test. However, the sample size has declined from forty to thirty- four and this may account for the difference in results. 214 differential between the I3 means is accounted for by the strong upward trend of import volume in developed countries; as before, the relevant comparison involves the trend-corrected measure. The greater instability in import volume in the underdeveloped countries may be the result of the greater instability in the income terms of trade of the underdeveloped countries. BIBLIOGRAPHY ARTICLES AND PERIODICALS Adelman, Irma, and Morris, Cynthia Taft. "A Factor Analysis of the Interrelationship Between Social and Political Variables and Per Capita Gross National Product," anrterly Journal of Eco- nomics, LXXIX (November, 1965), SSS-78. Allen, R. G. D. "Index Numbers of Volume and Price," International Trade Statistics. eds. R. G. D. Allen and J. Edward Ely. New York: John Wiley and Sons, 1953, 186-211. Baer, Werner. "The Economics of Prebisch and ECLA,” Economic Develop- ment and Cultural:ghange, X (January, 1962), 169-82. Baldwin, Robert E. "Secular Movements in the Terms of Trade," Ameri- can Economic Review, Papers and Proceedings, XLV (May, 1955), 259-69. Beckerman, Wilfred, and Bacon, Robert. ”International Comparisons of Income Levels: A Suggested New Measure," Economic Journal, LXXVI (September, 1966), 519-36. Bhagwati, Jagdish. "Immiserizing Growth: A Geometrical Note," Review of Economic Studies, XXVI (June, 1958), 201-05. ________ "International Trade and Economic EXpansion," American Economic Review, XLVIII (December, 1958), 941-53. Bruton, Henry J. "Growth Models and Underdeveloped Countries," The Economics of Underdevelopment. eds. A. N. Agarwala and S. P. Singh. New York: Oxford University Press, 1963, 219—41. Caine, Sydney. "Instability of Primary Product Prices—-a Protest and a PrOposal," Economic Journal, LXIV (September, 1954), 610-14. Devons, Ely. "Statistics of United Kingdom Terms of Trade," The Man- chester School of Economic and Social_§tudies, XXII (September, 1954). 258-75- Dorrance, G. S. "The Income Terms of Trade," Review of Economic Studies, XVI (1948—49), 50-56. Ellsworth, P. T. "The Terms of Trade Between Primary Producing and Industrial Countries," Inter-American Economic Affairs, X (Summer, 1956), 47-65. 215 216 Fleming, J. Marcus, and Lovasy, Gertrud. "FUnd Policies and Proce- dures in Relation to the Compensatory Financing of Commodity Fluctuations," International Nonetary Fund Staff Papers, VIII (November, 1960), pp. 1-76. Haberler, Gottfried. "Terms of Trade and Economic Development," Economic Development for Latin America. ed. Howard Ellis. London: MacMillan and Company, 1963, 275—97. Hirschman, Albert C. "The Paternity of an Index," American Economic Review, LIV (September, 1964), 761-62. Imlah, Albert H. ”Real Values in British Foreign Trade, 1796-1953," Economic Elements in the Pax Britannica. Cambridge: Harvard University Press, 1958, 20-41. "The Terms of Trade of the United Kingdom, 1796-1913," Economic Elements in the Pax Britannica. Cambridge: Harvard University Press, 1958, 82-113. Johnson, Harry G. "The Economic Expansion and International Trade," International Trade and Economic Growth: Studies in Pure Theory. London: George Allen and Unwin, 1958, 65-94. . "Optimum Tariffs and Retaliation," International Trade and Economic Growth: Studies in Pure Theory. London: George Allen and Unwin, 1958, 31-61. Kindleberger, Charles P. "Terms of Trade for Primary Products," National Resources and International Development. ed. Marion Clawson. Baltimore: Johns Hepkins Press, 1964, 339-65. ”The Terms of Trade and Economic DevelOpment," Review of Economics and Statistics, XL (February, 1958), 72-85. Kreps, Theodore J. "Import and Export Prices in the United States and the Terms of International Trade, 1880-1914," QEarterly Journal of Economics, XL (August, 1926), 708-20. Lewis, W} A. "werld Production, Prices, and Trade, 1870-1960," Th3 Manchester School of Economic and Social Studies, XX (May, 1952) 9 105-38 0 MacBean, Alasdair I. "Causes of Excessive Fluctuations in Export Proceeds of Underdeveloped Countries," Oxford University Institute of Economics and Statistics Bulletin, XXVI (November, 1964), 323-41. McLeod, A. N. "Trade and Investment in Underdeveloped Areas: A Comment," American Economic Review, XLI (June, 1951), 411-19. Martin, K. and Thackeray, F. G. "The Terms of Trade of Selected Countries, 1879-1938," Bulletin of the Oxford University Insti- tute ofgétatistics, X (November, 1948), 373-98. 217 Massell, Benton F. "Export Concentration and Fluctuations in Export Earnings: A Cross-Section Analysis," American Economic Review, LIV (March, 1964), 47-63. Mehta, Fredie A. "The Effects of Adverse Income Terms of Trade on the Secular Growth of Underdeveloped Countries,” Indian Economic Journal, IV (July, 1956), 222-40. Meier, Gerald M. ”The Terms of Trade and Economic Development: Comment," Review of Economics and Statistics, XL (February, 1958), 88-90. Morgan, Theodore. "The Long-run Terms of Trade Between Agriculture and Manufacturing," Economic Development and Cultural Change, VIII (October, 1959), l-21. "Trends in Terms of Trade, and Their Repercussion on Pri- mary Producers," International Trade Theory in a Developing Werld. ed. Roy Harrod. New York: St. Martin's Press, 1963, 52-95- Mundell, R. A. "The Pure Theory of International Trade,” Aggrican Economic Review, L (March, 1960), 67-110. North, Douglas C. "Ocean Freight Rates and Economic Deve10pment," Journal of Economic History, XVIII (December, 1958), 537-55. Pesek, Boris P. "Economic Growth and its Measurement," Economic Development and Cultural Change, IX (April, 1961), 293-315. Prebisch, Raul. "Commercial Policy in the Underdeveloped Countries," American Economic Review, Papers and Proceedings, XLIX (May, 1959). 251-73- Robertson, Dennis H. "The Terms of Trade," International Social Sgience Bulletin, III (Spring, 1951), 28-33. Rosenbluth, Gideon. "Measures of Concentration," Business Concen- tration and Price Policy. Princeton: Princeton University Press, 1955. 57-94. Rostow, W} W. "The Historical Analysis of the Terms of Trade," Economic History Review, IV (1951), 53-76. "The Terms of Trade in Theory and Practice," Economic History Review, III (1950), 1-20. Schiff, Erich. "Direct Investments, Terms of Trade, and Balance of Payments," Quarterly Journal of Economics, LVI (February, 1952), 307-20. Singer, Hans W. "The Distribution of Gains Between Investing and Borrowing Countries," American Economic Review, Papers and Pro- ceedings, XL (May, 1950), 473-85. 218 "The Terms of Trade and Economic Development: Comment," Review of Economics and Statistics, XL (February, 1958), 85-88. Staehle, Hans. "Some Notes on the Terms of Trade," International Social Science Bulletin, III (Spring, 1951), 33-37. wallich, Henry C. "Stabilization of Proceeds from Raw Materials Exports," Economic Development for Latin America. ed. Howard Ellis. London: MacMillan, 1963, 342-61. "Underdeveloped Countries and the International Monetary Mechanism," Money, Trade, and Economic Growth. New York: MacMillan, 1951, 15-32. 'Wright, C. M. ”Convertibility and Triangular Trade as Safeguards Against Economic Depression," Economic Journal, LXV (September, 1955). 422-35- BOOKS Adelman, Irma. Theories of Economic Growth and Development. Stanford, California: Stanford University Press, 1961. Atallah, M. K. The Terms of Trade Between Agricultural and Indus- trial Products. Rotterdam: Netherlands Economic Institute, 1958. Cairncross, A. K. Home and Foreign Investment, 1870-1913. Cam- bridge, England: Cambridge University Press, 1953. Caves, Richard E. Trade and Economic Structure: Models and Methods. Cambridge: Harvard University Press, 1960. Clement, M. 0., Qt g1, Theoretical Issues in International Economics Boston: Houghton Mifflin Company, 1967. Coppock, Joseph D. International Economic Instability. New York: MCGraw-Hill, 1962. Currie, Lauchlin. Accelerating Development: The Necessity and the Means. New York: McGraw-Hill, 1966. Haberler, Gottfried. A Survey of International Trade Theory. lst ed. rev. Princeton, New Jersey: Princeton University Press, 1961. The Theory of International Trade. Trans. Alfred Stonier and Frederic Benham. New York: The MacMillan Company, 1936. Higgins, Benjamin. Economicgpevelopment. New York: W2 W3 Norton and Company. 1959- 219 Hirschman, Albert 0. National Power and the Structure of Foreign Trade. Berkeley: University of California Press, 1945. The Strategy of Economic Development. New Haven: Yale University Press, 1958. Horowitz, David. Hemispheres North and South. Baltimore: The Johns Hopkins Press, 1966: Johnston, J. Econometric Methods. New York: McGraw-Hill, 1963. Kemp, Murray C. The Pure Theoryyof International Trade. Englewood Cliffs, New Jersey: Prentice-Hall, 1964. Kindleberger, Charles P. Economic Development. 2nd. ed. New York: McGraw-Hill, 1965. Foreign Trade and the National Economy. New Haven: Yale University Press, 1962. . International Economics. 3rd. ed. Homewood, Illinois: Richard D. Irwin, 1963. The Terms of Trade: A Eurppean Case Study. New York: Technology Press of Massachusetts Institute of Technology and Wiley, 1956. Kuznets, Simon. Economic Growth. New York: The Free Press of Glencoe, 1961. . Six Lectures on Economic Growth. New York: The Free Press of Glencoe, 1959. Leibenstein, Harvey. Economic Backwardness and Economic Growth. New York: John Wiley and Sons, 1957. Lipsey, Robert E. Price and anntity Trends in the Foreign Trade of the United_§tates. Princeton, New Jersey: Princeton University Press, 1963. Lloyd, Peter. International Trade Problems of Small Nations. Durham, North Carolina: Duke University Press, 1968. MacBean, Alasdair I. EXport Instability and Economic Development. Cambridge: Harvard University Press, 1966. Meade, James E. A Geometry of International Trade. London: George Allen and UnWin, 1952. Meier, Gerald M. International Trade and DevelOpment. New York: Harper and Row, 1963. Meyer, F. V. The Terms of Trade. COpenhagen: Munksgaard, 1962. 220 iichaely, Michael. Concentration in International Trade. Amsterdam: North-Holland Publishing Company, 1962. Pepelasis, Adamantios, gt 31. Economic Development. New York: Harper and Brothers, 1961. Royal Institute of International Affairs, New Directions for World Trade. London: Oxford University Press, 1964. Schote, Werner. British Overseas Trade from 1700 to the 19303. Trans. W; 0. Henderson and W. H. Chaloner. Oxford: Blackwell, 1952. Siegel, Sidney. Nonparametric Statistics. New York: McGraw-Hill, 1956. Sodersten, Bo. A Studyyof Economic Growth and International Trade. Stockholm: Almqvist and Wiksell, 1964. Stigler, George J. The Theoryyof Price. New York: The MacMillan Company, 1952. Taussig, Frank W. International Trade. New York: The MacNillan Company, 1927. Viner, Jacob. International Trade and Economic Development. Oxford: The Clarendon Press, 1953. Syndies in the Theory of International Trade. New York: Harper and Brothers, 1937. Wallich, Henry C. Monetary Problems of an Export Economy. Cambridge: Harvard University Press, 1950. PUBLIC DOCUMENTS International Monetary Fund. International Financial Statistics. ‘Washington, various issues. International Financial Statistics, Supplement to 1965/66 Issues. washington, 1965. League of Nations. Industrialization and Foreign Trade. Geneva, 1945. United Nations. "Commodity Trade and Policies in the Post-war Period," World Economic Survey. New York, 1959. . Instability in Export Markets of Under-DevelOped Countries. New York, 1952. Monthly Bulletin of Statistics. New York, various issues. 221 . Proceedings of the United Nations Conference on Trade and Development. New York, 1964. Volumes I-VIII. Relative Prices of Exports and Imports of Under-Developed Countries. Lake Success, 1949. Restrictive Business Practices. New York, 1953. . Standard International Trade Classification, Revised. New York, 1961. Sgpplement to the Menthly Bulletin of Statistics, Defini- tions and Explanatory Notes. New York, 1964. Yearbook of International Trade Statistics. New York, various issues. . Yearbookyof National Accounts Statistics; 1964. New York, 1965. United Nations, Economic Commission for Latin America. The Economic Development oprgtin America and its Principal Problems. Lake Success, 1950. Economic Suryey of Latin Aggricaj 1949. New York, 1951. United Nations, Report by the Secretary-General (Raul Prebisch) of the United Nations Conference on Trade and Development. Towards a New Trade Policy_£pr Development. New York, 1964. UNPUBLISHED MATERIAL Graham, Linda Lee. "An Analysis of an Index of International Trade Concentration." Unpublished master's thesis, Department of Economics, Oklahoma State University, 1965. Montgomery, Sarah S. "The Terms of Trade of Primary Products and Manufacturing Goods in International Trade, 1870-1952." Unpublished Ph. D. dissertation, University of Wisconsin, 1960.