PLACE IN RETURN BOX to temovo this Man! from your record. To AVOID FINES return on Of before due we. DATE DUE DATE DUE DATE DUE THE UNITED STATES TRADE RELATIONS WITH THE PACIFIC BASIN NEWLY INDUSTRIALIZING COUNTRIES BY Won Kwang Paik A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Political Science 1989 ABSTRACT THE UNITED STATES TRADE RELATIONS WITH THE PACIFIC BASIN NEWLY INDUSTRIALIZING COUNTRIES BY Won Kwang Paik The present study examines the Pacific Basin Newly Industrializing Countries’ (NICs’) import penetrations--those of Singapore, Korea, Hong Kong, and Taiwan--into the United States for the time period between 1967 and 1985. By following the three major theoretical perspectives of international political economy (the interdependence, dependency, and mercantilist perspectives), the study develops the models and the determinants of international trade. Each model of international trade is tested empirically to assess its adequacy in explaining the NICs’ penetration into the U.S. markets. The analyses of the present study show that the interdependence model, which focuses on comparative advantage, has the highest explanatory power among the individual models; confirming the conventional wisdom of international trade. More importantly, the analyses also show that each of the Pacific Basin NICs is profoundly different from one another. Each individual NIC has a distinct set of determinants for its success in penetrating the U.S. markets. The study concludes with policy implications for both the United States and the NICs which allow them to anticipate and adapt to a rapidly changing international trade system. For Kyoung and Nicholas, whom I have traveled together the entire path of this journey and will continue to travel together for many more to come. For my beloved parents, whom I miss the most to share the fruit of their labor. This dissertation is dedicated to them for their spiritual presence. ACKNOWLEDGEMENTS I owe considerable debts to many people for the completion of this project. To my major advisor, Professor Jack Knott, I cannot repay in full for his sage advises and countless encouragements. Jack always gave me the critical and professional feedbacks, and for these, I am eternally grateful. For the other members of my committee, Bob Jackman, David Jones, and Cleo Cherryholmes, who read earlier drafts of this manuscript, I thank them for their time, advice, patience, and support. I also owe much debts to Don Crone and Norm Graham who served on my committee briefly but ably. To Peter Koper who edited my earlier drafts, I am very grateful for his generous help. Bill Browne and Barb Ringquist who read earlier versions of this dissertation, I appreciate their helping hands not just as professionals but as friends. Most of all, I would like to acknowledge the many "unspoken" helps from my family. To my wife, Kyoung, I thank her for being the caretaker and the understanding partner of life. To my brother, Jonathan, I owe much thanks for his curious inquires of dissertation that motivated my dull conscience. iv TABLE OF CONTENTS LIST OF TABLES ................................................... viii LIST OF FIGURES .................................................. x LIST OF ACRONYMS ................................................. xi CHAPTER I. THE STUDY OF INTERNATIONAL TRADE RELATIONS ........... 1 Introduction ............................................ 1 Review of Literature .................................... 6 Interdependence Theory ............................. 6 Dependency Theory .................................. ll Mercantilist Theory ................................ 16 What is to be done? ..................................... 21 CHAPTER II. THE DETERMINANTS OF IMPORT PENETRATION .............. 24 Introduction ............................................ 24 The Trend of Import Penetration ......................... 25 The Causes of Import Penetration ........................ 39 Interdependence Hypotheses ......................... 39 Dependency Hypotheses .............................. 41 Mercantilist Hypotheses ............................ 44 CHAPTER III. A METHODOLOGY FOR TRADE RESEARCH ................... 48 Introduction ............................................ 48 Formulation of Trade Dispute Variable ................... 48 Types of Trade Disputes ............................ 49 Procedures of Concept Formation .................... 50 Methodological Concerns ................................. 55 Autocorrelation .................................... 57 Diagnosis of Autocorrelation ....................... 59 Corrections for Autocorrelation .................... 64 Decision Chart for Analysis ........................ 68 CHAPTER IV. OPERATIONALIZATION AND MODELS OF ANALYSIS ........... 71 Introduction ............................................ 71 Operationalization of the Variables ..................... 72 Models of Import Penetration ............................ 86 Interdependence Model .............................. 89 Dependency Model ................................... 90 Mercantilist Model ................................. 92 Composite Model .................................... 94 CHAPTER V. RESULTS OF THE ANALYSIS .............................. 96 Introduction ............................................ 96 Presentation and Interpretation of Results .............. 97 Descriptive Analysis Result ........................ 97 Unadjusted Ordinary Least Squares Estimation ....... 99 Visual Diagnostic Result ........................... 105 Estimation on the Adjusted Data .................... 108 A Comparative Analysis of Individual and Composite Model .............................................. 123 CHAPTER VI. CONCLUDING REMARKS .................................. 134 Introduction ............................................ 134 Suggestions of the Analysis ............................. 135 Policy Implications of the Study ........................ 141 For the United States .............................. 141 vi For the NICs ....................................... 145 Theoretical Implications of the Study ................... 147 Caveats and Prospects ................................... 153 APPENDIX A - UNITED STATES IMPORTS BY LEVELS OF PROCESSING ....... 157 APPENDIX B - A SIMULTANEOUS MODEL OF INTERNATIONAL TRADE ......... 159 APPENDIX C - CORRELOGRAMS OF AUTOCORRELATION AND PARTIAL AUTOCORRELATION FUNCTIONS ........................... 162 APPENDIX D - RESULTS OF COEFFICIENT ESTIMATES FOR THE INDIVIDUAL MODELS .............................................. 165 APPENDIX E - COEFFICIENT ESTIMATES FOR THE COMPOSITE MODEL ANALYSIS ............................................ 174 APPENDIX F - CHANGES IN COEFFICIENT ESTIMATES BETWEEN THE INDIVIDUAL AND COMPOSITE MODELS ..................... 186 APPENDIX G - RESULTS OF F-TESTS BETWEEN THE INDIVIDUAL AND COMPOSITE MODELS .................................... 190 APPENDIX H — UNITED STATES BUDGET AND TRADE DEFICITS ............. 192 BIBLIOGRAPHY ..................................................... 194 vii Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table .10 .11 LIST OF TABLES Annual Total U.S. Imports and Imports from the NICs, 1967-1985 ............................................ 26 Number of Petitions for Annual Totals and for Commodity Clusters by Filing Dates, 1967-1985 ........ 52 Number of Affirmative Petitions for Annual Totals and for Commodity Clusters by Filing Dates, 1967-1985.... 53 Percentage of Petitions for Six Commodity Clusters by Filing Dates, 1967-1985 ........................... 54 Descriptive Analysis Results for Select Variables.... 98 Expected Results and Percentage of Unadjusted OLS Estimates with Correct Signs for All Equations ....... 100 Percentage of Unadjusted OLS Estimates with Correct Signs at Regional and National Level of Analysis ..... 101 Percentage of Unadjusted OLS Estimates with Correct Signs for Each NIC ................................... 102 Percentage of Unadjusted OLS Estimates with Correct Signs for Commodity Level of Analysis ................ 104 Diagnostic Results from Autocorrelation and Partial Autocorrelation Functions ............................ 106 Correct Equations for the Analysis Based on the Visual Diagnostics ................................... 107 Estimates from the Adjusted Data with Correct Signs for Overall, Regional and National Level ............. 108 Estimates from the Adjusted Data with Correct Signs for Commodity Level of Analysis ...................... 110 Coefficient Estimates for the Interdependence Equations at Regional and National Level ............. 112 Coefficient Estimates for the Interdependence Equations at Commodity Level of Analysis ............. 113 viii Table Table Table Table Table Table Table Table Table Table .12 .13 .14 .15 .16 .17 .18 .19 .20 .21 Coefficient Estimates for the Dependence Equations at Regional and National Level ....................... 114 Coefficient Estimates for the Dependence Equations at Commodity Level of Analysis ....................... 115 Coefficient Estimates for the Mercantilist Equations at Regional and National Level ....................... 117 Coefficient Estimates for the Mercantilist Equations at Commodity Level of Analysis ....................... 118 A Comparative Evaluation of Individual Models for Their Predictive Performance for Various Types of Analysis ............................................. 120 A Comparative Evaluation of Individual Models From the Composite Analyses for Their Predictive Performance for Various Types of Analysis ............ 124 Changes in the Signs of Coefficient Estimates Between Individual and Composite Analysis .................... 127 Changes in the Signs of Coefficient Estimates Between Individual and Composite Analysis for Each Variable.. 127 Absolute Differences of Coefficient Estimates Between the Individual and Composite Model ........... 129 Absolute Differences of Coefficient Estimates Between the Individual and Composite Model for Each Variable ........................................ 131 ix Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure LIST OF FIGURES NICs’ Total Imports Into the United States as a Percent of Total U.S. Imports, 1967-1985 ............. Each NIC’s Total Import as a Percent of Total U.S. Imports, 1967-1985 ................................... NICs’ Imports for the Six Levels of Commodities, 1967-1985 ............................................ Each NIC’s Imports for the First Level Commodities, 1967-1985 ............................................ Each NIC’s Imports for the Second Level Commodities, 1967-1985 ............................................ Each NIC’s Imports for the Third Level Commodities, 1967-1985 ............................................ Each NIC’s Imports for the Fourth Level Commodities, 1967-1985 ............................................ Each NIC’s Imports for the Fifth Level Commodities, 1967-1985 ............................................ Each NIC’s Imports for the Sixth Level Commodities, 1967-1985 ............................................ The Decision Tree Chart for the Analysis ............. The NICs’ Tendencies Toward Import Penetration Into the United States .................................... 27 29 30 32 33 34 35 36 37 69 139 LIST OF ACRONYMS Autocorrelation Autocorrelation Function First Order Autoregressive Process Kth Order Autoregressive Process Association of South-Eastern Asian Nations Comparative Advantage Domestic Capital Formation Dependency Estimated Generalized Least Squares United States Dollar Exchange Rate the General Agreement of Tariffs and Trade Government Expenditure United States Gross Nation Product Growth Rate Hong Kong Interdependence Republic of Korea Mercantilist Newly Industrializing Country Ordinary Least Squares Partial Autocorrelation Function Singapore Standard International Trade Classification Taiwan Trade Dispute Tariff Schedules of the United States Annotated United States Unemployment Rate United States International Trade Commission United States Total Aid xi CHAPTER I THE STUDY OF INTERNATIONAL TRADE RELATIONS Introduction For many years, international trade has been an essential part of international relations. Traditionally, economists rather than political scientists have paid the most attention to international trade and national policies. Recently, however, with the emergence of "new Protectionism" and the resurrection of international political economy as a legitimate field of international relations, international trade and trade relations have become the focal points of political science inquiries. Given its varied intellectual attractions, the central theme of international trade remains constant: What are the causes of trade relations? Specifically, what factors determine a flow of commodities from country A to country 8? There are basically three major types of answers to this question. The first type is a neo—classical economic explanation which originates in the writings of Adam Smith and David Ricardo. In this explanation, the amounts and types of commodities that are traded are not determined by any one individual or any single enterprise. Rather, exchanges are determined by the advantages and disadvantages which are naturally endowed to a particular producer. If a particular person, business firm, or country has a naturally endowed advantage in 2 producing a particular commodity, then that commodity will be exported to another place with a lesser amount of the endowed factor. The second type of explanation stems loosely from the writing of Karl Marx. Exchanges of commodities are not naturally determined, rather they are structurally distorted. Because of the dialectical nature of the international system, trade relations are bound by a nation’s status in the system as either an undeveloped or a developed economy. For an undeveloped nation, the nature and types of trade relations will be conditioned and defined by a developed trading partner. In short, trade decisions are made by developed nations and are determined by the conditions surrounding the international system. The third type is a political explanation which presupposes the self-interest of nation-states. Political interests, or more broadly speaking national interests, dictate the characteristics of trade relations between countries. What is to be gained or relinquished in trading or not trading with other countries? Would trade serve the national interests? Would it enhance national power through the elevation of national wealth? Based on such self-interested calculations, a nation will pursue trade policies to fulfill its politicized interests. If an exchange of particular commodities is within the domain of perceived national interests, then trade will be allowed. If not, the exchange will simply be interrupted. Although many attempts have been made to expand our knowledge of international trade, empirical analyses of these explanations, especially for the events of the recent decades, are very sparse and inadequately carried out. The overwhelming tendency is to treat the 3 phenomenon of international trade from a single, nomological dimension. Economic analyses, such as those of Bhagwati (1983), Leamer (1984), and Helpman and Krugman (1985) seek to explain international trade in terms of difference curves, demand functions, and price elasticities. At the same time, the political analyses of Goldstein (1986), Odell (1985), and Cline (1984) investigate the institutions of protection, the outcomes of trade conflicts, and the domestic causes of protectionism. The divergence between economic and political analysis is obvious and understandable. However, this study seeks to recognize the saliency of both economic and political factors in international trade. Increasingly, trade relations are becoming political events-- trade decisions are politically determined--coinciding with the perceived increase in trade frictions and a fear of trade wars. In an effort to lessen the gap between two disjointed disciplines, this study seeks to test empirically the three general explanations of international trade with reference to a specific situation: an increased import penetration from the Pacific Basin Newly Industrializing Countries (NICs) into the United States in the 19705 and 805. The main objective of the present study is to: (1) develop and test the hypotheses of international trade relations, especially between the United States and the Pacific Basin NICs; (2) evaluate the explanatory power of the hypotheses by identified in the hypotheses; (3) re-evaluate the soundness of each theoretical perspective; and (4) examine the plausibility of a new theory of international trade, based on the findings of the study. Given these intentions, the present study examines inter- dependence, dependency, and mercantilist theories of trade relations. Additionally, the study seeks to investigate these theories from the several levels of analysis. Import penetrations from the N105 into the United States will be investigated using the national total and the totals from the processing levels of commodities.1 Empirical tests will be performed at the aggregated regional level, the aggregated national level, and the clusters of commodities level. In so doing, I am seeking not only to evaluate each theory’s relevance at the national (macro) level, but also to determine its explanatory power at the commodity (micro) level of analysis. The study investigates trade relations between the Pacific Basin NICs, (i.e., Singapore, South Korea, Hong Kong, and Taiwan) and the United States. The Pacific Basin NICs were chosen for several reasons. First of all, the Pacific Basin is an economically dynamic region. Linder (1986z4) tells us that the Pacific Basin NICs "...spectacular growth shows greater resilience than that of the South European and Latin American NICs." Moreover, Cline (1984) maintains that, among the developing countries, South Korea, Hong Kong, Taiwan, and Singapore showed the most dramatic growth in shares in the market for manufactured imports in industrial countries. Clearly, the Pacific Basin NICs (sometimes called the Gang of Four or the Four lsittle Japans) are the most rapidly developing and the most actively traading economies among devel0ping countries. __ 1The six "levels of processing of exports" developed by Firebaugh and Bullock (1986) will be followed. This categorization of exports 3 based on the amount of technology and capital required to process CO'IIIIodities for exports. Exports are classified into three levels of pr1 mary products and three levels of manufactured goods. 5 Secondly, the Pacific Basin is a politically important region. It has an undeniable strategic importance. Sneider (1986:83) claims that ”the region is dependent for its security on United States military power....The United States, in turn, must constantly reinforce the credibility of its commitments and maintain a consistent involvement.“ Indeed, the United States has a fundamental commitment to the Pacific Basin to prevent hostile powers such as the Soviet Union and North Korea from expanding their control over the region and to maintain sufficient power to thwart any direct or indirect threat. Thirdly, although the Pacific Basin NICs share many common characteristics, there are also considerable differences among them. For instance, each N10 has a distinct historical origin of statehood.2 Each stems from individualized history, culture, and tradition. Each pursues different economic policies toward development. And currently, each NIC faces profoundly different domestic and internat- ional demands and conditions. In short, the Pacific Basin offers an interesting empirical referent for trade theory. It is economically vigorous, politically important, and sufficiently distinct in many ways. Analyzing import penetrations by the Pacific Basin NICs requires the consideration of the similarities and the dissimilarities within the region. It offers enough variation to be meaningful, but not so much that comparisons become extremely complicated and data collection difficult. 2Harris (1986: 59) affirms that although Hong Kong lacks a state 1" ‘the development sense, the government (of Hong Kong) does more than ‘t acknowledges, more than what officials call ‘sitting on our hand’. THEORETICA R VIEW In the field of international political economy, there are three major theoretical perspectives concerned with trade relations. They are, Interdependence (Liberalism), Dependency (Marxism), and Mercantilism (Realism). At the risk of over-simplification, I will summarize the main points of each perspective in an attempt to isolate the consistent and prevalent causes of import penetration. letereepengence Theory3 Interdependence theory maintains that an open trading system characterized by the free flow of goods with minimum political intervention (i.e., less protectionism) produces the most efficient economic system. Based on the liberal view of human nature, interdependence theory assumes that individuals are the center of creative and productive activity and that the state is an inefficient and restrictive body which curtails economic activity. Interdependence theory argues that the state should not regulate economic activity because state regulations interrupt economic growth, whereas a free market (trade) leads to innovation and growth through competition. Such a system of free trade, in turn, will produce 3The term "interdependence" was introduced in the eighteenth century by the physiocrats in France and by bourgeois critics of governmental economic policy in Great Britain. They were critical of Se115-interested, mercantilist policies pursued by their governments and introduced the term as an ideal to be achieved in a new liberal PO] itical and economic order. See, Morse (1976: 660-661) for more detailed description of this origin of the term. 7 maximum economic welfare for the trading nations and for the world as a whole. Gilpin (1985:279) notes that Liberals argue...the state should not interfere with economic transactions across national boundaries. Through free exchange of commodities, removal of restrictions on the flow of investment, and an international division of labor, everyone will benefit in the long run as a result of a more efficient utilization of the world’s scarce resources. Likewise, Walleri (1978:592-593) affirms that "in its original formulation, the liberal world view stressed the notions of laissez- faire capitalism, the rights of property, limited government, and Social Darwinism. Assuming perfect competition, the market would produce maximum economic growth, development, and general welfare." Waltz (1954:98-99) describes the main argument of interdependence theory as follows: 00 Michigan and Florida gain by trading freely the automobiles of the one for the oranges of the other? Or would Michigan be richer growing its own oranges under glass, instead of importing the produce of ”foreign” labor? The answer is obvious.... Each side gains from trade, whether between individuals, corporations, localities, or nations. Otherwise, no trade would take place. In a similar vein, Morse (1976:661) states that "the system of individual liberty in a free market would enable society as a whole to acffieve a higher level of material benefit than would be feasible under mercantilism." Strange (1985) makes a stronger assertion that "tJIe main tenet of liberal economics regarding international trade is that the less governments intervene to obstruct the flow of trade, the better.” Strange (1985) adds that as liberal policies toward foreign comDietitian are adopted, national welfare and global welfare will be better served. Based on the assumption that "an invisible hand" seems to be working in the area of economic policy as well as in the market place, interdependence theory promotes the doctrine of free trade. It argues that free trade allows the most effective allocation of resources in the production of goods and services, thus producing maximum national and global welfare. Walleri (1978:594) concludes that "combined with unrestricted movement of capital, free trade would maximize world production through optimal efficiency in resource allocation. Any maldistribution resulting from global economic transaction, conducted according to liberal principles, is considered to be only minor." Essentially, interdependence theory stems from the classical economic writings of Adam Smith and David Ricardo. In his Wealth of Nations (1776), Adam Smith had ridiculed the fear of trade by comparing nations to households. Smith (1776:424-425) argued that since every household finds it worthwhile to produce only some of its needs and to buy others from his neighboring producers, the same logic should apply to nations: It is the maxim of every prudent master of a family, never to attempt to make at home what it will cost him more to make than to buy. The taylor does not attempt to make his own shoes, but buys them from the shoemaker....What is prudence in the conduct of every private family, can scarce be folly in that of a great kingdom. If a foreign country can supply us with a commodity cheaper than we ourselves can make it, better buy it of them with some part of the produce of our own industry, employed in a way in which we have some advantage. Following Adam Smith’s liberal economic view, David Ricardo For“mulated the doctrine of comparative cost to explain international trade. The value of goods is determined in each country by the amount 9 of labor required to produce them, but this need not be the same in different countries. Labor may be more productive in one country than in another, so that different amounts of labor are required to produce given goods (Snider, 1967:33). Ricardo showed that the gains from trade accrue to both sides even when a country has no absolute advantage whatsoever. As long as the price ratio differs at all between countries in the absence of trade, every country will have a comparative advantage, an ability to produce some goods at lower relative cost disadvantage than other goods. The country should export these goods in exchange for some of the others. The notion of comparative advantage, as argued by David Ricardo, is formally presented as the Heckscher-Ohlin model in the international trade literature. In short, the model states that a country’s trading relations with the rest of the world depend on its endowments of productive factors, identified as land, labor and capital. Kindleberger and Lindert (1978:30) summarize the Heckscher-Ohlin explanation of trade patterns as follows: ...different goods require different factor production, and different countries have different relative factor endowments; therefore countries will tend to have comparative advantages in producing the goods that use their more abundant factors more intensively; for this reason each country will end up exporting its abundant-factor good in exchange for imported goods that use its scarce factor more intensively. More analytically, Leamer (1984:xiii) outlines the two major assumptions of the comparative advantage model: (1) there are factors 10 of production that are immobile between countries, and (2) these factors are used in different combinations to produce different goods. Accordingly, the model posits that a country will possess a comparative advantage in good X if the country is relatively well endowed with factors that are used intensively in the production of X. This law of comparative advantage views trade as essentially a way for countries to benefit from their differences. Because countries differ in climate, culture, skills, resources and so on, each country will have a comparative advantage in producing goods for which its particular character suits it. Consequently, the comparative advantage model tells us that a country will export the commodity that uses intensively its relative abundant factor. Countries with relatively large labor forces have a comparative advantage in the commodity that exhibits relatively large scale economies (Leamer, 1984:47). The principle of comparative advantage, Krugman (l986:7) maintains, "...leads one to expect to see trade dominated by exchanges that reflect the particular strengths of economies-~for instance, exports of manufactures by advanced countries and exports of raw materials by underdeveloped countries." In a similar vein, Brander (1986) contends that trade relations and patterns are determined by 1comparative advantage, and free markets are the best way of exploiting comparative advantage. As such, interdependence theory explains that filternational trade relations are stimulated by comparative advantages and that the system of free trade creates maximum economic growth and benefit. 11 e e heo Dependency theory argues that the world trading system is characterized by an unequal exchange of goods between developed and undeveloped countries. In the Marxist tradition, dependency theory focuses on the modes of production. Each mode of production defines a pair of opposed classes, a class of producers (proletariat) and a non- producing class (bourgeoisie) that exploits them. Marx posits that the modes of production determine the structure of society and state. More precisely, Marx argues that society is shaped by the dominant mode of production, and the relationship inherent in that mode will determine the role of the state. Marx articulates that capitalism is a stage in the continuous progression of human economic organization which is preceded by communalism and feudalism and will eventually be replaced with communism. Emerging from these basic notions, dependency theory articulates that a capitalist world economy (liberal or otherwise) will continue to breed underdevelopment in undeveloped countries (Walleri, 1978:589). In its original formulation, dependency theory claims that the economic and political institutions in underdeveloped countries are the product of the historical development of the world capitalist $.Ystem. This system is delineated by the economic and political domination of developed countries over undeveloped countries. The main argument of the dependency theory is that 12 The ‘metropolis’ or ‘core’ exploits the ‘satellites’ or ‘peripheral’ by direct extraction of profit or tribute, by ‘unequal’ exchange or through monopolistic control over trade. Underdevelopment is not a state of original backwardness, but is the result of the imposition of a particular pattern of specialization and exploitation in the periphery (Brewer, 1980:17). Similarly, Gilpin (1975bz44) affirms that The argument of dependencia [dependency] thesis is that the economic dependence of the underdeveloped periphery upon the developed core is responsible for the impoverishment of the former. Development and underdevelopment are simultaneous processes: the developed countries have progressed and have grown rich through exploiting the poor and making them poorer. Lacking true autonomy and being economically dependent upon the developed countries, the underdeveloped countries have suffered because the developed have a veto over their development. Dos Santos (1985:303) defines dependence as a situation in which the economy of certain countries is conditioned by the development and expansion of another economy to which the former is subjected. A dependent relation exists when some countries (the dominant, core ones) can expand and can be self-sustaining, while other countries (the dependent, peripheral ones) can do this only as a reflection of that expansion. Dos Santos (1985:304) says historic forms of dependence are conditioned by: (1) the basic form of the capitalist world economy which has its own law of development; (2) the type of economic relations dominant in the capitalist center and the warp in which the 'latter expand outward; and (3) the type of economic relations existing itiside the peripheral countries which are incorporated into the Situation of dependence within the network of international economic r91 ations generated by capitalist expansion. 13 Frank (1966:19) contends that ”the economic, political, social, and cultural institutions and relations we now observe there (in underdeveloped countries) are the products of the historical devel0pment of the capitalist system no less than are the seemingly more modern or capitalist features of the national metropolis of these underdeveloped countries." Moreover, Frank (1966) argues that, in the ‘metropolis-satellite’ structure, each of the satellites serves as an instrument to extract capital or economic surplus out of its own satellite and to channel part of this surplus to the world metropolis. In addition, each metropolis serves to impose and maintain the monopolistic structure and exploitative relationship of this system as long as it serves the interests of the metropolis which take advantage of this structure to promote their own development and enrichment of their ruling classes. The dependent relationship is characterized by the domination of big capital in the hegemonic centers and its expansion abroad through investment in the production of raw materials and agricultural products for consumption in the hegemonic center. A productive structure in the dependent countries is conditioned by the export of these products. And as such, the production is determined by demand from the hegemonic center and the internal productive structure is characterized by rigid specialization and monoculture in the entire region. In this vein, Dos Santos (1985:310) concludes, ...(a dependent economic system) reproduces a productive system whose development is limited by those world relations which necessarily lead to the development of only certain economic sectors, to trade under unequal condition, to domestic competition with international capital under unequal conditions, to the imposition of relations of super- exploitation of the domestic labor force with a view to 14 dividing the economic surplus thus generated between internal external forces of domination. Ultimately, dependency theory maintains that such relationships facilitate more dependent economies and backwardness in the peripheral countries. The development of dependent capitalism benefits very narrow sectors, encounters unyielding domestic obstacles to its continued economic growth (with respect to both internal and foreign markets), and leads to the progressive accumulation of balance of payment deficits, which in turn generate more dependence and more superexploitation. Dependent development, as the modified version of dependency theory, introduces the concept of the semi-periphery.4 This version views the international trading system as an exploitative system consisting of the core, semi-periphery, and periphery. Still, in such an exploitative trading system, power and decision are lodged in the urban financial and industrial cores and the lack of true autonomy and economic dependence are inevitable for the peripheries and the semi- peripheries. More importantly, the strategy of capital accumulation for a dependent state is conditioned by its relation to the international economy. Although the modified version of dependency theory recognizes the possibility of development conjoined with dependence, there are many 4The concept of semi-periphery originates from Wallerstein (1974, 1979). In his world system analysis, Wallerstein equates capitalism and world economy: "Capitalism and a world economy (that is, a Single division of labor but multiple polities) are obverse sides of the same coin." Although Wallerstein undertakes a holistic approach (few are likely to disagree), I shall limit the scope of this study to dependency theory since Wallerstein’s arguments have a great deal in COmnnon with those of dependency. 15 distortions associated with dependent development. Examining the distortions generated by the "tripod" alliance of multinational, state, and local capital in Brazil, Evans (1979:50) concludes that ”...(dependent development) excludes most of the national bourgeoisie from political participation just as it excludes the mass of the population.“ Similarly, Cardoso (1973) argues that dependent development will lead to (1) more dependent economies, as shown by increased foreign indebtedness, (2) income inequalities as the gap between rich and poor grows wider, and (3) authoritarian regimes since states are ”captured" by foreign capital which seeks controlled climates for investment. Modified dependency theory, as an economic explanation, argues that economic factors motivated by capitalist interests are the major cause of dependent relationships and rejects the significance of nation-state as an independent actor. Evans (1978) claims that the collaboration within the "tripod" alliance is based on a common respect for the logic of profitability. Obviously, dependency theory sees the world trading system not as a collection of individual nation states working together to foster mutual benefits, but as a structured capitalist system operating under the principles of unequal exchange of goods and benefits. Spero (1981:139) explains that Trade between North and South is a process of unequal exchange, as control of the international market by developed capitalist countries leads to a declining price for the raw materials produced by the South and a rising price for industrial products produced by the North. Thus the terms of trade of the international market are structured against the South. In addition, international trade encourages the South to concentrate on backward forms of production which prevent development. 16 Emphasizing "monopolistic" structures and "exploitative" relationships, dependency theory argues that the trade relations between the core and the peripheries are devised to serve the economic interests of the core countries. The major implication of dependency theory is that a dependent economy is conditioned and controlled by the development and expansion of core economies. In other words, the economic conditions and circumstances of core nations will dominate trade relations. Changes in the core economic conditions, as seen by improved or worsened general economic conditions, will be directly related to the import performance of developing countries. Moreover, as the consequences of dependent relationships, a dependent state will be associated with the development of only certain economic sectors, limited export commodities, unequal trade conditions, increased balance-of-payment deficits, and foreign indebtedness. Usmmtilism In contrast to interdependence and dependency theory, mercantilism emphasizes the primacy of politics over economic factors. Gilpin (1985: 286) boldly states that ”international economic relations are in reality political relations." Historically, mercantilism was a rationale for the policies of the dynamic ruling groups who formed the modern European nation-states. Morse (1976:661) tel ls us that mercantilism ...justified the elimination of internal tariffs and taxes that provided revenues for lesser feudal groups who were viewed as potential rivals. The dynamic rulers were 17 interested in creating effective customs unions within their newly formed states as a means of enhancing their own power and building up unified economic and political systems. Although mercantilism as an economic doctrine does not have a coherent body of literatures, the principal contention is that the nation state and the interplay of national interest (as distinct from corporate interest) are the primary determinants of the world economy (Gilpin, l975b:45). Gilpin (l975b:45) defines mercantilism as "the attempt of government to manipulate economic arrangements in order to maximize its own interest, whether or not this is at the expense of others." These interests may be related to domestic concerns (i.e., full employment, price stability) or to foreign policy (i.e., security, interdependence). Basically, according to mercantilism, each nation will pursue economic policies that reflect domestic economic needs and external political ambitions without much concern for the effects of these policies on other countries or on the international economic system as a whole. Specifically, mercantilism stresses the importance of maximizing power, which is identified with wealth. According to Morse (1976:661), this mercantilist desire is based on the assumptions that (I) wealth is an absolutely essential means to power, whether for security of for aggression; (2) power is essential or valuable as a means to the . acquisition or retention of wealth; (3) wealth and power are each proper ultimate ends of national policy; and 5Early mercantilist writers were not interested in constructing thewaretical explanations. Rather they were interested in attaining Spec:ific political and economic goals that would elevate the state’s powe r and wealth . 18 (4) there is long-run harmony between these ends, although in particular circumstances it may be necessary for a time to make economic sacrifices in the interest of military security and therefore also of long-term prosperity. Likewise, Gilpin (1985:289) asserts that mercantilism means the reciprocal and dynamic interaction in international relations of the pursuit of wealth and pursuit of power. In the short run, the distribution of power and the nature of the political system are major determinants of the framework within which wealth is produced and distributed. In the long run, however, shifts in economic efficiency and in the location of economic activity tend to undermine and transform the existing political system. Walleri (1978) notes that mercantilism emphasizes the role of the state in promoting economic growth and development rather than emphasizing the role of the self-. regulating market. In particular, the state should eliminate internal barriers to trade, subsidize the creation of infrastructure, and promote home industry and the export of manufactures. Analyzing the economic rationale of mercantilism, Hirschman (1980) elaborates the possible use of foreign trade as an instrument of national power policy.6 In their quest for wealth, mercantilists assume that a nation can increase its wealth only by decreasing the wealth of other nations. As such, Hirschman (1980:4-5) constructs a syllogism to demonstrate the connection established by the mercantilists between wealth and national power: 6See Krasner (1976) for a discussion of "state-power” theory and its relationship with the structure of international trade. Krasner countends that trade relations can only be understood within the corntext of the power and interests of states. 19 Meje: premiee: An increase of wealth of any country is an increase of its absolute power, and vice versa. Minor premjee: An increase of wealth of any country, if brought about by foreign trade, is necessarily a loss of wealth for other countries. Cenclgeien: An increase of wealth through foreign trade leads to an increase of power relative to that of other countries--precisely the political aim of mercantilist policy. Hirschman (1980) concludes that trade policies are molded by the interests of power policy and are implemented to enhance the state’s power position against the other states. By viewing trade as the means of accumulating wealth which is convertible to power, mercantilism focuses on the attempt of governments to manipulate economic arrangements in order to maximize their own interest. Accordingly, the state plays a central role in overseeing all facets of trade, with the ultimate purpose of building the power of the state. Economic doctrines of mercantilism are closely related to more traditional notions of realism. Broadly, mercantilism can be viewed as a subset of the realist paradigm, encouraging economic nationalism and a strong central government while sharing fundamental assumptions. Essentially, three assumptions are integral to the realist vision: (1) states as coherent units are the dominant actors in world politics; (2) force is a useful and effective instrument of policy; and (3) there is a hierarchy of issues in world politics, headed by questions of'military security (Keohane and Nye, 1989:23-24). Realists depict a world in which politics is continually characterized by active or potential conflict among states, with the usea of force possible at any time. Each state attempts to defend its ter-ritory and interest from real or perceived threats. In such a 20 system, political integration is minimized and lasts only as long as it serves the interests of particular states. Subsequently, according to Keohane and Nye (1989z6), "...(realism) insists that national security is the primary national goal and that in international politics security threats are permanent." The central theme of mercantilism is that although the economic and technical substructure partially determines and interacts with the political superstructure, political values and security interests are crucial determinants of international economic relations, including trade relations. In short, politics determines the framework of economic activity and channels it in directions which tend to serve the political objectives of dominant political groups and organizations (Gilpin, 1985). Furthermore, Keohane and Nye (1989) claim that although international economic activity appears to be nonpolitical, it does not mean that political power is unimportant. Indeed the effect of politics may be indirect; it may determine the relationship within which day-to-day processes take place. Gilpin (1985:286) unequivocally points out that "the primary actors in the international system are nation-states in pursuit of what they define as their national-interest." The mercantilist perspective implies that the structure of international trade is determined by the interests and power of states acting to maximize national goals. Theoretically, the state is viewed to be an autonomous actor and the objectives sought by the state are regarded as the national interests. With respect to trade relations, the state must appraise two distinct sets of national interests: the national security interest and the mercantilistic economic interest. 21 On the one hand, the state must examine its national security objectives in dealing with trade questions. For trade to be possible, there must be a convergence of the national security objectives. That is, two states must share some common ideas, some common world view or ideological commitment.7 Otherwise, trade relations will be held back to the minimum level. On the other hand, the state must also consider its mercantilistic economic objectives. What economic arrangements can it manipulate to maximize its interest? Which trade policy will bring forth the greatest benefit? The state, with the ultimate aim of elevating its wealth, will follow the trade policy which will reflect its domestic economic needs, i.e., full employment, sustained growth rate, price stability, etc. fleet is to Be Done? This survey of the literatures that attempts to explain international trade relations has demonstrated the multi-faceted nature of import penetration. Interdependence theory views import penetration simply as a economic phenomenon driven by the forces of comparative advantages. The theory implies that micro-economic factors, especially those naturally endowed production and labor factors have a significant cause of the NICs’ import penetration into the United States. Dependency theory argues that the macro economic 7This condition explains the inherent difficulty associated with (or the lack of) East-West trade relations. Clearly, it is not an economic problem, rather a divergence in the national security interests held by the United States and the Soviet Union which explains the current relations. 22 conditions and the circumstances surrounding the international capitalist system will determine trade relations. In other words, the external demands and conditions which surround the NICs will determine their rate of penetration into the U.S. market. Mercantilist theory considers import penetration as a result of shrewd political calculations master-minded by the “Prince-like" political actors. Convergence of the national security interest with the U.S. coupled with the mercantilistic export policies of the Pacific Basin NICs will the key determinants in explaining the NICs’ import penetration into the United States. At this point, it seems reasonable to provide an overview of the chapters that follow and the caveats of the present research to allow the readers some guidelines to navigate its research direction. First, the linkages between the theoretical perspectives and their hypotheses will be constructed inasmuch as the theories permit. For instance, interdependence theory is very precise in terms of its theoretical rigor, making a clear connection between the processing level and import penetration. Dependence and especially mercantilist theory, on the other hand, are considerably less precise, only vaguely suggesting that the relationship at the processing level of exports (perhaps) exists. Second, the operationalizations of key concepts will be performed to best approximate the theoretical constructs.8 Once identified, the indicators will be constructed based on the available data and the 8By construct, I mean the universal knowledge, a broader hypothetical entity when validity is involved. It is usually a term not directly observable but may be shaped by its observations. 23 appropriate means to execute the computations.9 Third, the estimates of the models will be checked for the various methodological concerns (i.e., trends and autocorrelations), to ensure unbiased parameters. Finally, the interpretations of the results and the implications of the study will be presented and elaborated. However, the findings of the present study are by no means universal, nor final. The present research deals with the Pacific Basin NICs within a specific time period. Its findings may or may not be generalized to other places or other time periods--those tests are remain to be done. 9For instance, I have employed Planperfect, a spread sheet program to compute import penetration, Dbase 111 Plus to tally trade disputes, Microsoft Chart and Quattro to generate various graphics, and SPSS/PC+ Version 2.0 to execute data transformations and the estimation procedures. CHAPTER II THE DETERMINANTS OF IMPORT PENETRATION Introduction The previous chapter argued that import penetration is a complex phenomenon, conceptually linked to various causal factors. One might distinguish these causal factors on the basis of their general orientations and the scope of analysis. First, some are primarily economic factors while others are predominantly political. Second, the causal factors have either international (external/macro) or domestic (internal/micro) emphasis. Taken together, causal factors can be categorized into four broad classifications: international economic, domestic economic, international political, and domestic political. This overall categorization will serve as the guidepost in formulating operational hypotheses within each theoretical perspective. A descriptive analysis of import penetration into the United States by the NICs, at different level of processing indicates that while the general imports of the United States have grown substantially, the NICs’ share of U.S. imports have increased at 24 25 faster rate.1 Moreover, the trends in levels of processing show that the NICs’ shares for the commodities which require medium processing skills (i.e., Levels 4 and 5 commodities) had the most impressive growth rates, whereas the lower processing goods (i.e., Level 1 and 2 commodities) produced the minimum increases. The Trend of Import Penetretjon The annual trends of import penetration display several interesting points. In terms of dollar values, the U.S. general imports have increased consistently. As indicated in Table 2.1, the gross value of U.S. general imports has gone up from 15 billion dollars in 1967 to 345 billion in 1985, more than a 20 fold increase. For the NICs, their share of U.S. imports sustained a faster rate of increase than that of U.S. imports as a whole. In fact, during the time period of the study, the NICs’ share went up by 91 times. Singapore displayed the highest rate of increase, registering a 533 fold increase, followed by Taiwan (188), Korea (164) and Hong Kong (31). The import penetration from the Pacific Basin NICs is a significant portion of U.S. imports and the N105 became increasingly important trade partners for the U.S. 1Appendix A shows a complete listing of a number of commodities which have been categorized into six levels of processing. The list is derived from Firebaugh and Bullock (1986) as mentioned earlier. My categorization, however, is slightly different since I employed U.S. Commerce data, while Firebaugh and Bullock utilized U.N. data. There were some tradeoffs in deciding which data set to use. The U.N data is simpler but did not have all the NICs (e.g., Taiwan), whereas the 11.56 commerce date is more comprehensive, thus more cumbersome to work wit.. 26 Table 2.1: Annual Total U.S. Imports and Imports from the NICs, 1967-1985 (in million 3) Year U.S. NICs SIG KOR HK TAW 1967 15339 428 8 61 272 87 1968 33114 1134 29 199 637 270 1969 36052 1549 55 291 815 388 1970 39963 1945 81 370 945 549 1971 45602 2406 136 462 991 817 1972 55555 3515 265 708 1249 1293 1973 69121 4646 459 971 1444 1772 1974 100972 5758 553 1460 1637 2108 1975 96940 5494 534 1442 1573 1946 1976 121793 8544 697 2440 2408 2999 1977 147022 10324 881 2883 2894 3666 1978 172025 13495 1103 3747 3474 5171 1979 206327 15421 1467 4047 4006 5901 1980 240834 17653 1920 4147 4736 6850 1981 261305 20732 2114 5141 5428 8049 1982 243952 22264 2195 5637 5540 8893 1983 258048 27614 2868 7148 6394 11204 1984 325726 36366 3979 9353 8266 14768 1985 345276 39066 4260 10013 8396 16396 Source: U.S. Department of Commerce, U.S. General Imports: World Area by Commodity Groupings. Measured as the percent of U.S. general imports, Figure 2.1 displays the NICs’ import penetration between 1967 and 1985. The NICs’ share, while controlling for the size of U.S. imports, has risen from 2.8 percent to 11.3 percent. In other words, since 1983, more than 10 percent of everything that we import originates from the "Gang of Four." This is the largest record of U.S. import penetrations shown by any group of NICs, including the Latin American and the South East Asian NICs. Furthermore, with the exceptions of 1974 and 1979, the NICs enjoyed increased import penetrations each year over 27 .......... v Y : V""""V' .oonno ,u-nono-uv...l...-\ ' ' o o . ° ' o o . . n . . . ' ' . o u c ' o o u ' I u a I - u u o n o ' u o . ' n n . ‘ - o n : . o . a . a n . o o a : a . u . . . . 0 e I . l - n . - . o n u . u - - . . . . . . o . . . . ‘ . z . . I . n . . ~ - c . . u . u g . o a o a . o o 0 . _ o o o . . . u u n . . . .......... . {raunhu-uu-o. u... . nu..."- . . - . I . r o C D . O . a 11‘ 1+4 19 19 19 19 19 19 1'9 19 19 19 19 19 19 19 19 1919 19 19 67 68 69 78 7172 73 74 75 7677 78 79 8881 82 83 84 85 YEAR Figure 2.1 NICs’ Total Imports Into the United States as a Percent of Total U.S. Imports, 1967-1985 28 the previous year’s total. The exceptions are probably associated with the global economic hardship caused by the OPEC oil crisis in 1974 and the severe economic recession experienced by the U.S. in 1979. As indicated by Figure 2.2, Hong Kong began as the largest trading partner among the NICs, only to be overtaken by Taiwan in 1972. Hong Kong and Korea competed for the number two spot during the late 19705, but Korea emerged as the solid second after 1982. Singapore, meanwhile, continues to be the smallest of the Four, probably because it is the smallest economy among the Pacific Basin NICs. During this time period, Taiwan recorded the most remarkable growth, elevating its U.S. market share from 0.6 percent to about 5 percent. The increases for the other NICs were: Korea, from 0.4 percent to 2.9 percent; Hong Kong, from 1.8 percent to 2.4 percent, and Singapore, from 0.05 percent to 1.2 percent. Hong Kong, which began as the top exporting NIC, had the lowest rate of growth of import penetration, eventually becoming a distant third among the Four. The data on levels of processing of exports (commodities that belong to the six levels of processing are given in Appendix A) provide some insights as to why. At the aggregate NICs level, Figure 2.3 displays that Level 1 and Level 2 commodities, which are agricultural and primary products, experienced the lowest growth rate of import penetration. Level 5 (medium-high processed products) demonstrated the most impressive growth, very closely followed by Level 4, which consist of labor- intensive commodities. Levels 3 and 6 commodities have had only 1narginal growth rates from the aggregate perspective. In all, the 29 +Hmmm 4-14 051131? : ; . z z '. ; 4—v4 DWIEJCO 'fké ............. g ............. émgg ..... gumumuénmnmé .......... g ........... Seq—Ian ODIN! . ' 5 i 5 t , «Hmmw wfimmm mfimhmw : «fimhm é ........... .1.,4 gm p.[m. 2 “Hmwm FE: {C Emil : ; “”74 m Phi...” "M .311 .......... .......... ........... ..L. H m r... 1.1-4 ' . ' «Hawhm -«-'4 U“ D-Ikl 4fimhfi 'm-rfi OW P~I£3 «4.44 CD ufl£T1 4-'* UW'JJCD ii" -u-v4 Ch untx- ED Figure 2.2 Each NIC’s Total Import as a Percent of Total U.S. Imports, 1967-1985 3O O4 mi mi 09 $3 131-1 L...) 07 MJ ICFUS :3 CO CD 1»ch CO 23 1114 Q3 nun-1 31!}: Figure 2.3 NICs’ Imports for 1967-1985 [-7H CD CD m1 mwmmm - 1- 11-4 13'“: 1:12] (“1'1 -T4 55 ODIN] ...HU'YCL'IIW'I "~14 UWCIJUD --v4 CO Cw OW .._vfi flhtk.flm -fimrupfl =-WH UH DWIJJ -Ho1h01 «amp-1r ...- H :11 n- m -wflmnm ~4- M ch 111-- ...... ----*- 1"‘1 0‘1 E"- 1253! ~- F‘l LT"! 1,13 12,71 IIIIIIII ;. - i H cm 1.1.11 17m ED the Six Levels of Commodities, fifl {I 'H Ed .1. '1‘ M 31 trends in levels of processing signify that much of the NICs’ import penetrations can be attributed to commodities with a medium level of processing, such as wearing apparel, footwear, steel, chemicals, plastic materials, and metal products, whereas commodities with a low level of processing, i.e., food products, leather goods, textile fibers, and fertilizers, did not enhance the N105’ trading position against the U.S. Figures 2.4 through 2.9 illustrate the individual NIC’s import penetrations, controlling for the levels of processing. It seems that Singapore is the only NIC exporting any significant amount of food products (which are Level 1 commodities) into the U.S. For Level 2, all four NICs showed a negligible amount of import penetrations, accounting for less than 2 percent of total penetrations in 1985. All four NICs had low rates of penetration for Level 3 commodities, each reporting them as no more than 5 percent of total exports. For Level 4 commodities, although only Hong Kong had an increased rate of penetration, the magnitude of import penetrations for the N105 is definitely noteworthy: Korea and Hong Kong had 42 percent, Taiwan had 34 percent, and Singapore registered 10 percent of total penetration. For Level 5 and 6 commodities, the trends show consistently increasing rates of penetrations. Among the Four, Taiwan showed the most dramatic growth rates for Level 5, while Singapore led in Level 6 commodities. The figure also indicates that the individual NICs rely heavily on Level 6, which are high-tech commodities, for their U.S. import penetrations. By 1985, Level 6 commodities accounted for the following percentages of total penetrations: Singapore (74 percent), Taiwan (46 percent), Hong Kong (43 percent), 255nm.mwmmmmmm,WWHWMMwmmmmnmwemwwm"mmmmmmmwumm 32 mi rm :3 mil EQEPUS 1 1""! 1m} 1:4? :3 Fl... :9 End Figure 2.4 Each NIC’s Imports for the First 0 I: 1967-1985 -...r“' ““74 DD 'I"‘"I v o o ' ' ............................................................. ‘I‘ ooooooooooooooooooooo un- .- u 0 ,4 Ll I I .1': r‘ 1 T I EH-F} ooooooooooooo eh} Icy-q 1 -~~ww OWIDJLO :'-14 :0 OO‘W‘ ,1... 1:11 1:131:11 2~+.r4 OH Oath] 3,44 unuxamm! E ------ -wH Uh OOIMD «Pimlfifi D~iTi 4L 1—4 :71 15- E13 z-1.44 g1 b.[m. -—quwm .....qt.. H Ch [5. mgr" : ”f .............. 9;J.q4 gm gnarl ..rf' .......... ;_._.:..F': .................... .....1.. 1-1 Cl": 1.4:! 1:12) . J... ' ”if 0M“... . ",4— ,..r+1r“‘ P“ I'm! ET"! [am PM“ no?" [314' Level Commodities, (21:: 4E hi he 33 N MT SIGPUSLZ -+ KORPUSLE HKPUSLZL «e anpus -::+::- ............ ............. ............. ............ ............. ............. ............. ............. ............. ......... . :32. 1'... UN. :0 Ln 3 3 : 1 I : : t : : : : ; : : ; - : - - .. ........................... .........................................'.............;...........................;.............:..........| p... H m m w c - . - ' . . ' . . a". 0 D ' o o I . O o . ‘ ' u ' ' a e ' n n i 1 : i 3 : : 1 : : 1 I : 3 i : : 3 : : : OOOOOOOOOOOO z OOOOOOOOOOOOO . OOOOOOOOOOOOO :...IIDIIDOOO: OOOOOOOOOOOOO . OOOOOOOOOOOOO ‘OCCOCOIOCOIOQ: OOOOOOOOOOOOO . OOOOOOOOOOOOO T ...... .I'..'. fl m m - - : - - : : . : : (I: O p . I o . ' o . . : : - : : - - ' '11] ............ ............. ............. ............ ............. ............. ............. ............. ........ "3).... H m m N ............ ............. ............. ............ ............. ............. ............. ............. ............. ...... . 3.1]. up: In a: H ............. ............. ............. ............. ............ ............. ............. ............ ..... 1.1.1.4 m m m K ............. ............. ............. ............. ............. ............ ............. ............. ............ 11' ........ (.33... fl 0‘", ['5- U7 :1. ............ ............. ............. ............. ............. ............ ............. ............. ............ 2.... ..... .u. H m p. m 3. ............ ............. ............. ............. ............. ............ ............. 3}. “a": ...... g ............ ....... H.156. 1.4 0"} r1... urn-1| 1 m 13: L11 rm Ln {:1 pm on r4 mm Figure 2.5 Each NIC’s Imports for the Second Level Commodities, 1967-1985 ‘t'EHR 34 .l—h da 1.: SIQPUSL “e KORPUSLB HEPUSLB «- TewPUSLB - +1’ - *4 Ch :0 U7 *4 CD CD 1F HUNCDIT’J #4 Ch CD 04 Haitian-I wd Ch to G: HUWI‘hU‘i *4 Ch I» DD *4 0‘ PM D- 14 Ch PM 00 14 Ch PM U9 '4 Ch Pu 1M *4 Oh Pb 61 44 Oh Pb 04 rd EPIPH v4 '4 Ch Pb 80 r# Eh an OW Hmmcm HU‘II-flf‘h to d: m; m '12:: 1:11:1“'~.r:: 11* N 1111 H H Figure 2.6 Each NIC’s Imports for the Third Level Commodities, 1967-1985 35 L HHPUSLé ~wmmm ~Hommw ~fimmm ~fimmm «Hmmw rfimmm Pfimhm numb-on ~Hmww “Hmhwfl ~fimhm ~Hmhw ~fiuwum wfimwwu rwfimfuw ammo: ~fimmm ~wmmm Hmww Figure 2.7 m Each NIC’s Imports for the Fourth Level Commodities, 1967-1985 [33:11} Q J“ 36 "wmlflfi 031$) - w an on w ~—*4 0“ DBL?) -~rfl Ch flJEHJ -« 1"‘l [TI m M! ‘ ' - . - . . '“ *4 m CD :33: g ............. w ............... Li? ........ ................ g ................ g ................ a“ ............ g ............... Twmwm : ,n : " ? wwmmm 5 z 5 wfVTTmQEWWNMCETTrhwmiqg ....... 'Lgfimw. ...... ”*4 UhtE-ru ................ ................ ................ 1... ............ ,' ................ ......... . ...... .u vi 0*, [~— ED ................ ................ ................ ................ ..... "h; .. ............ 3,; ............... ........ ...mm 1...] U"! r... m .41de ................. El.“ H m r... v.31 g ................ 3 .............. g ................ g ................ gmuumuméfixfiq: ..... é ................. g .......... .mwqu an ph.nj g ................ é ................ g ................ g ................ E ................ inggfyuimfix+quéu ......... Ix -q4 Ch p.|34 ................ ................ ................. ............ g!” ............. ......... l .......... .. H m m m L i i i i E- i : H_mk_mflw..q gm.¢3rm. MD ”‘3'" IN] £533 KC. all] “‘13" [mu] $533 "H‘. #- Figure 2.8 Each NIC’s Imports for the Fifth Level Commodities, 1967-1985 If: a: l‘ IL 5le IS; M 37 mmmnmaah a, mgmzmxx iv ,Mmmmmflm i: mmmnmwmw J- Each NIC’s Imports for the Sixth Level Commodities, 1967-1985 Figure 2.9 38 and Korea (39 percent). To recapitulate, first, the NICs’ overall import penetrations show higher rates of growth than that of U.S. imports as a whole. Second, at the regional level, the success of the NICs’ import penetrations can be attributed to the impressive growths in commodities with a medium level of processing, including wearing apparel, footwear, steel, and chemicals and much less so in high-tech or primary commodities. Third, for individual NICs, the trends indicate that: (l) the bulk of Singapore’s exports is composed of Level 6 commodities, (2) Korea relies mostly on Level 4 and 6 commodities for its U.S. import penetration, (3) Hong Kong, while dropping to the third spot among the Four, continues to concentrate on Level 4 and 6 commodities, and (4) Taiwan, which replaced Hong Kong in the top spot, focuses on Level 4, 5, and 6 commodities as its main stake in import penetrations. This analysis points out several interesting patterns of the NICs’ import penetration, but the questions still remain: Why is it the case that the NICs experienced increased import penetrations? Why are there variations in import penetrations at different levels of processing? And why do different NICs have different rates of penetration at different levels? To deal with these questions, causal factors, suggested by the competing theoretical perspectives will be isolated, in order to formulate the operational hypotheses of import penetrations. 39 Ihg Causes of Import Penetratipn Interdependence Hypotheses The Interdependence perspective assumes that the relations between states are fundamentally harmonious, not conflictual. Simply put, international trade is a positive-sum game; everyone gains, and no one needs to lose, from a proper ordering of trade relations. Interdependence emphasizes the significance of economic interests and actors. For example, business corporations and other economic interests are conceived as the significant actors in shaping the structure of the international trading system. In this dynamic system, the national interest is best served by generous and cooperative interactions among economic interests--"the pursuit of self—interest in a free, competitive economy achieves the greatest good for the greatest number in international no less than in the national society (Gilpin, 1985:279)." As such, interdependence contends that economic conditions, motivated by a free exchange of commodities and international division of labor, will determine the nature and level of trade relations. As explained by Adam Smith, David Ricardo, and Heckscher-Ohlin, at the national and micro-economic level, the competitiveness of actors will determine the direction and amount of trade flows. Trade will be dominated by exchanges that reflect the particular comparative advantages of given nation-states and also of given products. Destler (1986) confirms that "...in high-technology industries in which U.S. 40 comparative advantage continued to increase, the U.S. trade surplus grew from $15 billion in 1973 to $52 billion in 1980." In their final analysis, Kindleberger and Lindert (1978) affirm that "...most economists regard the Heckscher-Ohlin explanations of trade as broadly true.” Clearly, interdependence theory emphasizes domestic economic causes of import penetration. The theory contends that if a NIC manufacturer can produce a particular commodity at a cheaper cost thus possessing a comparative advantage over a U.S counterpart, we would expect the importation of that particular good to the U.S. market. Conversely, if a NIC’s comparative advantage decreases we would expect a decrease in import penetration in the U.S. Under the comparative advantage argument we would anticipate a positive relationship between a NIC's comparative advantage and the magnitude of U.S. import penetration. And furthermore, we would expect the relationship to be true at the commodity level of analysis. Delineated as such, the theory of interdependence suggests the following: Hypothesis la: A change in comparative advantage will be associated with a change in the level of import penetration. Hypothesis lb: An increase in NIC’s overall comparative advantage will be associated with an increase in the total U.S. import penetration, while a decrease in such advantage will be associated with a decrease in NIC’s total penetration. 41 Hypothesis 1c: An increase in NIC’s comparative advantage on a given level of processing will be associated with an increase in the magnitude of U.S. import penetrations for that export level, whereas, a decrease in such advantage will be associated with a decrease in import penetration. The interdependence hypotheses are based on the assumption of harmonious economic interactions among different actors which will eventually produce the most efficient allocation of resources. Trade relations, to be mutually benefiting, are dictated by the economic principles of comparative advantages under a system of free trade. If a country possesses comparative advantage in a particular good, then that country will export that particular good to other countries which are less competitive. These flows of trade are (or ought to be) carried out in purely economic, market conditions--divorced from political regulations or interventions. en n H t e es Dependency theory begins with the premise that economic relations are essentially conflictual. There is no underlying harmony; one actor’s gain is another’s loss. In such a "zero-sum" international system which consists of a core, semi-periphery, and periphery, economic interactions are motivated by the logic of profitability. There is a tendency toward disequilibrium in the system since power 42 and decisions are lodged in the financial and industrial core. Developing countries are conditioned by their relations to the international economy which is dominated by the developed countries. More specifically, economic conditions and performances of core countries will determine the nature and amount of trade relations. Improved economic conditions in the developed countries will be directly related to the economic performance of these countries. If the core countries such as the United States experience expansion of their financial and industrial outputs, the NICs would experience expansion of their trade relations with the core countries. Conversely, if the core countries suffer economic decline, the NICs would simultaneously encounter diminished trade relations with those core countries. Since dependent relationships exist within the broader, international capitalist system, general economic circumstances will have an undeniable effect on the NIC’s trade relations with the core countries. Focusing on the international and intra-national economic factors, dependence perspective suggests these hypotheses: Hypothesis 2a: A change in the United States’ overall economic conditions will cause a change in NIC’s overall import penetration into the United. States. Hypothesis 2b: An improvement in the overall economic condition of the United States will produce increased NIC’s import penetration of U.S. markets, 43 whereas a degenerating United States’s economy will decrease the NIC’s import penetration. Hypothesis 2c: A change in international economic circumstances will determine the amount of the NIC’s import penetration. Hypothesis 2d: An improved international financial situation in a NIC will bring forth increased import penetration to the U.S. market, while a worsening situation will be associated with decreased import penetration of the NIC. The dependency hypotheses of trade relations share common ground with the interdependence hypotheses. In explaining the causes of import penetrations, they are not directly competing hypotheses, rather they are complementary to one another. Both hold that trade relations are motivated by economic factors and are fueled by the notions of profitability. Nonetheless, the dependency hypotheses highlights international economic circumstance, whereas the interdependence hypotheses emphasize domestic economic causes. Specifically, since dependent economies are subjugated to the expansion or decline of dominant economies, NIC’s import penetrations to the United States will be determined by such conditions as the United States’ employment rate, growth rate, etc. Further, since a dependent economy is a prisoner of the international capitalist system, international financial and monetary circumstances will guide 44 the magnitude of trade relations. Although dependency theory pays special attention to the structural and economic effects of dependent relationships, my main intention here is to investigate the reasons for import penetrations. Consequently, in this study, I am reluctant to make a further inquiry into the effects or purported distortions of dependent trade relations.2 Hergantjlist Hypotheses Mercantilism claims that contrary to the attitude of liberalism and Marxism, international trade relations are in reality political relations. Huntington (1973) affirms that the prediction of the death of the nation-state by the interdependence theory is premature. He offers a contrary argument that, as the number and scope of trade activities increase, the value of one particular resource (i.e., power) almost exclusively under the control of a national government would also increase. Hhat factors are important in determining trade relations between the U.S and NICs from the mercantilist perspective? The foundation of the mercantilist description of trade relations is the state’s thirst for power and wealth. Gilpin (1985) points out that the distribution of power and the nature of the political system are the major 2As mentioned earlier, the effects (not the causes) of dependent trade relations include limited export commodities and unequal trade terms. The latter effect are often argued to be shown through the increased trade deficits and increased foreign indebtedness. Furthermore, some advocates (Evans, 1979; Cardoso, 1973) have contended that dependent relations will lead to increased domestic inequalities and authoritarian regimes. 45 determinants of the system under which wealth is produced and distributed. Since wealth can easily be converted to power, mercantilism attempts to manipulate economic arrangements and situations to maximize the state’s own interest. Unmistakably, the primary actors are the nation-states and the fundamental factor is the nation-state’s pursuit of national interests, especially those of national security interests.3 Driven by its quest for power and national interest, the United States will, first of all, manipulate the trade relations with the NICs to maximize its own security interest. If the United States has a crucial security interest with a NIC, the United States may allow certain import penetrations from that NIC in order to maintain and maximize its national security interests. All things being equal, the saliency of the United States’ security interests in a NIC will dictate the level of NIC’s import penetration. Additionally, the United States and the NICs must contemplate their mercantilistic economic interests. The United States’ economic interests as indicated by its import policies will have a significant input in regulating NIC’s import penetration. A more restrictive U.S. import policy will curtail the NIC’s import performances. A more liberal import policy, on the contrary, will boost their performances. Furthermore, the NIC’s export policies will have a certain undeniable effect on their import penetration into the United States. If a particular NIC pursues a "strong" export policy, we would expect an 3This is embedded in the traditional realist assumption. Realism assumes that there is a hierarchy of issues in international relations, headed by questions of military security. In other words, the "high” politics of military security dominates the ”low" politics of economic and social affairs. 46 increased import penetration, while a less robust export policy would produce a decreased penetration. He would expect a positive relationship between the United States’ security interest (i.e., "high“ politics and the primary national goal) and the level of NIC’s import penetration. Secondly, we would expect a significant relationship between the United States import policies (i.e., "low" politics, motivated by the mercantilistic notion of economic nationalism) and the NICs’ import penetration. Thirdly, we would also anticipate a moderately strong relationship between the NICs’ export policies and their import performances. The mercantilist hypothesis, for the trade relations between the United States and the Pacific Basin NICs, can be postulated as the following: Hypothesis 3a: A shift in national interest considerations will bring forth a shift in NICs import penetration. Hypothesis 3b: An increase in the U.S.’s security interest will be associated with an increase in NIC’s import penetration whereas a decrease in such interest will be associated with a decrease in import penetration. Hypothesis 3c: A more restrictive change in the United States’ import policy, as indicated by increased trade 47 disputes, will result in a decrease in NIC’s import penetrations, whereas a more liberal change in such policy, as shown by decreased disputes, will produce an increase in import penetration. Hypothesis 3d: A strengthening in NIC’s export policies will produce a positive change in its magnitude of import penetrations to the United States, while a deterioration in the export policies will produce a reduction in import penetration. The mercantilist hypotheses are based on the assumption that political factors are the primary determinants of international economic relations. The hypothesis 3b focuses on the national security interests and the hypotheses 3c and 3d concentrate on the mercantilistic economic interests of the states. It is assumed here that each nation state is motivated to maximize its national interests, whether or not this is at the expense of others. The mercantilist hypotheses that import penetration is caused by: (l) the convergence/divergence of national security interest between the United States and the NICs, (2) restrictive/liberal U.S. import policies, as indicated by an amount of trade disputes and (3) the strengthened/weakened NICs’ export policies as approximated by a change in governmental role in export industries toward the NICs’ penetration into the U.S. markets. CHAPTER III A METHODOLOGY FOR TRADE RESEARCH Introduction This chapter presents the methodological issues regarding trade research. More specifically, the present chapter discusses the formulation of new trade variables and the methodological concerns of autocorrelations, diagnostic tests for autocorrelations, and correction of autocorrelation. After discussing the concept formation and the concerns of autocorrelations, the decision chart for the present analysis will be presented and elaborated in order to provide a guideline in analyzing international trade relations. Formulation of Trade Dispute Variable Over the years, political analyses of international trade have focused, either, on trade policies or the institutions of protectionism. The former analyses, such as Cline (1984) and Baldwin (1987), have dealt with tariff and non-tariff barriers of international trade, whereas the latter analyses, including Goldstein (1986), Odell (1985), and Nivola (1986), have examined the institutions of trade policies. In analyzing the trade relations between the Pacific Basin NICs and the United States, the present 48 49 study offers a conceptualization of trade disputes, as a new indicator of U.S. import policy. The formation of trade disputes is based on the assumption that trade disputes are filed as petitions with to the United States International Trade Commission (USITC). Created by Act of Congress as the United States Tariff Commission in 1916, the name was changed to the United States International Trade Commission by the Congress under the Trade Act of 1974. The Commission’s major organizational mission is to ”...[provide] technical assistance and advice to the President, the Congress, other Government agencies, and the public on international trade issues (United States International Trade, 1980:l).' The Commission is an independent, quasi-judicial, fact- finding agency that determines whether U.S. industries are materially injured by imports that benefit from various unfair and illegal activities from abroad. T T d D' 3 Basically, there are four types of petitions filed against foreign competitors to the Commission: (1) adjustment assistance/escape clause, (2) unfair import practice, (3) anti- dumping, and (4) countervailing duties.l Petitions filed under adjustment assistance seek certain types of import relief, such as an increase in duties, establishment of quantitative restrictions, or specified types of adjustment 1These four types of petitions will be discussed more extensively in Chapter IV. 50 assistance. For unfair trade petitions, the Commission may issue orders, excluding the articles which employ unfair methods of competition or unfair acts, from the entry into the United States, or issue cease and desist orders. Petitions for anti-dumping and countervailing duties are formal complaints by U.S. industries that imports are being sold at less than fair value or being subsidized by exporting governments. Upon investigation, the Commission may issue a special dumping or countervailing duty, if the Commission’s determination is affirmative. Procedures of Concept Formation Petitions filed to the Commission from 1967 through 1985, aside from the four types mentioned above, are categorized in terms of: (1) date of filing, (2) date of settlement, (3) final outcome, and (4) level of processing. Each petition has been coded for its dates of filing and settlement. Since the present study deals with trade relations on an annual basis, both dates have certain meanings. The filing date of the petition signifies that a U.S. industry felt the need for certain adjustment, (i.e., import relief, desist order, or special duty), against foreign competition. A determination date, especially if the finding is affirmative, means that a specific corrective measure has been initiated against U.S. imports. Upon investigation, the Commission determines an affirmative or negative finding for each petition. An affirmative finding denotes that there is a justifiable and reasonable indication of injury, while 51 a negative finding means that the Commission finds no justifiable basis for its complaint. It is apparent that we need to distinguish those petitions which are reasonable, and the others which are not. As such, each petition is coded with respect to its final outcome--the Commission’s affirmative or negative determination.2 In addition, each petition is assigned a processing code, following the six clusters of commodities. For instance, a processing code of l is assigned for primary commodity, 4 for consumer commodity, and 6 for high-tech commodity. The purpose of this assignment is to allow the micro (commodity) level of analysis. Taken together, petitions filed with the U.S. International Trade Commission are categorized in terms of type of petition, dates filed and determined, a final determination, and a level of processing. This classification of petitions filed with the U.S. International Trade Commission constitutes the U.S. import policy, as expressed through trade disputes. Table 3.1 presents the total numbers of petitions, from 1967 through 1985, categorized by their filing dates. 2This is not, always, an easy task. Commission’s investigation process may take as little as 60 days or as long as 3-4 years. Often times, length of investigation is affected by court litigation (for licence agreement and trademark) and changes in Commissioners. 52 Table 3.1: Number of Petitions for Annual Totals and for Commodity Clusters by Filing Dates, 1967-85 YEAR TOT L1 L2 L3 L4 L5 L6 1967 10 l 0 3 4 2 0 1968 8 1 0 l 5 1 0 1969 13 2 0 l 4 4 2 1970 58 3 0 5 30 5 15 1971 109 3 1 24 49 9 23 1972 87 4 4 13 30 8 28 1973 104 ll 0 6 44 21 22 1974 65 2 3 6 29 7 18 1975 34 5 2 2 12 6 7 1976 30 2 2 6 5 7 8 1977 44 2 5 4 11 10 12 1978 69 5 3 9 22 19 11 1979 45 4 7 0 14 6 14 1980 143 39 4 5 52 13 29 1981 87 7 3 7 37 13 20 1982 263 7 5 4 212 15 20 1983 140 4 2 11 53 33 37 1984 168 10 4 4 94 30 26 1985 74 4 3 6 30 24 7 Tot 1551 116 48 117 737 233 299 (%) 100.0% 7.5% 3.1% 7.5% 47.5% 15.0% 19.3% Table 3.1 shows, for the time period between 1967 and 1985, a total of 1551 petitions were filed to the Commission as the formal complaints against foreign competitions. The largest number of petitions is filed during 1982, followed by 1984 and 1980. More specifically, from 1980 to 1985, the Commission received 875 petitions which is about 60 percent of all petitions. On the average, the‘ highest percent of complaints was directed toward Level 4 commodities (47.5 percent), which consist of wearing apparel, footwear, and chemicals, followed by Level 6 commodities (19.3 percent), which 53 include electrical, office, and telecommunication machines. Table 3.2 shows the numbers of affirmative petitions in terms of their filing dates for the same time period. Table 3.2: Number of Affirmative Petitions for Annual Totals and for Commodity Clusters by Filing Dates, 1967-1985 YEAR TOT L1 L2 L3 L4 L5 L6 1967 3 o o 1 1 1 o 1968 8 1 o 1 5 1 o 1969 9 o o o 4 3 2 1970 35 3 o 3 17 4 8 1971 34 3 o 14 8 3 5 1972 \21 4 1 4 16 2 14 1973 7 o 2 15 11 6 1974 21 o 1 4 9 2 5 1975 15 3 1 1 6 2 2 1976 12 2 1 2 2 2 3 1977 20 o 5 2 2 4 7 1978 35 4 3 1 14 9 4 1979 22 1 7 o 6 3 5 1980 34 3 1 2 l6 2 10 1981 42 3 3 1 23 5 7 1982 158 2 4 4 135 8 5 1983 83 1 2 4 46 15 15 1984 86 6 2 1 49 21 7 1985 so 4 o 1 26 15 4 Tot 749 47 31 48 400 113 110 (%) 100.0% 6.3% 4.1% 6.4% 53.4% 15.1% 14.7% Table 3.2 indicates that there was a total of 749 affirmative petitions between 1967 and 1985. This means that about half of all petitions (i.e., 48 percent) filed to the Commission for the time period have affirmative findings. The largest number of affirmative petitions is filed during 1982, followed by 1984 and 1983. Affirmative petitions during the time period from 1980 through 1985 54 account for more than 60 percent of all affirmative petitions. Once again, the bulk of affirmative petitions can be found for Level 4 (53.4 percent), Level 5 (15.1 percent), and Level 6 (14.7 percent) commodities. To examine the trend of petitions on an annual basis, Table 3.3 presents percentages of petitions broken down by the level of processing. Table 3.3: Percentage of Petitions for Six Commodity Cluster by Filing Dates, 1967-1985 YEAR L1 L2 L3 L4 L5 L6 1967 10.0% 0.0% 30.0% 40.0% 20.0% 0.0% 1968 12.5% 0.0% 12.5% 62.5% 12.5% 0.0% 1969 15.4% 0.0% 7.7% 30.8% 30.8% 15.4% 1970 5.2% 0.0% 8.6% 51.7% 8.6% 25.9% 1971 2.8% 0.9% 22.0% 45.0% 8.3% 21.1% 1972 4.6% 4.6% 14.9% 34.5% 9.2% 32.2% 1973 10.6% 0.0% 5.8% 42.3% 20.2% 21.2% 1974 3.1% 4.6% 9.2% 44.6% 10.8% 27.7% 1975 14.7% 5.9% 5.9% 35.3% 17.6% 20.6% 1976 6.7% 6.7% 20.0% 16.7% 23.3% 26.7% 1977 4.5% 11.4% 9.1% 25.0% 22.7% 27.3% 1978 7.2% 4.3% 13.0% 31.9% 27.5% 15.9% 1979 8.9% 15.6% 0.0% 31.1% 13.3% 31.1% 1980 27.3% 2.8% 3.5% 36.4% 9.1% 20.3% 1981 8.0% 3.4% 8.0% 42.5% 14.9% 23.0% 1982 2.7% 1.9% 1.5% 80.6% 5.7% 7.6% 1983 2.9% 1.4% 7.9% 37.9% 23.6% 26.4% 1984 6.0% 2.4% 2.4% 56.0% 17.9% 15.5% 1985 5.4% 4.1% 8.1% 40.5% 32.4% 9.5% Over the years, as shown in Table 3.3, the highest percent of petitions is submitted for Level 4 commodities. For instance, over 80 percent of all petitions filed during 1982 focused on Level 4 commodities. To a lesser extent, U.S. firms producing Levels 5 and 6 55 commodities have filed a considerable number of complaints against foreign competition. In fact, during 1976 and 1977, Level 6 commodities have the highest share of all petitions filed with the U.S. International Trade Commission. The trend of trade disputes suggests that U.S. industries that manufacture consumer goods (e.g., wearing apparel and footwear) and semi-processed products (e.g., chemical products and semi-processed steel) have filed the highest number of complaints against foreign competition. They have filed the most complaints and made the most demands for specific governmental interventions to curtail the penetrations of U.S. markets by foreign competitors. Methodological Concerns The true behavioral relationship for the multivariate model is assumed to be v, = 81x11 + Bzxiz + --- + ka,k + u, (3.1)3 where Y: is the dependent, or endogenous, variable for observation i; xik is the ith observation on the kth independent or exogenous variable; Bk is the coefficient for the kth variable or a constant slope for all i observation of the given variable Xk; and U1 is the stochastic element or random error for observation i. 3This equation is a generic representation of models which will be employed in Chapter IV--the hypotheses derived from the theoretical discussions will be converted into functional forms. 56 The value of the dependent variable, Y: is determined by adding a systematic component (lei1 + 82x12 + ~-- + BkXik) and a random component (U4). The only observable quantities are the values of the explanatory variables and the final outcome for the dependent variable. It is not possible to observe the random component or the behavioral parameters. From our observations of Y1 and Xi’s, we wish to estimate the parameters of 81, 32, and Bk. For the multivariate model, it is necessary to make the following assumptions about the restrictions on the model and the U1 in order to produce the best linear unbiased (BLU) estimator (Hanushek and Jackson, 1977:46-54). (l) | rjk l < 1.0 for all j # k, (2) Fixed X, (3) EM) = 0. (4) E(u,2) - U2 for all 1, and (5) E(Uin) = 0 for all i f j. Assumption (1) states that there is no exact linear combinations among the regressed exogenous variables (i.e., X11, X12,...Xik). Simply put, there is no perfect multicollinearity. Each of the exogenous variables have some considerable independent variation. Assumption (2) signifies that the values of xik are fixed and that only the random error terms, Ui’s vary from one sample to another. Ideally, we could obtain many different samples for the values of Xik and estimate the parameter estimates for each and every sample. Assumption (3) indicates that since the error term is a random variable, its specific value for each observation is determined strictly by chance. For a large number of Ui’s, the mean value or 57 population mean would be zero. In other words, the error terms associated with each observation over all replication of the experiment have a zero mean. Assumptions (4) and (5) pertain to the variances of error terms. The fourth assumption implies that all error terms have the same variance. This is the assumption of homoskedasticity, that one does not expect the error term to be greater for higher values of the exogenous variables than for lower values. The fifth assumption is that of no autocorrelation. All the error terms are assumed to be drawn independently of each other so that all the possible error terms associated with one observation are independent of, and thus uncorrelated with, the error terms of other observations. For ordinary least square estimates (OLS) to be unbiased, Assumptions (1) to (3) must be true, and for OLS to be the BLUE (Best Linear Unbiased Estimator), Assumptions (4) and (5) must be satisfied. In addition, we need to make the assumption that the disturbance or stochastic error term is normally distributed around zero in order to allow statistical inference such as hypothesis testing on estimated coefficients for their significance. Aptpgprpelptipn Time-series analyses, e.g., international trade analyses, are commonly associated with a violation of Assumption (5), where error terms are correlated across different observations. This violation denotes that errors are not independently distributed. That is, when 58 observations are made over time, the effect of the disturbance occurring at one period carries over into another period. Kmenta (1971:299) explains that Autocorrelation of the disturbances can be compared with the sound effect of tapping a musical string: while the sound is loudest at the time of impact, it does not stop immediately but lingers on for a time until it finally dies off. This may also be the characteristic of the disturbance, since its effect may linger for some time after its occurrence. But while the effect of one disturbance lingers on, other disturbances take place, as if the musical string were tapped over and over, sometimes harder than at other times. If the disturbances are autocorrelated, E(UtUt-s) # O (for some t > s.)4. This means that the disturbance occurring at time t is related to the disturbance occurring at (t-s). A correction for autocorrelation requires more precise information about the nature and behavior of disturbances. Most of the work on the correctional procedures of auto- correlation is based on the assumption that the regression disturbance follows a first-order autoregressive scheme, abbreviated as AR-l. If the disturbance of an observation (Ut) is correlated with the disturbance of the previous observation (Ut-1), we have Ut = pUt_1 + Et (for 311 t) (3.2) where p is an autocorrelation coefficient whose absolute value is less than one, and Et is a normally distributed random error with zero mean and constant variance. It is assumed that Et which is known as pure 4For relationships estimated from observations over time, the variables are given a subscript t (for time) rather than the subscript i that is used in the general case. 59 "white noise" is independent of Ut-1. That is, E(EtUt-1) - 0 for all t. This implies that each current disturbance (Ut) is equal to a portion of the preceding disturbance (Ut-1) plus a random effect represented by Et. Johnston (1984:310) notes that applying OLS to estimate a relationship with autocorrelated disturbances produces unbiased but inefficient estimation and yields invalid inference procedures. It has been shown that the least squares estimators are unbiased even when the disturbances are autoregressive but the estimators are no longer BLUE (Kmenta, 1971:302-303; Johnston, 1984:311). Likewise, it has also been shown that the least square estimators are not asymptotically efficient, although consistent. The estimated variances of OLS estimators, i.e., var(b), are biased and would cause a serious overestimation or underestimation of t statistics and significance levels in conventional inference procedures. To conclude, the consequences of autocorrelated disturbances are: (l) the least square estimators of the regression coefficients are unbiased and consistent but have no other desirable properties, and (2) the estimated variances of the least squares estimators are biased and conventionally calculated confidence intervals and tests of significance are not valid (Kmenta, 1971:311). Detectipn for Autpgprrglation It is a commonly accepted presumption that relationships estimated from observations over time often involve autocorrelated disturbances. Although there are several diagnostic procedures for 60 detecting autocorrelation,5 the two most common types are: (l) the Durbin-Hatson test and (2) visual diagnostics. The Durbin-Natson statistic is a straightforward diagnostic procedure for autocorrelation which makes use of the residuals from an OLS estimate of a model. It assumes that the residuals from the OLS estimation are consistent estimates of the underlying error terms and that they provide some indication of the structure of the error process in small samples. In the context of first order auto- correlation, we compute the value of the Durbin-Natson statistic (d) as d - SUM(et - et_1)2/ SUM(et)2, (3.3) where the et are residuals from OLS estimators. As the sample size becomes large, the different ranges of summation in the numerator and the denominator have a diminishing effect and d “ 2(1-p), where p is the autocorrelation parameter. Johnston (1984:315) tells us that the range of d is from O to 4: d < 2 for positive autocorrelation of the e’s; d > 2 for negative autocorrelation of the e’s; and d . 2 for zero autocorrelation for the e’s. Unfortunately, the distribution of the Durbin-Hatson statistic depends not only on the sample size and number of coefficients being estimated, but also upon the sample values of the explanatory 5For example, King’s locally optimal bounds test, the Berenblut- Hebb test, Geary’s sign change test, the Breusch-Godfrey test, and the Hallis test for fourth-order autocorrelation. See, Judge, et al. (1985:324-330) and Johnston (1984:317-321) for more complete discussions of these tests. 61 variables. Nevertheless, the distribution does have upper and lower bounds that are functions only of the sample size, the number of exogenous variables, and p. There are two major drawbacks in using the Durbin-Hatson statistic to test for the presence of autocorrelated disturbances. First, the test is valid strictly for nonstochastic explanatory variables. According to Johnston (1984:316), it is not applicable if there is a lagged dependent variable, and it can be shown that the combination of a lagged Y variable and a positively autocorrelated disturbance term will bias the Durbin-Hatson statistic upward and thus give misleading indications. Second, there is some ambiguity if the test statistic falls between the upper and lower bounds. The inconclusive range is an awkward problem and it becomes fairly large at low degrees of freedom, i.e., small sample sizes. One of the possible consequences is to accept the null hypothesis although it is false, which is a more serious problem than incorrectly assuming autocorrelation to be absent. Another test for the presence of autocorrelated disturbances involves visual diagnostics. Chatfield (l984:7) states that "anyone who tries to analyze a time series, without plotting it first, is asking for trouble." A typical visual diagnostic is to plot OLS residuals against time. If the residuals tend to run in certain streaks, it is necessary to analyze it further in more sophisticated manner by plotting the autocorrelation function which is known as the correlogram. The autocorrelation function (ACF) is a set of correlation coefficients between Ut and Ut_1, Ut and Ut_2, and so on, as a 62 function of the lag involved. As mentioned earlier, the AR-l autocorrelated disturbance is written as Ut - pUt_1 + Et, for all t, and this gives the autocorrelation function as pk a rk/ro, k = 1, 2,... (3.4) Since |p| < l and p>p2>p3, the autocorrelation function will decay exponentially. Furthermore, the correlogram will geometrically decline when pk is positive and it will oscillate in sign if pk is negative.6 In addition to the ACF, the partial autocorrelation function (PACF) is an important tool to describe the stochastic properties of autocorrelated disturbances. Judge et al. (1985:227) note that the kth partial autocorrelation coefficient measures the correlation between a stochastic process yt and Yt-k’ given Yt-l’---’Yt-k+l° In other words, the partial autocorrelation coefficient {rpk} measures the excess correlation at lag k which is not accounted by an AR (t-k) model and when plotted against the autocorrelation coefficient (pk) gives the partial autocorrelation function. Customarily, it is assume that the regression disturbance follows a first-order autoregressive process (AR-l). However, if we assumes an AR-l when in fact the disturbances have a more complicated structure, i.e., AR-Z, AR-3, then we may have estimates that are even 6See Appendix C for the examples of the correlograms for positive and negative autocorrelations. 63 less efficient than the OLS.7 Thus to determine the correct autoregressive process, it is necessary to examine and visually inspect the correlograms of the partial autocorrelation functions. Chatfield (1984:69) affirms that values of partial autocorrelation coefficients rp which are outside the two standard error limits are significantly different from zero. It can be shown that the PACF of an AR (p) process ‘cuts off’ at lag p so that the ‘correct’ order is assessed as that value of p beyond which the sample values of (rpj} are not significantly different from zero. This means that the partial autocorrelations have a cutoff point at lag p. Visually, the PACF correlogram will be truncated beyond the confidence intervals for the particular autoregressive process. For instance, if it truncates only at the first lag, then AR-l model may be appropriate. If it truncates at both the first and second lags, then the true model may be a combination of the first and second autoregressive processes. As the general guidelines for interpreting the correlograms, Chatfield (1984:25-30) describes the following tendencies of autocorrelation coefficients: (1) A random series--If a time series is completely random, then for large N, r 3 O for all non-zero values of k. This means that 19 out of 20 of the values of rk can be expected to lie between two standard error limits, (2) short-term correlation--This is characterized by a fairly large value of r followed by 2 or 3 more coefficients which, whi1e significantly greater than zero, tend to get successively smaller. Values of rk for longer lags tend to be approximately zero. 7Engle (1974) examines the specification of the disturbances for efficient estimation by comparing the estimates derived from autoregressive least squares (ALS) and OLS. He states that "only when ALS represents a minor misspecification of the true disturbance process is it necessarily better than OLS." Otherwise, OLS is superior and vastly more efficient than ALS. 64 (3) Alternating series--If a time series has a tendency to alternate, with successive observations on different sides of the overall mean, then the correlogram also tends to alternate. The value of r1 will be negative, whereas the value of r2 will be positive as observations at lag 2 will tend to be on the same side of the mean. (4) Non-stationary series--If a time series contains a trend, then the values of rk will not cage down to zero except for very large values of the lag. (5) Seasonal fluctuations-~If a time series contains a seasonal fluctuation, then the correlogram will also exhibit an oscillation at the same frequency. Chatfield (1984) cautions that a visual inspection of the correlogram is not an easy task--it takes some training and experience to detect the presence of autocorrelation and determine the correct order of autoregressive process. gprrgction for Autocorrelation To apply an appropriate estimation procedure for models with autocorrelated disturbance, we must (1) determine the order of autoregressive process, and (2) transform or adjust the data, if necessary. As we have discussed in the previous section, the correct specification of disturbances is an essential requirement for an efficient autoregressive estimator. Furthermore, it may be necessary to transform or adjust the data to remove any trend or seasonal fluctuations. If the evidence from the correlogram suggests that the data contain a seasonal trend, it is necessary to modify the series 8Broadly speaking, a time series is said to be stationary if there is no systematic change in mean (no trend), if there is not systematic change in variance, and if strictly periodic variation has been removed. 65 through a transformation or an adjustment. The common types of data transformations include logarithmic and square root transformation. As pointed out by Chatfield (1984:14- 16), there are three main reasons for making a transformation. The first is to stabilize the variance. If there is a trend in the series and the variances appears to increase with the mean then it may be necessary to transform the data. The second type is to make the seasonal effect additive. If there is a trend in the series and the size of the seasonal effect appears to increase with the mean then it may be advisable to transform the data so as to make the seasonal effect constant. And finally, it may be required to transform to make the data normally distributed. Adjustment procedures are usually performed on the data with long-term trend. If the data contain a long term trend, the values of autocorrelation coefficients will not come down to zero except for very large values of the lag. The data contain non-stationary series and one must remove this trend prior to running any analysis. One of the most common adjustment techniques for non-stationary series is to transform the data through "differencing." Differencing is a special type of filtering, which is particularly useful for removing a seasonal or annual trend. This procedure simply differences a given time series until it becomes stationary. As noted by Kmenta (1971), taking the first difference is sufficient to remove any trend.9 One rarely needs to take second or higher differences. After taking the first difference, the adjusted data should be checked again by 9Kmenta calls this the "method of first differences" to deal with the problem of autoregression in disturbances. See, pp. 321-322 for the econometric treatment of this procedure. 66 plotting the ACF and PACF to detect the presence and determine the order of autoregressive process, if any. Once the correct diagnostics are made concerning the auto- regressive process, we can apply a number of estimation procedures. The most widely accepted procedure is known as the generalized least squares (GLS) estimation, pioneered by Cochrane-Orcutt. For the first autoregressive process, suppose that the equation to be estimated is Yt I 30 + let + Ut, (3.5) where Ut - pUt-1 + Et. By multiplying this equation with the known value of p, we get th-1 = p80 + pBIXt-1 + pUt_1. (3.6) Subtracting the second equation from the first yields Yt - th-1 = 30(1-p) + 31(Xt - pXt_1) + (Ut - pUt-1), (3.7) and this can re-written as vt* - 80* + let* + vt. (3.8) Now Vt will no longer be autocorrelated and the least squares 67 estimators are best linear unbiased estimators.lo Hhen p is unknown, we need to estimate the value of p with the least squares residuals. The least squares estimator of p is given as p . SUM(etet_1)/ sun(et-1)2, (3.9) where et and et_1 denote OLS residuals. Applying OLS to equation (3.14) after replacing p with p is known as the Cochrane-Orcutt two- step method. Kmenta (1971:315-319) asserts that this procedure requires two successive applications of OLS: the first to obtain p and the second to obtain estimates of Bo and BI and their estimated standard error. If the two-step procedure is repeated until convergence, the resulting estimators of Bo and 81 are known as the Prais-Hinsten iterative estimators. These estimated general least squares (EGLS) can be shown to be best linear unbiased estimators and asymptotically consistent and efficient. To conclude about the EGLS estimation, Kmenta (1971:323) affirms that (1) unless the value of p is quite small (say, less than 0.3), the Prais-Hinsten estimators perform better than the OLS estimators, (2) generally, the iteration procedure helps in improving the performance of the estimators, and (3) the tests of significance, even when performed with transformed data, may be unreliable. 10This same transformation technique can be apply to the errors with a more complicated autocorrelation structure. For example, if the error terms have a second order autoregressive process so that Ut - p1Ut_ + p Ut- + E , then Xt must be transformed to Xt - plxt-1 - pzxt-2 (Kennegy, 1985: 09). 68 The Deeieion Chart The discussion of the various methodological concerns has disclosed that there are several prerequisites for the present analysis. In order to avoid methodological traps and since the analysis will be based on annual data, we need to determine if the data contain any annual trends. If so, the data must be adjusted by implementing the appropriate procedures, such as differencing. In addition, we need to determine whether or not the data contain autoregressive disturbances. If the diagnostic procedures of the Durbin-Hatson statistics and the autocorrelation function confirm the presence of autocorrelations, we would need to specify the correct autoregressive process (AR-1, AR-2), by examining the partial autocorrelation function. With the correct diagnosis of disturbances, we would then apply the appropriate estimation procedures, such as the Prais-Ninsten iterative procedures. The actual procedures of the present analysis can be summarized in the following decision tree chart: 69 Run OLS T Check Durbin-Natson Statistics and the autocorrelation functions I I l l Stationary Non- Stationary Non- with no AC stationary with AC stationary with no AC with AC STOP Data Data adjustment adjustment Run OLS Check Partial Autocorrelation functions STOP AR-l l l AR-k I Run EGLS Run EGLS for AR-I for AR-K STOP l I STOP Figure 3.1: The Decision Tree Chart for the Analysis 70 Figure 3.1 shows that there are four possible types of analyses based on the stationarity of series and on the presence/absence of autocorrelations: (1) stationary series with no autocorrelation, (2) non-stationary series with no autocorrelation, (3) stationary series with autocorrelation, and (4) non-stationary series with autocorrelation. For detection of non-stationary series and autoregressive disturbances, the correlograms of the autocorrelation functions and the Durbin-Natson statistics will be employed as the essential diagnostic tools. Likewise, the specification of the autoregressive process will be determined by the correlograms of the partial autocorrelation functions. A correct specification of the model in terms of trends and autocorrelations will be one of the primary concerns for the present analysis. And this will allow us to produce unbiased and efficient estimators of population parameters while providing valid hypothesis tests for significance and confidence intervals. CHAPTER IV OPERATIONALIZATION AND MODELS OF ANALYSIS Introduction Chapter 11 presented the causes of and hypotheses about import penetration suggested by three competing theoretical perspectives. To test these causal factors, this chapter will provide the operationalizations of the variables and the functional forms of the models under investigation. The variables to be defined and to be studied in the functional forms include: (1) import penetration, (2) comparative advantage, (3) macroeconomic condition, (4) United States’ import policy, (5) NICs’ export policy, and (6) United States’ national security interest. After defining these variables, I will present the functional forms of the individual and integrative models of import penetration, that is, interdependence, dependence, mercantilist, and composite models. Subsequently, I will discuss the expected results of these models in accordance which the suggestions of the respective theoretical perspectives. 71 72 Operationalization of the Variables Impprt Penetretion Import penetration denotes the amount of imports from country A to country B. Generally, it means the actual value of commodities transferred across national boundaries. The amount or the value of imports from the NICs into the U.S. can be measured as a national total for a NIC and also as a total for each of six levels of processing, classified in terms of the amount of labor skills and capital necessary to produce the commodities. For the time period of the study, import penetration will be estimated as follows: (a) the total amount of commodities imported from an individual NIC, measured as a percent share of imports, into the United States, and (b) the amount of imports from an individual NIC, measured as a percent share of imports, for a particular "level of processing" of commodities. Furthermore, the Pacific Basin regional totals will be computed as the aggregate of NICs’ national totals, and their totals for six commodity clusters. The official U.S. import statistics are compiled initially by the Bureau of the Census in terms of the commodity classifications in "Tariff Schedules of the United States Annotated" (TSUSA). These classifications are rearranged as the codes of Schedule A, which is based on the Standard International Trade Classification (SITC) of 73 commodities imported into the United States.1 Customarily, every item that is imported into the United States is assigned a 7-digit identification code. The 7-digit codes, which account for over 10,000 items, are part of a series of six progressively broader product categories up to the 1-digit level which separates U.S. imports into 10 general categories. For this study, I have decided to use a 2-digit SITC code2 which classifies imported commodities into 64 categories. This particular code allows sufficient variations of commodities to be meaningful while maintaining a data set of manageable size. Each of the 64 two-digit categories is assigned a particular processing code, following the Firebaugh and Bullock method of classifying commodities which distinguishes three levels of primary products and three levels of manufactured goods. For example, live animals were assigned to the first level of processing, organic chemicals to the fourth level, and office machines to the sixth level. This categorization presupposes that there is a qualitative difference between different types of imports. Since each commodity serves diversified consumers and since each faces different U.S. manufacturers, it seems reasonable to assume that each cluster of imported commodities from NICs will produce a distinct impact and penetration pattern. 1For a more detailed description of commodity classifications, see the "explanation of Statistics” in U.S. Generel Imports (FT 155) puinshed by the U.S. Department of Commerce, 1980, 1981. 2This seems to be a conventional procedure. For example, Leamer (1986) bases his analysis on the data set consisting of 10 aggregations from 61 two-digit SITC codes. 74 The import penetration variables are constructed as follows: NICPUS--NICs’ total imports as a percent of total U.S. imports. NICPUSLl-L6--NICs’ total imports as a percent of US total for each level of processing. SIGPUS--Singapore’s total imports as a percent of total U.S. imports. SIGPUSLl-L6--Singapore’s total for each level of processing. KORPUS--Korea’s total imports as a percent of total U.S. imports. KORPUSLl-L6--Korea’s total for each level of processing. HKPUS--Hong Kong’s total imports as a percent of total U.S. imports. HKPUSLl-L6--Hong Kong’s total for each level of processing. TAHPUS--Taiwan’s total imports as a percent of total U.S. imports. TAHPUSLl-L6--Taiwan’s total for the level of processing. Comperetive Advantage As a concept, comparative advantage offers a micro-economic rationale for import penetration. Essentially, it is based on a particular actor’s endowment with productive factors. Generally, comparative advantage is determined by the amounts of human and capital resources required to produce certain commodities. Previous analyses (Cline, 1984; Leamer, 1986) have shown that comparative advantage can be measured in a number of ways. For example, "value added per worker" as a proxy measure, indicates labor intensity of production. High values added indicate low labor intensity, and vice versa. The other measures include "physical capital per worker" which is the ratio of physical capital such as plant and equipment to employment in the industry, and "human capital per worker" which is the value of education and skills, or human capital, in the work force, measured by using the difference between the average wage and the wage of unskilled labor (Cline, 1984:46-47). 75 For the Pacific Basin NICs, the lower labor costs of production offer a commonly accepted rationale for their comparative advantages over the United States. If a NIC manufacturer can produce a particular commodity with cheaper labor costs than a U.S. counterpart, then that NIC manufacturer has a comparative advantage for that commodity. Labor costs are computed from the wage and output data available from national statistics and also from United Nations’ publications. The United Nations’ Industriel Stetistics Yearbooks from 1967 through 1985 provide a consistent basis for computing comparative advantage measures for Singapore, Korea and Hong Kong, while the measure for Taiwan is computed based on its national statistics. U.N. Industrial Statistics contains estimates of wages and salaries including all payments3 in cash or in kind made to employees and operatives during the reference year. From these data, comparative advantages are computed as the total of labor costs divided by the total value of industrial outputs for particular commodities. The following commodities represent the corresponding levels of processing: Level 1--Food, Food product, and Tobacco. Level 2--Textile and Leather. Level 3--Glass and Nonferrous metals Level 4--Hearing apparel, Footwear, and Steel. Level 5--Chemical, Plastic, and Metal products. Level 6--Machinery and Electrical Machinery. 3The payments include: (a) direct wages and salaries; (b) remuneration for time not worked; (c) bonuses and gratuities; (d) housing allowances paid directly by the employer; and (e) payments in kind. 76 The labor cost for each level is computed by taking the average of the selected commodities. This cost for each NIC is then divided by the cost in the United States for the same commodities. Computed as such, comparative advantage is a relative measure indicating the differences between the costs to produce those commodities for the NICs and the United States. A smaller value in comparative advantage represents an advantage for a NIC over the United States, while a larger value indicates a comparative disadvantage. The variables for comparative advantages are coded as follows: NICCA--N1Cs’ average comparative advantage. NICCALl-L6--NICs’ average comparative advantage for each level of processing. SIGCA--Singapore’s average comparative advantage. SIGCALl-L6--Singapore’s average comparative advantage for each level of processing. KORCA--Korea’s average comparative advantage. KORCALl-L6--Korea’s average comparative advantage for each level. HKCA--Hong Kong’s average comparative advantage HKCALl-L6--Hong Kong’s average comparative advantage for each level. TAHCA--Taiwan’s average comparative advantage. TAHCALl-L6--Taiwan’s average comparative advantage for each level. Maerpeconomic Conditions Macroeconomic conditions are relatively easy concepts to define. Major economic indicators, which illuminate international and domestic economic conditions are readily available and extensively quantified. However, the problem for the present study is deciding which macroeconomic indicators to adopt. Dornbusch and Frankel (1987) note that "the most important (potential macroeconomic determinant) is probably the exchange rate: an overvalued currency induces a rise in 77 import competition." In addition to the exchange rate, macroeconomic conditions are reflected by the real production (GNP) growth rate. Simply put, an improved core economy at the macro level will produce a higher consumption rate, holding the other factors constant. Another possible indicator of macroeconomic conditions of the core country is the rate of unemployment. Unemployment rates represent the domestic reactions to import penetrations. A high unemployment rate indicates that the domestic labor forces are being overwhelmed by the foreign competitions. A lower unemployment rate signifies that the core economy is maximizing its available labor forces for overall expansion. My analysis will use the dollar exchange rate, the United States’ GNP growth rate, and the rate of unemployment to evaluate the import penetrations from the NICs. The variables for macroeconomic conditions are coded as follows: NICER--NICs’ average exchange rates over the U.S. dollar. SIGER--Singapore’s avg. exchange rates over the U.S. dollar. 4/4 KORER--Korea’s avg. exchange rates over the U.S. dollar. HKER--Hong Kong’s avg. exchange rates over the U.S. dollar. _// TAHER--Taiwan’s avg. exchange rates over the U.S. dollar. ‘IUSGNP--U.S. Gross National Product growth rate. _./.USUEP--U.S. unemployment rate. U.S. Import Policy As an indicator of U.S. import policies, trade disputes have a certain impact on NICs’ import penetrations. If a NIC is charged with dumping a particular good, the penetration of the disputed good may be substantially curtailed. Conversely, the absence of trade disputes will be associated with unrestricted import penetrations. A 78 restrictive import policy, as indicated by an increase in trade disputes, is more likely to produce a decreased penetration, whereas a liberal import policy, as shown through a decrease in trade disputes, will lead to an increased import penetration. Data on trade disputes is available from the United States International Trade Commission (USITC) which publishes an annual report of petitions filed against foreign competitors. The petitions are filed and categorized in terms of (1) adjustment assistance/escape clause, (2) unfair import practice, (3) anti-dumping, and (4) countervailing duties. Domestic U.S. manufacturers may request import relief assistances under the adjustment assistance and escape clause petitions. The Trade Expansion Act of 1962 provides means whereby industries or firms that are seriously injured or threatened with serious injury because of increased imports may seek relief. According to the Annual Report (1973:5), following an investigation and affirmative finding by the Commission, the President may, ...under specified circumstances, increase rates of duty or impose other restrictions on imports which are causing or threatening to cause serious injury to a domestic industry, negotiate orderly marketing agreements with foreign countries, or certify adversely affected firms or groups of workers as eligible to apply for adjustment assistance. The adjustment assistance may take several forms. Actual adjustment assistance may be provided in terms of loans, technical assistance, and tax benefits to firms for unemployment compensation and retraining. Beginning in 1976, Section 201 of the Trade Act of 1974 provides the means whereby relief can be sought by domestic manufactures under 79 the escape clause. Relief may be sought by a trade association, a firm, a certified or recognized union, or a group of workers. Nhen petitioned, the Commission determines whether an article is being imported into the United States in such increased quantities as to be a substantial cause of serious injury or threat to the domestic industry. The Annual Report (1976:4) affirms that If the Commission determines in the affirmative, it must find the amount of the increase in, or imposition of, any duty or restriction on such article which is necessary to prevent or remedy such injury, or recommend the provision of adjustment assistance to firms, workers, or communities. The Commission conducts investigation under section 337 of the Tariff Act of 1930 to determine whether unfair methods of competition and unfair acts exist in the importation of articles into the United Sates. The Commission determines whether there is a tendency to substantially injure a domestic industry, efficiently and economically operated, in the United States. The Annual Report (1982:11) states that if the Commission finds a violation, it can issue orders excluding the violating goods from entry into the United States, unless it determines that such orders should not be issued in view of a public-interest consideration. Commission orders go into effect 60 days after issuance unless disapproved by the President. Section 201 (a) of the Antidumping Act, 1921, makes provision to deal with imported articles sold at less than their fair values. The Annual Report (1979:14) asserts that ...whenever the Secretary of Treasury advises the Commission that a class or kind of foreign merchandise is being, or is likely to be, sold in the United States or elsewhere at less than its fair value, the Commission shall determine within 3 months whether an industry in the United States is being or 80 is likely to be injured, or is prevented from being established, by reason of the importation of such merchandise. At the conclusion of its investigation, the Commission notifies the Secretary of the Treasury of its determination. If the Commission determines in the affirmative, the Secretary of Treasury issues a finding that dumping has occurred, and the described imports become subject to special dumping duties. In addition, the U.S. antidumping law, which is set forth in sections 731-740 of the Tariff Act of 1930, provides for the levying of special duties to offset sales at less than fair value (LTFV) by foreign producers. According to the Annual Report (198124), The Department of Commerce must determine whether an imported article is being sold at less than fair value, and the Commission must determine whether an industry in the United States is materially injured or threatened with material injury, or whether the establishment of an industry in the United States is materially retarded, by reason of alleged sales at less than fair value. In these investigations, the Commission is obliged to examine a domestic industry in order to assess the impact of imports on the industry’s economic health. Section 303 of the Tariff Act of 1930 provides the provision dealing with countervailing duties directed towards U.S. imports. The Annual Report (1980:20) states that ...whenever any country, dependency, colony, province, or other political subdivision of government, person, partnership, association, cartel, or corporation shall pay or bestow, directly or indirectly, any bounty or grant upon the manufacture or production or export of any article or merchandise manufactured or produced in such country, dependency, colony, province, or other political subdivision of government, then upon the importation of such article or merchandise into the United States, whether it is imported directly from the country of production or otherwise and whether it is imported in the same condition as when exported from the country of production or has been changed in condition by remanufacture or otherwise, there shall be 81 levied and paid, in all such cases, in addition to any duties otherwise imposed, a duty equal to the net amount of such bounty or grant, however the same be paid or bestowed. Effective January 1, 1980, the Trade Agreements Act of 1979 became the law governing responses to countervailing duties. The Annual Report (1980:3) states that the new statute added a requirement. The Commission must determine that an injury to a domestic industry has occurred before imposition of countervailing duties. In addition, the benefit of a Commission’s injury determination is extended to subsidy cases involving merchandise subject to duty provided that the merchandise originates in a signatory to the Subsidies Code of the General Agreement of Tariffs and Trade (GATT). The Annual Report (1983:3) tells us that under title VII of the Tariff Act of 1930, as added by the Trade Agreements Act of 1979, the Commission ...conducts preliminary and final investigations to determine whether there is a reasonable indication that (in preliminary investigations) or whether (in final investigations) a U.S. industry is materially injured or threatened with material injury, or the establishment of such an industry is materially retarded, by reason of imports of merchandise which is being sold at less than fair value or is benefiting from foreign subsidies. The Department of Commerce determines whether dumping or subsidies exist and, if so, the margin of dumping or amount of the subsidy. The Commission finding concerns only whether there is injury by reason of that dumping or subsidy. If the Commission finds that such dumping or subsidies are causing damage to a U.S. industry, the Commission recommends an amount of dumping or countervailing duties equal to the 82 amount of the foreign subsidy. The above categories of petitions, compiled by the Commission, are aggregated annually for the overall total and the totals for six processing clusters of commodities. Although a petition indicates a domestic demand for protectionism and an alleged or an actual unfairness from the foreign manufactures, petitions have varying degrees of success and importance. As such, each petition is coded with respect to its date of filing, date of settlement, and final outcome (an affirmative or negative finding). Trade disputes, as an approximation of the U.S. import policy, are coded as follows: TDlTOT—-The total number of petitions coded by the date filed. TDlLl-L6--The total number for each level. TDlYTOT--The total number of petitions filed coded by the date filed and with affirmative findings. TDlYLl-L6--The total for each level. TDZTOT & TDZLl—L6--The total number of petitions coded by the date decided a the total for each level. TDZYTOT & TDZYLl-L6--The total number of petitions coded by the date decided with affirmative findings 8 the total for each level. NI§§’ Expert Policy Export policies are those policies actively promoted by the NICs’ governments in pursuit of market access in the United States. Theoretically, since trade relations are bilateral, an inclusion of NIC export policies is essential in providing a complete explanation of U.S./NICs trade relations. Okita (1986) maintains that "effective government policies in NICs buttressed by close cooperative efforts by government, industry, and academia are the most important reason for the recent NICs’ development." The World Bank Study (1988:29) affirms 83 that Korean industrial policy is notable for the prominent role of government in the economy, the boldness of policy changes, and extraordinary results. The study adds that Korea’s export takeoff would not have been possible without decisive and innovative government policies.4 It appears that effective government policies are consolidated around financial issues. Generally, an effective export policy denotes strong governmental assistance in exporting production. Governmental assistance may be in the form of direct government subsidies, low interest rate loans to export manufactures, and lower taxes for the manufacturers. Effective government policies can be viewed as a function of the resources and intentions of government to finance, and thus to facilitate, exports and export industries. Governmental assistance data are usually confidential and apparently very difficult to obtain. Due to the unavailability of this data, export policies of NICs will be approximated from the rates of domestic capital formation and domestic governmental expenditure. Okita (1986:24) claims that the NICs recent development is accountable to "...a high rate of investment, specifically investment backed by dramatically higher domestic saving rates (or domestic capital formation).' Given this, rates of capital formation are adopted as the proxy measure of export policies, since they are relatively reasonable indicator of the amount of investment capital available to the NICs (and thus the NICs’ governments) for investment or export purposes. 4See, Kprea; Managing the Industrial Transition published by the World Bank, 1988. 84 The rates of government expenditure are also used to measure export policies of the NICs. These rates of expenditure reflect the financial resources a NIC government’s disposal. Before a NIC’s government can assist export productions, it must have financial resources to support its intentions. A higher rate of government expenditure implies a higher financial capacity to assist, while a lower rate denotes a decreased capacity to promote export production. The data for each NIC’s domestic capital formation and government expenditure are adopted from the United Nations’ National Aecpgpts Statisties for the time period of the study. The NICs’ export policies are coded as follows: NICCF--NICs’ average domestic capital formation as the percent of gross domestic products. SIGCF--Singapore’s capital formation. KORCF--Korea’s capital formation. HKCF--Hong Kong’s capital formation. TANCF--Taiwan’s capital formation. NICGE--NICs’ average government expenditure as the percent of gross domestic products. SIGGE--Singapore’s government expenditure. KORGE--Korea’s government expenditure. HKGE--Hong Kong’s government expenditure. TANGE--Taiwan’ government expenditure. Natipnal Securjtv Interest According to the realist vision, national security interest is the most important political determinant of economic relations. A high politics of national security is the primary source of concern about NICs’ import penetration. Conventional realism stipulates that a state will manipulate its economic arrangements to promote its security interests. From this posture, Sneider (1986:77) asserts that 85 ”American security interests are [as] deeply engaged in Asia as they are in Europe...American interests remain constant--to prevent hostile powers from expanding their control over the region and to maintain sufficient power to thwart any direct or indirect threat to the United States." National security interest is a complex phenomenon and it has acquired many meanings over the years. Keohane and Nye (1989:35) affirm that "national interests will be defined differently on different issues, at different times, and by different governmental units." For the present study, the U.S. security interest in a particular NIC will be approximated by measuring the amount of U.S. aid given to that NIC. In other words, we can define the American security interest as the amount of U.S. foreign aid given to a NIC on an annual basis. The amount of U.S. economic and military aid is thus a direct indication of the degree and magnitude of American security interests bestowed on a particular NIC5. In this way, the American security interest is coded as follows: NICUSTA--The average of the total U.S. aid given to the NICs. SIGUSTA-—Singapore’s total aid from the United States KORUSTA--Korea’s total aid from the United States HKUSTA--Hong Kong’s total aid from the United States TANUSTA--Taiwan’s total aid from the United States 5Although this indicator does not capture the U.S. economic interests in those NICs, more reliable data was not possible to reconstruct. I’ve tried but failed to come up with the data which summarizes economic and financial interests of the U.S. (at least during the time period that I am investigating). Hong Kong reports the percentage of the U.S. direct investment but other NICs, especially Korea and Taiwan, do not. 86 Models of Import Penetration Models of import penetration present functional forms for analysis based on the causal factors suggested by the dominant theoretical perspectives. The interdependence model emphasizes a microeconomic cause of import penetration. Following the classical theory of international trade, the functional form of the interdependence model consists of the "labor" comparative advantage of NICs. Comparative advantage is measured as an absolute difference in labor costs between a NIC and the United States in producing a particular commodity. The interdependence model will be tested on three levels of analysis: (1) a regional level, an aggregate sum for the all NICs, (2) a national level, a national total for each NIC, and (3) a commodity level, a total for each of six clusters of commodities. The dependency model focuses on the macroeconomic determinants of import penetration. The model will be comprised of the exchange rate, GNP growth rate and unemployment rate of the United States. Exchange rates between the NIC’s currency and the U.S. dollar reflect the overall financial condition of the international capitalist system. The United States’ GNP growth and unemployment rates indicate a change in the core nation’s overall economic conditions and the rate of expansion to which the peripheries (the NICs economies) are subjected. The mercantilist model is concerned with the self-centered policies of trading nations. The model stems from the notion that a nation-state seeks to manipulate trade relations to maximize its own 87 interests. Each nation-state pursues a distinct set of objectives under the rubric of national interest. The functional form of the mercantilist model will simultaneously consider the national interests of the United States and those of the NICs. It will include, (1) trade disputes which approximate U.S. import policy toward the NIC, (2) NIC’s domestic capital formation and government expenditure to indicate NICs’ export policy toward the United States, and (3) U.S. total aid to the NIC as an indicator of U.S. security interest in that NIC. Contrary to the interdependence hypothesis, the dependency and mercantilist hypotheses are not explicitly concerned with the micro or commodity level of analysis. Their emphasis is on the aggregate or macro level of analysis. These hypotheses do not preclude, however, the possibility that might exist at the commodity level of analysis. Whether this focus on macro relationship is caused by the nature of their theoretical frameworks or simply by a lack of attention to the micro level is unclear. Testing the dependency and mercantilist hypotheses at the commodity level of analysis may provide some information regarding their lack of precision. The results of the tests may demonstrate the inherent limitations of dependency and mercantilist hypotheses in dealing with the micro level of analysis. On the other hand, the tests may lead to some tangible suggestions as to how to change or modify their assumptions and arguments to perhaps better address the present circumstances. In sum, these tests may lead to a synthesis of theories or suggestion for a new theory of international political economy. 88 In addition to the individual models of import penetration, the composite model will be presented and tested. The determinants of all three individual models will be pooled together and this will be evaluated from the regional, national and commodity level of analysis. The purpose of analyzing the composite model has two main purposes: (1) to determine whether the individual models (independently analyzed) are theoretically sound in explaining the NICs’ import penetration into the United States, and (2) to examine the possibility of developing a new theoretical framework, to improve our understanding of international trade relations. The composite model analysis will provide a certain insight as to whether the interdependence, dependency, and mercantilist models are adequate explanations of the NICs’ import penetration. For instance, if the interdependence model is a theoretically sound explanation of the NICs’ import penetration, we would expect to find consistent results in both the individual and composite model analyses.6 If inconsistent, we would need to re-think about its theoretical soundness and question its adequacy as a theory of international trade relations. In this way, the composite model analysis is an additional test that critically evaluates the individual models and their determinants of the NICs’ import penetration. Moreover, based on the findings of the composite model analysis, we may provide some suggestions for a new theoretical framework of international trade relations. The findings may yield some new ideas for theory construction and theoretical synthesis. At any event, the composite 6This assumes that the independent variables are uncorrelated from one another. 89 model will be an important part of the present study, providing a new means for analyzing the individual models and their determinants, and yielding some concrete suggestions for theory construction and theoretical synthesis of international trade relations. Interdependence Model The interdependence perspective is based on the fact that different commodities require different factors for production, and different countries have different quantities of factors of production. A country will possess a comparative advantage in good X if the country is relatively well endowed with factors that are used intensively in the production of X. A country will export the commodity that uses intensively its relatively abundant resources. If a NIC manufacturer can produce a particular commodity at a lower cost, thus possessing a comparative advantage over a United States counterpart, we would expect the importation of that particular commodity into the U.S. market. A change in comparative advantage will be associated with a change in the level of import penetration. Particularly, an increase in a NIC’s overall comparative advantage will be associated with an increase in its overall import penetration, whereas a decrease in such an advantage will be linked to a decrease in U.S. import penetration. Moreover, an increase in a NIC’s comparative advantage on a given level of processing will cause an increase in U.S. import penetration for that level of processing. The functional form of the interdependence model can be stated as 90 Y; 8 80 + 31x11 + U; (4.1) where Y; - Import Penetration X41 = "Labor" Comparative Advantage. Because "labor" comparative advantage is measured as an absolute difference between the labor costs in producing a commodity, we would expect a negative relationship between a NIC’s labor cost and its import penetration. He would anticipate that a lower labor cost will covary with a higher propensity to achieve import penetration of that commodity. Conversely, a higher labor cost for a particular commodity will be associated with less likelihood of that commodity being imported into the U.S. market. 0 n enc Mod l The dependency perspective is based on the idea that trade relations are conditioned by the developed economies in particular, and by the international system in general. For a NIC, the types and amounts of import penetration will be conditioned and defined by United States economic situations. If the U.S. economy experiences an expansion in financial and industrial output, a NIC would also experience an expansion in its exports to the United States. In addition, since the dependent relationships between the NICs and the United States exist within the broader international capitalist system, international economic indicators such as the dollar exchange rates will have an undeniable impact on the NIC’s imports into the U.S. market. The importations of commodities from the NICs into the United States will be determined by the U.S. 91 macroeconomic conditions and by circumstances surrounding the international capitalist system. Thus, this perspective maintains that Y; = 80 + 82x12 + B3Xi3 + B4Xi4 + U; (4.2) where Y; - Import Penetration, X12 = Dollar Exchange Rate, X13 = U.S. Gross National Product Growth Rate, and X14 = U.S. Unemployment Rate. Accordingly, we would expect a positive relationship between the NICs’ exports to the United States and their dollar exchange rate. A higher dollar exchange rate means that a NIC’s own currency (i.e., Korea’s won, Taiwan’s dollar, etc.) is under-valued relative to the U.S. dollar, thus it is cheaper to import commodities into the U.S. market. A lower exchange rate denotes that a NIC’s currency is over- valued relative to the U.S. dollar, making its commodities more expensive in the U.S. market. He would also expect a positive relationship for the GNP growth rate. A positive growth rate, an expansion in the United States economy, will be associated with increased import penetration, while a negative growth rate will coincide with a lower, more sluggish rate of penetration. The U.S. unemployment rate would be negatively related to the NIC’s import penetration. If the United States domestic labor market is damaged by the NIC’s imports, we would expect a reduced demand for those imports. And if the United States experiences full employment, we would expect an increased demand for commodities from the NICS. 92 Mereantilist Model The mercantilist perspective is based on the idea that domestic political interests under the rubric of national interests will shape the trade relations between nations. The trade policies and political concerns of participating states will determine the exchanges of commodities across national boundaries. In particular, self- interested United States import and NIC export policies, coupled with overriding security interests, will overshadow the NICs’ import penetrations. Driven by its quest for wealth and power, the United States will manipulate trade relations with the NICs to maximize its own economic and political interests. First, the United States’ mercantilistic economic interests, as indicated by its import policies, will have a significant impact on the NIC’s import penetration. A restrictive U.S. import policy, defined as an increase in trade disputes between the United States and a NIC, would curtail the level of import penetration, while a liberal policy, demonstrated as a decrease in trade disputes, would assist in promoting a higher rate of NIC’s import penetrations. Second, the United State’s security interest in a particular NIC would determine that NIC’s propensity to penetrate the U.S. market. The amount of economic and military assistance denotes the U.S. security interest and commitment to that NIC which, in turn, allows a higher rate of import penetration. A lower or decreasing amount of United States assistance to a NIC signifies that the U.S. security 93 interests and commitments are dissipating with respect to that particular NIC, curtailing its rate of penetration. Third, the mercantilist economic interests of the NICs, as shown by their export policies, will be causally related to their propensities toward import penetrations in the U.S. market. A NIC’s export policy, defined as domestic capital formation and government expenditure will be positively related to import penetration. An increase in a NIC’s domestic capital formation and government expenditure will be a necessary condition for an increase in import penetration. A decreased capital formation and expenditure will cause a reduction in the NIC’s import penetration. The functional form of the mercantilist model can be written as follows: Y1 = Bo + 35X15 + 35Xi5 + 37x17 + ngig + U; (4.3) where Y; = Import Penetration, X15 a Trade Dispute, Domestic Capital Formation, Government Expenditure, and United States Total Aid. X a. 05 I X17 X18 From this model, we expect a negative relationship between NIC’s import penetration and trade dispute. Simply put, more trade dispute (i.e., friction) mean fewer import penetration. From increased domestic capital formation, government expenditure, and total U.S. aid, we would expect a positive relationship with a NIC’s import penetration. In other words, increased domestic capital, government expenditure and U.S. assistance will have a positive impact on a NIC’s propensity to successfully export to the United States. Holding other factors constant, export stimulating policies by a NIC, coupled with a 94 convergence of security interests, will allow increased commodity penetration into the U.S. market. Composite Model The composite model of import penetration pools the determinants of the three individual models of international trade into a single equation. Alike the individual models, the composite model will be analyzed from the regional, national and commodity level of analysis, and it can be written as follows: 1 = 31X1§ + 82X1§ + B3X13 + B4X14 + 85X15 + B5X15 (4.4) 87 i7 + 88 i8 + i where Y1 = Import Penetration X11 - "Labor" Comparative Advantage. X12 = Dollar Exchange Rate, X13 = U.S. Gross National Product Growth Rate, and X14 = U.S. Unemployment Rate. X15 = Trade Dispute, X15 = Domestic Capital Formation, X17 = Government Expenditure, and X13 2 United States Total Aid. As mentioned earlier, the composite model analysis will provide an additional, critical test of the individual models and their determinants of the NICs’ import penetration. This analysis will focus on the consistency of the interdependence, dependency, and mercantilist models as an adequate explanation of the NICs’ import penetration into the United States. If an individual model is a consistent explanation, we would expect to find a similar result between the individual and the composite analysis. If the results are 95 inconsistent, we would seriously question an individual model’s theoretical adequacy. Additionally, the composite model analysis will provide some concrete suggestions as to "how to" conceptualize international trade relations, probing the possibility of a new theory and/or a theoretical synthesis of international trade. For the composite model, we expect the same directions of the relationships between the determinants of the individual models and the NICs’ import penetration into the United States. In short, the comparative analysis of the individual and the composite models has an important implication for the evaluation of the causes of the NIC’s import penetration into the United States, and more importantly, for the study of international trade relations. CHAPTER V RESULTS OF THE ANALYSES Introduction Based on the functional forms described in Chapter IV and paying special attention to the methodological concerns discussed in Chapter 111, this chapter will present and interpret the results of the analyses. First, I will discuss the results of the descriptive statistics, emphasizing the essential variables’ central tendencies and dispersions. Second, I will report the results of the unadjusted OLS estimation, appraising the predictive powers of the individual models. Third, I will present the results of the visual diagnostics and identify the correct functional forms for the analysis. Fourth, the results of the estimations on the adjusted data will be presented and elaborated to evaluate the explanatory and predictive powers of the individual models for the regional, national and commodity levels of analysis. Finally, I will provide a comparative analysis of the individual and the composite models in order to determine the individual models’ overall performances and their "goodness of fit" in explaining the NICs’ import penetration into the United States. The presentation and discussion of the results will focus on the explanatory powers of the individual models and their determinants at the various stages of estimations--yielding some new insights into how 96 97 to think about international trade relations. Presentation and Interpretation of Results Descriptive Analysis Result The present study tests the models of import penetration based on the data from the Pacific Basin NICs for the period between 1967 to 1985. The data set consists of 99 variables, including: 35 import penetration variables, 35 comparative advantage variables, 5 exchange rate variables, 2 variables for United States GNP growth and unemployment rates, 7 trade disputes variables, and 5 each for government expenditure, capital formation, and U.S. total aid. Table 5.1 presents the descriptive statistics (e.g., mean, standard deviation, minimum value, and maximum value) for the select variables. On the average, the Pacific Basin NICs account for about 7 percent of the total U.S. imports. Taiwan has the highest share with 2.56 percent, followed by Hong Kong (2.09 percent), Korea (1.69 percent), and Singapore (.61 percent). For comparative advantage, the NICs’ average labor cost is about 61 percent of the United States’ labor cost. Korea has the highest advantage over the United States with 53.33 percent, Taiwan is the second with 59.75 percent, followed by Singapore (64.78 percent) and Hong Kong (65.90 percent). The dollar exchanges rates show that Hong Kong has the weakest currency relative to the U.S. dollar with the average index of 121.6, Korea is the second with 105.1, Singapore is the third with 104.7, and Taiwan has the strongest dollar exchange rate with the index value of 98 102.6. For the period between 1967 and 1985, the U.S. gross national products have grown at 2.76 percent annually and unemployment rates have averaged to 6.4 percent. Table 5.1: Descriptive Analysis Results for Select Variables Variable Mean Std. Dev. Minimum Maximum N NICPUS 6.95 2.42 2.793 11.31 19 SIGPUS .61 .36 .052 1.233 19 KORPUS 1.69 .74 .396 2.900 19 HKPUS 2.09 .27 1.621 2.538 19 TANPUS 2.56 1 19 .569 4.748 19 NICCA 60.53 7 13 50.8 72.8 18 SIGCA 64.78 9.15 49.4 84.2 19 KORCA 53.55 4.58 46.8 63.8 19 HKCA 65.90 4 34 58.5 73.7 18 TAHCA 59.75 14 61 41.1 78.2 19 NICER 108.51 11.24 98.0 135.4 19 SIGER 104.67 16.12 88.2 128.7 19 KORER 105.12 38.69 56.0 179.8 19 HKER 121.62 20.47 98.1 164.6 19 TANER 102.61 3.80 95 5 106.3 19 USGNP 2.76 2.41 -2 5 6.4 19 USUEP 6.41 1.84 3 5 9.7 19 TDlTOT 81.63 63.61 8 263 19 NICGE 54.79 1.57 51.9 57.2 18 SIGGE 58.08 4.27 52.2 67.4 18 KORGE 36.98 2 81 32.1 42.0 18 HKGE 36.98 2.81 32.1 42.0 18 TANGE 87.08 5.55 75.3 97.9 18 NICCF 158.28 24.68 112.1 197.3 18 SIGCF 208.96 47.90 111.3 282.4 18 KORCF 149.12 17.41 109.2 180.3 18 HKCF 132.48 30.98 82.8 188.2 18 TAHCF 142.51 20.30 116.9 192.0 18 NICUSTA 69.83 60.49 -66 197 19 SIGUSTA 5.79 31.57 ~53 93 19 KORUSTA 248.05 144.27 58 698 19 HKUSTA 2.53 11 12 -14 32 19 TANUSTA 57.68 135 27 -255 392 19 99 American domestic industries, interest groups, and trade unions-- among others--have filed about 82 petitions per year to United States International Trade Commission. The actual numbers of petitions ranges from the minimum of 8 petitions and the maximum of 263 petitions. On the average, the NIC’s government expenditure is approximately 55 percent of that of the United States Taiwanese government has the highest spending tendency with 87.1 percent, surpassing Singapore (58.1 percent), Korean and Hong Kong (each with 37.0 percent). For domestic capital formations, the NICs, on the average, account for about 158 percent of the U.S. capital formations. Individually, Singapore has the highest capital formation rate with 209 percent, Korea is the second with 149 percent, Taiwan and Hong Kong averaged 143 percent and 132 percent, respectively. For U.S. military and economic assistance, Korea is the largest recipient with the annual average of $248 millions, Taiwan is the next largest with $58 millions, while Singapore and Hong Kong have received about $6 and $3 millions each. Unadjusted Ordinary Least Squares Estimationl For the individual models of import penetration, I ran ordinary least squares (OLS) to estimate their suggested effects. In this section, the results of unadjusted OLS estimations will be presented in terms of the percent of estimates with the correct signs to provide 1The result presented in this section is not the final result of the present analysis--it is merely a finding from the preliminary analysis. The final result of the analysis will be presented and elaborated after the data adjustment. 100 a general comparison of each model.2 The actual coefficient estimates--following the decision chart of the analysis--will be presented and discussed in details in the later sections after the visual diagnostics and subsequent data adjustment. Table 5.2 displays the expected results of the determinants of the individual models and the percentage of the unadjusted OLS estimates with the correct signs. Table 5.2: Expected Results and Percentage of Unadjusted OLS Estimates with Correct Signs for All Equations Modela Variable Expected Result OLS Result (%) INT CA 8 < 0 28.6 28.6 DEP ER 8 > 0 62.9 USGNP B > 0 45.7 USUEP B < 0 37.1 48.6 MER TDl B < 0 37.1 GE 8 > 0 60.0 CF 8 > 0 60.0 USTA B > 0 57.1 53.6 aNote that INT=Interdependence, DEP=Dependency, and MERsMercantilist Model. The results in Table 5.2 shows that comparative advantage, U.S. unemployment rate, and trade dispute are expected to have negative relationships with NICs’ import penetration, whereas dollar exchange rate, U.S. GNP growth rate, NIC’s government expenditure and capital 2The estimates with correct signs are those estimates that correctly predicted the expected directions (either positive or negative) of the causal relationships. The percentages of estimates with the correct signs provide a summary result of "fitness" between the theoretical models and the actual observations. 101 formation, and U.S. total assistance are expected to be positively related with the import penetration. The unadjusted OLS estimates for the individual models show that the mercantilist model has the highest overall prediction with 53.4 percent of estimates with the expected signs.3 The second highest is the dependency model with 48.6 percent and the last is the interdependence model with 28.6 percent of estimates predicted correctly. Among the individual determinants, the U.S. dollar exchange rate has the highest percent of estimates with the expected sign with 62.9 percent, followed by the NIC’s government expenditure and capital formation with 60.0 percent each. Tables 5.3 and 5.4 present the OLS estimates for the regional and the national level of analysis. Table 5.3: Percentage of Unadjusted OLS Estimates with Correct Signs at Regional and National Level of Analysis (%) Model Variable Regional Level National Level INT CA 0 0 0 0 DEP ER 100.0 50.0 USGNP 85.7 100.0 USUEP 0 61.9 0 50.0 MER TDI 57.1 0 GE 85.7 25 0 CF 85.7 50.0 USTA 57.1 71.4 50.0 31.3 3I have estimated 35 regression equations for each individual models for the total of 105 equations. The four NICs and a regional total--individually--have 7 equations: one for the aggregated sum for national and regional total plus six for the commodity clusters. 102 Table 5.4: Percentage of Unadjusted OLS Estimates with Correct Signs for Each NIC (%) Model SIG KOR HK TAN INT 0 0 0 0 DEP 33.3 66.6 66.6 33.3 MER 50.0 75.0 0 0 At the regional level, Table 5.3 indicates that comparative advantage is a poor explanatory variable of the NICs import penetration. None of the OLS estimates has the expected sign. This signifies that, at least at the macro level, comparative advantage cannot predict the NICs’ import penetration. At the national level, comparative advantage--once again-—has no impact on the each NIC’s import penetration into the United States Table 5.4 show that none of the NICs’ comparative advantage has the correct sign. It appears that the interdependence model does not explain the variations of the NICs import penetration when aggregated into a summary figure. For the dependency model, the dollar exchange rate and the U.S. GNP growth rate are positively related with NICs’ import penetration as expected. In fact, 100 percent of exchange rates and 86 percent of U.S. growth rate have correct OLS estimates. At the national level, dollar exchange rate is predicted correctly for Korea and Hong Kong, and U.S. growth rate is correct for all NICs. In other words, at the national level, the dollar exchange rate explains Korea and Hong Kong situations better than Singapore and Taiwan, confirming the speculation that Korea and Hong Kong have pursued the exchange rate 103 policies to keep the values of their currencies down (i.e., devalue) relative to the U.S. dollar to maintain their products’ competitiveness in the U.S. market, while Taiwan and Singapore did not. For the U.S. GNP growth rate, it seems that all the NICs are affected by the changes in the U.S. economy conditions. An improving U.S. economy, as indicated by a higher GNP growth rate, has a positive impact on NICs’ import penetration, whereas, a worsening U.S. economy curtails NICs’ import penetration. This is especially true for Taiwan, Singapore, and Korea, but not for Hong Kong. At the regional level, the mercantilist model predicted 71 percent of all coefficient estimates. Specifically, the NICs’ government expenditure and capital formation display 86 percent of estimates predicted correctly. This supports the stipulation that much of NICs’ import penetrations is internally generated by the NICs themselves. As expected, a higher government expenditure and capital formation is directly related to an increase in the NICs export into the United States At the national level, the model shows 31 percent of correct coefficient estimates. The model has the highest prediction rate for Korea (75 percent), followed by Singapore (50 percent), while generating the worst fits for Hong Kong and Taiwan (0 percent). Table 5.5 presents the percentage of the unadjusted OLS estimates with the correct sign for the commodity level of analysis. For the commodity level of analysis, the comparative advantage shows some predictability with 33.3 percent of estimates showing the expected signs. The dependency model, as a whole, has correctly predicted correctly 47.8 percent of the unadjusted OLS coefficient estimates. 104 The mercantilist model has the highest overall prediction with 55.8 percent of estimates predicted correctly. For Levels 1, 2 and 5 commodities, the mercantilist model has the highest predictive ratio among the individual models. For Levels 3 and 6 commodities, the dependency model registered the highest percentages with 60 percent and 53 percent of the expected signs. The interdependence and mercantilist model have the highest percent for Level 4 commodities with 60 percent of estimates with the correct signs. Table 5.5: Unadjusted OLS Estimates with Correct Signs for Commodity Level of Analysis Model Levels of Processing (%) Overall (7.) L1 L2 L3 L4 L5 L6 INT 20.0 40.0 40.0 40.0 60.0 40.0 33.3 DEP 33.3 40.0 60.0 53.3 40.0 46.7 47.8 MER 60.0 55.0 55.0 60.0 60.0 45.0 55.8 The OLS estimation on the unadjusted data has shown that for the regional and the commodity level of analysis the mercantilist model has the best predictive power among the individual models. Concurrently, the dependency model has the superior predictive power at the national level of analysis. Somewhat surprisingly, the interdependence model is the worst individual model, predicting none of estimates correctly at the regional and the national level, while explaining only a small variations (i.e., 33.3 percent) at the commodity level of analysis. According to the unadjusted OLS 105 estimations, import penetrations from the NICs can be attributed to the determinants of the dollar exchange rate, the U.S. GNP growth rate, the trade disputes, the NIC’s government expenditure and capital formation, and the United States total assistance. a i nost'c Result Upon examining the Durbin-Watson statistics and the scatter plots of residuals from the unadjusted OLS estimations, there appears to be some annual trends and/or autocorrelations. As discussed in Chapter III, to determine whether the data contain any non-stationary series and/or autoregressive disturbances, I have precedes to examine the correlograms of the autocorrelation functions of all 99 variables. Table 5.6 summarizes the result of the visual diagnostic based on the correlograms of autocorrelation and partial autocorrelation functions. The correlograms of the autocorrelation functions have confirmed the suspicion from the Durbin Natson statistics and the residual plots that the data contains certain annual trends and autoregressive disturbances. To determine the annual trends, the correlograms were checked to see if the values of autocorrelation coefficients move toward zero at the small or at the large value of the lag. The series which move toward zero at relatively large lag (e.g., non-stationary series) have been adjusted by taking the first differences. After taking the first differences, the correlograms are plotted on the adjusted series to see if these adjustments have successfully removed the annual trends. 106 The visual diagnostic of the correlograms and the adjustment procedures disclosed that 22.2 percent of the data had stationary series with no autocorrelation, while 72.7 percent had non-stationary series caused by the annual trends. None of the variables had just autocorrelation, but 5 variables4 have both non-stationary series and autocorrelations. Table 5.6: Diagnostic Results from Autocorrelation and Partial Autocorrelation Functions Model Variable 1a 2 3 4 Total PUS 7 26 0 2 35 INT CA 6 28 O l 35 DEP ER 0 4 O 1 5 USGNP l 0 0 0 1 USUEP O l O O l MER TDl 5 2 O O 7 GE 0 5 0 0 5 CF 0 5 O 0 5 USTA 3 2 0 0 5 Total 73 0 4 22 99 aNote that lastationary series with no autocorrelation, 2=non~stationary series with no autocorrelation, 3=stationary series with autocorrelation, and 4=non-stationary series with autocorrelation. In addition, to determine the correct autoregressive process for those five variables with autocorrelations, the correlograms of the partial autocorrelation functions are constructed and interpreted. 4The variables with both the non-stationary series and autocorrelations are KORPUSL5, TAHPUSL3, KORCAL3, NICER, and HKER. 107 These correlograms have disclosed that the partial autocorrelation coefficients are "cut off" at the first lag, telling us that correct error specification is the first order auto-regressive process. Based on the visual diagnostics, the appropriate functional forms (equations) for the present analyses are summarized in Table 5.7. Table 5.7: Correct Equations for the Analysis Based on the Visual Diagnostics Model 0LS(D=0) OLS(D=1) EGLSa INT NIC 0 7 0 SIG 1 6 0 KOR 0 5 2 HK 0 7 0 TAN 0 7 0 DEP NIC 0 0 7 SIG 0 7 0 KOR 0 6 1 HK 0 0 7 TAN 0 6 l MER NIC 0 7 0 SIG 0 7 0 KOR 0 6 1 HK 0 7 0 TAN 0 7 0 TOTAL 1 85 19 aNote that EGLS is Estimated Generalized Least Squares. Table 5.7 shows that only 1 equation out of 105 total equations belongs to unadjusted OLS estimation. 85 equations (81 percent of all equations) require adjusted OLS estimation and 19 equations (18 108 percent) necessitate EGLS estimation. The individual models have varying numbers of the appropriate functional forms. The interdependence model requires 32 adjusted OLS, 2 EGLS and 1 unadjusted OLS estimation. The dependency model necessitates 19 adjusted OLS, 16 EGLS, and zero unadjusted OLS estimation. The mercantilist model requires 34 adjusted OLS and l EGLS estimation. Estimation on the Adjusted Data After the data adjustment based on the diagnostic tests, I ran the second OLS estimation and also ran the EGLS estimation. The results from these estimations are quite different from the previous, unadjusted OLS estimation. Table 5.8 displays the results of the adjusted estimations in terms of the percentages of estimates with the expected signs for the overall, the aggregated regional, and the national levels of analysis. Table 5.8: Estimates from the Adjusted Data with Correct Signs for All Equations, Regional and National Level (%) Mod Var Overall Result Regional Level National Level INT CA 62.9 62.9 85.7 85.7 50.0 50.0 DEP ER 45.7 71.4 50.0 USGNP 65.7 85.7 100.0 USUEP 31.4 47.6 14.3 57.1 0 50.0 MER TDl 31.4 14.3 25.0 GE 54.3 57.1 50.0 CF 37.1 28.6 0 USTA 51.4 43.6 71.4 34.2 50.0 31.3 109 According to Table 5.8, the interdependence model has the highest overall percentage (62.9 percent) of estimates with the expected signs for the adjusted estimations. This is a direct opposite of the unadjusted OLS estimation, where the interdependence model had the lowest predictive rate with 28.6 percent of estimates with the correct signs. The second highest is the dependency model with 47.6 percent and the lowest overall is the mercantilist model with 43.6 percent of estimates with the expected signs. At the regional level, the interdependence model has the highest percentage of estimates with the correct signs with 85.7 percent. Once again, this is a drastic change since none of the unadjusted OLS estimates had the correct sign. The dependency model has the second highest prediction rate with 57.1 percent of estimates predicted correctly. However, without including the U.S. unemployment rate, the dependency model would have tied with the interdependence model for the highest predictive percentage. The mercantilist model has the lowest percentage of estimates with the correct sign with 34.2 percent, dropping from the 71.4 percent of the unadjusted OLS estimation. At the national level, the interdependence and dependency model correctly estimated 50.0 percent of their estimates, while the mercantilist model only predicted 21.3 percent correctly. Among the individual variables, U.S. GNP growth rate registered the highest rate with 100.0 percent of estimates with the expected signs. Table 5.9 summarizes the percentages of the adjusted estimates with the expected signs for the commodity level of analysis. For the 110 commodity level of analysis, once again, the interdependence model has the highest predictive rate with 63.3 percent of estimates with the expected signs. The dependency model has the second highest rate with 42.2 percent, and the mercantilist model has the lowest predictive rate with 45.0 percent of estimates with expected signs. For the individual levels of commodities, the interdependence model has the highest percentage of estimates with the correct sign for Levels 2, 4, 5 and 6 commodities.5 The mercantilist model has the highest predictive ratio for Level 1 commodities. The dependency model has the highest percentage for Level 3 commodities. Throughout the six clusters of commodities, the interdependence model has yielded the highest predictive rate followed by the dependency model and then by the mercantilist model. Table 5.9: Estimates of the Adjusted Data with Correct Signs for Six Levels of Commodities (%) Model Levels of Processing (%) Overall (7.) L1 L2 L3 L4 L5 L6 INT 0.0 100.0 40.0 80.0 80.0 80.0 63.3 DEP 33.3 40.0 53.3 40.0 46.7 66.7 46.7 MER 50.0 50.0 50.0 40.0 40.0 40.0 45.0 5To recapitulate the six levels of processing: Level 1 consists of primary commodities (e.g., agricultural product); Level 2 includes semi-processed primary commodities, such as processed food product; Level 3 consists of beverage, textile yarn, and nonmetallic mineral; Level 4 has medium processing commodities, including wearing apparel, footwear, chemical product, and steel; Level 5 contains medical product, plastic material, and metal product; and Level 6 includes electrical, office, and industrial machines. 111 The results of the estimations on the adjusted data are radically different from those of the unadjusted OLS estimation. Although the results of the dependency and mercantilist model are not drastically different, the results of the interdependence model are significantly different from the unadjusted results. In fact, the interdependence model has consistently transcribed the highest percentages of estimates with the correct sign. For the overall, regional, national and commodity level of analysis, the interdependence model has the highest predictive rates. In light of the results from the unadjusted OLS estimation, this is indeed an alteration of the results (in terms of the coefficient estimates with the expected signs) that requires a closer inspection. Aside from the percentages of estimates with the expected signs that I have discussed thus far, the succeeding set of tables presents the actual coefficient estimates in order to delineate the exact nature and strength of relationships between the NICs’ import penetration and the theoretical determinants of the individual models. The coefficient estimates for the interdependence model are presented in Tables 5.10 and 5.11. Table 5.10 shows that--at the aggregated regional and national level-~NICs, Hong Kong and Taiwan display negative relationships between their comparative advantages and import penetrations, as expected, while the results for Singapore and Korea are inconsistent indicating the positive (rather than negative) coefficient estimates. In addition, only one estimate (TANCA) is significantly different from zero at .05 significance level. 112 Table 5.10: Coefficient Estimates for the Interdependence Equations at Regional and National Level b°1 Variable (Stand3rd Error) (t-ratio) NICCA - 110 (.097) (-1.143) SIGCA .0015 (.004) (.328) KORCA .002 (.016) (.134) HKCA -.013 (.021) (-.595) TANCA _ -.049** (.021) (-2.327) **Significantly different from 0.00 at .05 level. Table 5.11 shows that with the exception of Level 1 commodities, all of the coefficients have the expected signs. The results confirm the interdependence hypothesis that comparative advantage is negatively related to the commodity penetrations from the NIC’s.6 However, this finding is met conclusive. None of the coefficients 6The coefficient estimates of the commodity clusters for each NIC (Singapore, Korea, Hong Kong and Taiwan) are reported in Appendix 0. They are excluded from the present chapter because of their vastness. 113 estimates are statistically significant. The estimated coefficients are not significantly different from zero at .05 significant level. Table 5.11: Coefficient Estimates for the Interdependence Equations at the Commodity Level of Analysisa bil Level (Standard Error) (t-ratio) NICCALl .0001 (.004) (.031) NICCAL2 -.001 (.011) (-.089) NICCAL3 -.023 (.027) (-.861) NICCAL4 -.025 (.087) (-.285) NICCALS -.035 (.063) (-.558) NICCAL6 -.039 (.080) (-.492) aAll coefficients not significantly different from 0.00 at .05 level. Tables 5.12 and 5.13 present the estimated coefficients of the dependency determinants of import penetration. The coefficients are displayed for the aggregated regional and national, and the commodity clusters. 114 Table 5.12: Coefficient Estimates for the Dependency Equations at Regional and National Level ER GNP UEP NIC b- b~3 b-4 (Stand.1Error) (Stand? Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) NICs .086** .334** .447 (.040) (.135) (.259) (2.133) (2.471) (1.730) SIG -.006 .019 .022 (.004) (.015) (.030) (-1.554) (1.239) (.732) KOR -.004 .024 .027 (.008) (.047) (.090) (-.461) (.501) (.300) HK .009 .052 .079 (.006) (.041) (.084) (1.473) (1.267) (.937) TAN .100*** .152** .144 (.024) (.040) (.079) (4.126) (3.788) (1.820) **Significantly different from 0.00 at .05 level. ***Significantly different from 0.00 at .01 level. Table 5.12 shows that at the aggregated regional and national level, the U.S. dollar exchange rates are positively related to import penetration for the NICs, Hong Kong, and Taiwan. More precisely, overvaluation of the U.S. dollar has a significant impact on import penetrations for the NICs, Hong Kong, and Taiwan, while Singapore and Korea show inconsistent results, suggesting that their import penetrations continue to increase in spite of the dollar devaluations. As expected, the U.S. GNP growth rate is positively related to import 115 penetration for the NICs and all four individual NICs. Once again, the coefficients are significant only for the NICs and Taiwan. Additionally, for all four NICs, the U.S. unemployment rate yields the inconsistent result, showing that the rate is positively related to the NICs’ import penetration into the United States. Table 5.13: Coefficient Estimates of the Dependency Equations for Regional Aggregates at Commodity Level of Analysis ER GNP UEP MIC 11' b'3 b'4 (Stand.1Error) (Stand? Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) L1 0.020 0.064 0.068 (0.012) (0.042) (0.081) (1.648) (1.519) (0.842) L2 -0.018 -0.135 -0.080 (0.024) (0.086) (0.164) (-0.756) (-1.570) (0.338) L3 0.045* 0.044 0.058 (0.027) (0.089) (0.171) (1.684) (0.488) (0.338) L4 0.049 0.295 1.455 (0.111) (0.413) (0.795) (0.442) (0.715) (1.832) L5 -0.044 0.286 0.782 (0.073) (0.274) (0.527) (-0.595) (1.044) (1.483) L6 0.007 0.142 0.139 (0.075) (0.252) (0.483) (0.094) (0.565) (0.289) *Significantly different from 0.00 at .10 level. 116 For the commodity level of analysis, as shown in Table 5.13, the results are similar to those from the regional and national estimates. The dollar exchange rate has the expected direction of relationship for the Level 1, 3, 4, and 6 commodities. The GNP growth rate has the most consistent results for all levels of commodities except for the second level. The unemployment rate shows the consistent estimate only for the second level of commodities--the remaining levels have inconsistent results. These results, however, are not conclusive, since all coefficients are not significantly different from zero at .05 a level. Tables 5.14 and 5.15 present the estimated coefficients of the mercantilist determinants of import penetration, (i.e., trade dispute, NIC’s domestic capital formation, NIC’s government expenditure, and United States total assistance) for the regional, national and commodity clusters. Table 5.14 discloses that at the aggregated regional and national level, trade dispute is negatively related as expected only for Hong Kong. The other NICs have positive coefficient estimates telling us that a restrictive U.S. import policy, as indicated by an increase in trade disputes fails to curtail the import penetration. The NICs’ penetrations increased even during the times of growing trade disputes. The NIC’s domestic capital formations, contrary to the hypothesized direction, display a uniform negative relationship with their import penetrations. In other words, the pattern of the NIC’s import penetration has little to do with its domestic savings rate, at least for the aggregated national totals. 117 Table 5.14: Coefficient Estimates for the Mercantilist Equations at Regional and National Level TDl CF GE USTA NIC b'5 0'5 0'7 b'8 (Stand! Error) (Stand. Error) (Stahd. Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) (t-ratio) NICs .002 -.039 .085 .004 (.002) (.012) (.089) (.003) (1.092) (-2.860) (.964) 1 (1.569) SIG .001 -.001 .0008 -.00002 (.0003) (.001) (.0057) (.0007) (1.736) (-1.359) (.135) (-.023) KOR .0002 -.002 .021 .0007** (.0008) (.004) (.022) (.0003) (.224) (-.610) (.981) (2.076) HK —.0003 -.006 -.035 .006 (.0008) (.004) (.031) (.007) (-.426) (-l.562) (-1.134) (.933) TAN .001 -.009 -.005 -.0003 (.001) (.005) (.018) (.001) (1.077) (-1.783) (-.254) (-.313) **Significantly different from 0.00 at .05 level. Tables 5.14 also shows that the NIC’s government expenditure is positively related-~as expected--for the aggregated NICs, Singapore and Korea, but not so for Hong Kong and Taiwan. Increased imports into the United States from Singapore and Korea are at least partially determined by the expansions in their government expenditures, while for Hong Kong and Taiwan the government expenditures are irrelevant in explaining their import penetrations. The U.S. security interest as approximated through the U.S. total assistance has positive impacts 118 for the aggregated NICs, Korea and Hong Kong. However, the results in Table 5.14 indicate that the relationship is negative for Singapore and Taiwan, which means that their exports into the United States have not been affected by the U.S. total assistance. Table 5.15: Coefficient Estimates for the Mercantilist Equations for Regional Aggregates at Commodity Level of Analysis TDl CF GE USTA NIC b' b°5 b 7 b 3 (Stand.1Error) (Stand? Error) (Stand. Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) (t-ratio) L1 -0.000 -0.008 0.028 0.001 (0.004) (0.003) (0.024) (0.001) (—0.103) (-2.492) (1.167) (0.975) L2 0.068 0.017 0.049 0.002 (0.065) (0.012) (0.083) (0.003) (1.049) (1.505) (0.586) (0.636) L3 0.016 0.003 0.074 -0.003 (0.020) (0.011) (0.073) (0.002) (0.804) (0.324) (1.009) (—1.110) L4 0.016 -0.002 —0.181 0.040** (0.012) (0.056) (0.414) (0.013) (1.333) (-0.032) (-0.438) (3.048) L5 0.051 -0.028 -0.212 0.017 (0.067) (0.047) (0.280) (0.010) (0.749) (-0.598) (-0.759) (1.596) L6 0.034 -0.019 -0.025 -0.007 (0.025) (0.020) (0.144) (0.005) (1.377) (-0.952) (-0.175) (—1.445) **Significantly different from 0.00 at .05 level. As displayed in Table 5.15, for the commodity level of analysis, the trade dispute variable, at only one level of commodity (Level 1), 119 has the hypothesized direction of relationship and the remaining levels have inconsistent (positive) relationships. For NIC’s domestic capital formation, only two levels (Levels 2 and 3) show the expected direction of relationship. NIC’s government expenditure has a positive impact on the NIC’s import penetration for Levels 1, 2 and 3 commodities, while the U.S. total military and economic aid has the correct coefficients for Levels 1, 2, 4 and 5 commodities. Among those coefficients with the correct direction of relationship, only one coefficient (U.S. total aid for Level 4) is significantly different from the null hypothesis of B=0. Overall, the mercantilist determinants of import penetration have a higher predictive power for the lower level (Levels 1 to 3) commodities than the higher level (Levels 4 to Level 6) commodities. The mercantilist determinants have 66.7 percent of coefficients with the consistent results for the lower level commodities, while only 16.7 percent of coefficients are consistent for the higher level commodities. The results of the adjusted estimations suggest several interesting points concerning the individual models and their theoretical determinants of import penetration. Among the individual models, the interdependence model which isolates comparative advantage as the major cause of import penetration consistently predicted the highest percentages of estimates with the correct signs. But none of the estimates for comparative advantage is statically significant. For the individual determinants, the dollar exchange rate and the U.S. GNP growth rate have the highest predictive ratios, with a few statistically significant coefficient estimates. The mercantilist model and its determinants (especially trade dispute and the NICs’ 120 capital formation) have failed to explain the variations in the NICs’ import penetration into the United States Table 5.16 summarizes the results of the adjusted estimations in terms of rank-orderings of the individual models (based on their predictive performances) and lists the variables with the highest predictive ratios for the regional, national and commodity level of analysis. Table 5.16: A Comparative Evaluation of Individual Models for Their Predictive Performance at Various Types of Analysisa Type of Analysis Ranking of Model Key Variables Overall INT > DEP > MER GNP (62.9%) Regional INT > DEP > MER CA ER GNP GE TA (100%) Singapore DEP > MER > INT GNP GE (100%) Korea MER > DEP > INT GE TA GNP (100%) Hong Kong INT > DEP > MER CA ER GNP TA (100%) Taiwan INT > DEP > MER CA ER GNP (100%) Regional-L INT > MER > DEP CA GNP (83.3%) Singapore-L INT > MER > DEP CA GNP GE TA (50.0%) Korea-L INT > MER > DEP CA (83.3%) Hong Kong-L INT---DEP > MER UEP (83.3%) Taiwan-L INT > MER > DEP GNP (83.3%) aThe models (I=Interdependence, D=Dependency, and M-Mercantilist) are rank-ordered based on their percentages of estimated coefficients with correct signs; the listed variables are tho?e with the highest predictive ratios for the given type of ana ysis. 121 Based on the presentation and interpretation of the results from the previous tables and from the summary results displayed in Table 5.16, we can conclude the following points concerning the Pacific Basin NICs’ import penetrations into the United States: (1) Overall, the interdependence model has the highest explanatory power regarding the NICs’ import penetration. The interdependence model which had the worst predictive ratio (28.6 percent of estimates predicted correctly) for the unadjusted OLS estimation has the highest predictive ratio (62.9 percent) for the adjusted estimation. By the same token, the mercantilist model regressed from the best (53.6 percent) to the worst (43.6 percent) after the data adjustment. The dependency model consistently scored the second highest rate for both estimations. This means that, on the whole, the interdependence model explains the most variations concerning the NICs’ import penetration. For the individual determinants, however, the U.S. GNP growth rate has the highest predictive ratio (65.7 percent), even out- performing comparative advantage (62.9 percent). Generally speaking, the U.S. macroeconomic condition determines the changes in the Pacific Basin NICs’ export performances. (2) For the aggregated regional level of analysis, on the whole, the interdependence model has the highest explanatory power. Nonetheless, a number of the determinants (including, comparative advantage, dollar exchange rate, GNP growth rate, NICs’ government expenditure, and U.S. total assistance) has certain impacts on the NICs’ import penetration. 122 Furthermore, the dependency determinants of the dollar exchange rate and GNP growth rate are statistically significant. Minimally, it appears that the NICs’ import penetration is a multifarious phenomenon, caused by many--not just one or two--determinants. (3) At the national level of analysis, although similar, it is clear that the Pacific Basin NICs’ are very much different from one another. The Gang of Four share a common sensitivity to U.S. macroeconomic conditions (e.g., the U.S. GNP growth rate), but this is where the similarity ends. Singapore and Korea’s import penetrations into the United States are conditioned by active governmental interventions (e.g., government expenditure), while Hong Kong and Taiwan are affected by their competitive labor costs and dollar exchange rates. These dissimilarities suggest that Singapore and Korea are in the "dependent-mercantilist" relationship, whereas Hong Kong and Taiwan are in the "Interdependent-dependent" trade relationship with the United States. (4) For the commodity clusters at the regional level of analysis, once again the interdependence model has the highest explanatory power, followed by the dependency model and the mercantilist model. Among the individual determinants, the U.S. GNP growth rate and the comparative advantage are tied for the highest predictive rate with 83.3 percent of coefficient estimates predicted correctly. These results suggest that the clusters of commodity from the Pacific Basin NICs into the United States are attributed to their competitive labor costs and improving U.S. macroeconomic conditions. 123 (5) For the import penetration of the commodity clusters at the national level of analysis, the interdependence model (for all four NICs) has the highest explanatory power. However, each NIC’s commodity penetration is caused by a number of the different theoretical determinants. Singapore’s commodity clusters are affected by its comparative advantage, the U.S. GNP growth rate, the government expenditure, and the U.S. total assistance. For Korea the single most important determinant the lower labor cost; for Hong Kong, the major determinant is the U.S. unemployment rate; and for Taiwan, it is the U.S. GNP growth rate. 0n the average, the Pacific Basin NICs import penetration at the commodity clusters are determined by the labor costs, the U.S. GNP growth rate and the NIC’s governmental expenditure. In sum, the results appear to confirm the notion that the Pacific Basin NICs’ commodity penetration is indeed a multifarious event. A Comparative Analysis of Individual and Composite Model The purpose of estimating the composite model, as mentioned in Chapter IV, is to lay the groundwork for a new theory of trade relations, especially between the United States and the Pacific Basin NICs. To inquire into the possibility of a new theoretical framework for the NICs’ import penetration, I ran the composite model equations following the results of the visual diagnostics.7 Table 5.17 7The visual diagnostics yield the following equations (0 unadjusted OLS, 19 adjusted OLS, and 16 EGLS equations) for the composite model analysis: 124 summarizes the findings of the composite model analysis.8 Table 5.17: A Comparative Evaluation of Individual Models from the Composite Analyses for Their Predictive Performance at Various Types of Analysisa Type of Analysis Ranking of Model Highest Variables Overall INT > DEP > MER CA (62.9%) Regional INT > DEP > MER CA ER UEP GE TA (100%) Singapore INT > DEP > MER CA GNP CF (100%) Korea INT > DEP > MER CA UEP TA (100%) Hong Kong INT > DEP > MER CA ER GNP GE TA (100%) Taiwan INT > DEP > MER CA ER GNP GE TA (100%) Regional-L INT---DEP > MER CA ER GNP UEP GE TA (50.0%) Singapore-L MER > DEP > INT ER TD GE TA (66.7%) Korea-L INT > DEP > MER CA UEP CF (66.7%) Hong Kong-L INT > MER > DEP CA UEP TA (66.7%) Taiwan-L INT > DEP > MER CA GNP UEP (66.7%) aThe models (I-Interdependence, D=Dependency, and M=Mercantilist) are rank-ordered based on their percentages of estimated coefficients with correct signs; the listed variables are tho?e with the highest predictive ratios for the given type of ana ysis. OLS(D=0) OLS(D=1) EGLS NIC 0 0 7 SIG 0 7 0 KOR 0 6 1 HK 0 0 7 TAN 0 6 1 8See Appendix E for the coefficient estimates for the composite model analysis. 125 Table 5.17 suggests that, with the exception of the commodity clusters for Singapore, the interdependence model consistently has the highest explanatory power. For the individual determinants of import penetration, comparative advantage displays a convincing predictive rate throughout the analyses. Clearly, the NICs’ lower labor costs have certain undeniable impacts on their import penetrations into the U.S. markets. At the aggregated regional and national level of analysis, the results suggest that the NICs’ import penetration is a complex event caused by a number of factors beside the lower labor costs. In particular, the NICs’ import penetrations are affected by the macroeconomic factors of the dollar exchange rate, the U.S. GNP growth rate, the U.S. unemployment rate, the NIC’s governmental expenditure, and the U.S. total assistance. Consequently, the results imply that the aggregated sums of the NIC’s exports into the U.S. markets are caused by a combination of lower labor costs, the U.S. macroeconomic conditions, and the NIC’s export policies. In addition, for the commodity clusters at the regional level of analysis, the interdependence and dependency model are tied with the highest explanatory power. Nhen the models are pooled together, the individual determinants of comparative advantage, the dollar exchange rate, the U.S. GNP growth and unemployment rate, the NICs’ government expenditure, and the U.S. total assistance have certain causal impacts on the NICs’ import penetrations into the United States Nhen the models are pooled for each individual NIC, the results are more diverse, confirming the results from the individual model 126 analyses that each NIC’s commodity penetration has a distinct rationale. For Singapore, the mercantilist model has the highest predictive ratio and its commodity penetration is affected by the dollar exchange rate, the trade dispute, the government expenditure, and the U.S. assistance to Singapore. Korea’s commodity penetration is conditioned by its lower labor cost and domestic capital formation, and the U.S. unemployment rate, which fits the "interdependent but still dependent" argument of trade relations. Hong Kong’s commodity penetration is stimulated by its lower labor cost, the U.S. unemployment rate and the U.S. total assistance. It seems that Hong Kong’s penetration is less dependent on the United States than originally assumed under the individual model analysis. For Taiwan’s commodity penetration, comparative advantage, U.S. GNP growth and unemployment emerge as the dominant causes, sharing many common points with Korea’s commodity penetration. Tables 5.18 and 5.19 present the changes in the signs of the coefficient estimates between the individual and composite analyses. Positive numbers denote improvements in predicting the expected relationships and negative numbers signify worsening results of the analyses. According to Tables 5.18 and 5.19, for the overall analyses, the interdependence model shows no improvement over the individual analysis, while the dependency model discloses the most improvement and the mercantilist model shows a slightly worsened result. Among the individual determinants, the U.S. unemployment rate has the highest improvement, while the U.S. GNP growth rate reveals the worst alteration with six additional miss-predictions. For the aggregated 127 regional and national level of analyses, the results of the composite analyses are not significantly different from those of the individual analyses. There are very little or no changes between the individual and composite analyses. Table 5.18: Changes in the Signs of Coefficient Estimates between Individual and Composite Analysis Type of Analysis INT DEP MER Overall 0 +4 -1 Regional 0 -1 0 Singapore +1 0 -1 Korea +1 0 0 Hong Kong 0 0 0 Taiwan 0 0 0 Regional-L -2 0 +3 Singapore-L -1 +3 +3 Korea-L -1 0 -2 Hong Kong-L +1 -1 -1 Taiwan-L +1 -1 -3 Table 5.19: Changes in the Signs of Coefficient Estimates between Individual and Composite Analysis for Each Variable Type of Variable Analysis CA ER GNP UEP TD CF GE TA Overall 0 +2 -6 +8 +2 -2 -1 0 Regional 0 0 —1 0 0 0 0 0 Singapore +1 0 0 0 0 -1 0 0 Korea +1 0 -1 +1 0 0 0 0 Hong Kong 0 0 0 0 0 0 0 0 Taiwan 0 0 0 0 0 0 0 0 Regional—L -2 0 -2 +2 +2 +1 0 0 Singapore-L -1 +2 0 +1 +2 -1 +1 +1 Korea-L -1 -1 -1 +2 -1 0 -2 +1 Hong Kong-L +1 0 0 -1 0 0 -1 0 Taiwan-L +1 +1 -1 +3 -1 -1 +1 -2 128 The changes are most obvious for the commodity cluster analyses. At the aggregated regional level, the mercantilist model shows the highest improvement while the interdependence model records the worst results. The U.S. unemployment rate and the U.S. trade dispute illuminate the highest improvement over the individual analyses, and at the same time, the U.S GNP growth rate connotes the worsening expected result. For each NIC, the changes are very subtle and distinct. For Singapore, the dependency and mercantilist models are improved over the individual analyses. In fact, the U.S. exchange rate and the U.S. trade dispute display the highest improvement, whereas the labor cost and the domestic capital formation have worsened predictive powers. For Korea, both the interdependence and mercantilist models exhibit worsened estimation results, while the dependency model remains constant. The unemployment rate shows the highest improvement, whereas the NIC’s governmental expenditure demonstrates much worsened results. The composite model analysis for Hong Kong shows indicates that the interdependence model has improved slightly and the dependency and mercantilist model have worsened over the individual model analyses. In addition, the comparative advantage and government expenditure recorded the best and worst changes, respectively. For Taiwan, the dependency model exhibits the highest improvement, whereas the mercantilist model registers the regressive alteration. In particular, the U.S. unemployment rate denotes the highest improvement and the U.S. total assistance has gotten significantly worse. 129 To further the comparative evaluation of the individual and composite model analysis, Tables 5.20 and 5.21 present the percent changes in the coefficient estimates between the analyses. The differences in coefficient estimates between the individual and composite estimations are reported as the absolute differences.9 These differences would indicate the degree to which the individual models and their determinants are sufficient explanations in dealing with the NIC’s import penetration.lo Table 5.20: Absolute Differences of Coefficient Estimates between the Individual and Composite Model (%) Type of Analysis INT DEP MER Overall 290 905 207 Regional 89 115 23 Singapore 500 27 50 Korea 800 348 96 Hong Kong 46 53 34 Taiwan 18 49 61 Overall, the dependency model shows the highest average difference (905 percent) between the coefficient estimates of the individual and composite analyses. The interdependence model 9The differences are reported as the percent differences in absolute values. They are computed as the absolute differences between the individual and composite coefficient estimate over the individual coefficient estimate: Difference - |(B1 - B-)/B1|, where 81 . individual estimate and 81 - composite estim te. 10The main concern is to determine whether (1) the individual model analyses are theoretically sound in explaining the NIC’s trade relations or (2) we need to develop a new theoretical framework to improve our understanding of the phenomenon. 130 discloses the second highest difference (290 percent) and the mercantilist model has the lowest difference (207 percent). For the aggregated regional and national level of analysis, the results in Table 5.20 reveal that the differences at the regional level are similar to the overall results, the individual NICs, however, display unique differences.11 The differences are somewhat alike for Singapore and Korea on the one hand, and for Hong Kong and Taiwan on the other hand. Singapore and Korea have the highest deviations from the interdependence model with 500 percent and 800 percent, respectively. The second highest differences are: for Singapore, the mercantilist model with 50 percent, and for Korea, the dependency model with 348 percent. Hong Kong has the highest difference for the dependency model with 53 percent, followed by the interdependence model (46 percent) and the mercantilist model (34 percent). Taiwan shows the highest differences for the mercantilist model with 61 percent, ensued by the dependency model (49 percent) and the interdependence model (18 percent). These differences indicate that at the aggregated national level of analysis: (1) Singapore and Korea present sizable differences, while Hong Kong and Taiwan show much smaller differences, and (2) the magnitude of differences across the models are much higher for Singapore and Korea than Hong Kong and Taiwan. For each individual determinant, Table 5.21 shows that for the overall and the aggregated regional analysis, the highest differences are reported by the dependency determinants--the U.S. GNP growth, the 11The differences for the commodity level of analysis are presented in Appendix F. 131 unemployment, and the dollar exchange rate. For the aggregated national analysis, several determinants, including comparative advantage, the dollar exchange rate, the U.S. unemployment rate, the NIC’s government expenditure, have shown to deviate from the individual estimates. For Singapore and Korea, the highest deviations are registered by the comparative advantage. For Hong Kong, the U.S. total assistance shows the highest deviation, and the government expenditure unfolds the highest deviation for Taiwan. All four NICs disclosed the least amount of deviation for the trade disputes, and the mercantilist determinants. And consistently, Singapore and Korea produced the more divergent deviations of the coefficient estimates than Hong Kong and Taiwan.12 Table 5.21: Absolute Differences of Coefficient Estimates between the Individual and Composite Model (%) Type of Variable Analysis CA ER GNP UEP TD CF GE TA Overall 290 624 1151 938 143 328 149 133 Regional 89 98 117 131 50 17 10 25 Singapore 500 17 37 27 0 100 100 0 Korea 800 300 292 452 0 250 33 100 Hong Kong 46 89 15 10 0 0 34 100 Taiwan 18 84 42 22 0 44 200 0 12In addition to the changes in the signs and the differences between the individual and composite model estimates, the results of the F-tests between the models are given in Appendix G. 132 On the whole, the composite model analysis confirms the individual model analyses that the Pacific Basin NICs’ import penetration into the United States are affected several determinants, including comparative advantage, the U.S. macroeconomic conditions, and the NIC’s export policies. Although similar, there are a few differences between the individual and composite model analyses which require further investigation. First, the composite model analysis shows that Singapore is much more mercantilistic than originally conceived. Clearly, Singapore practices an aggressive export-oriented industrial policy and this becomes more apparent only in the composite model analysis. Second, the U.S. unemployment rate, which was one of the lowest predictive variables for the individual model analysis, became one of the most consistent predictors for the composite model analysis. One possible explanation is that the U.S. unemployment rate is highly correlated with the GNP growth rate and it is only when these two determinants are analyzed along with the third determinant (or even the combinations of several determinants) that we can distinguish its actual impact. Third, the comparative evaluation of the coefficient estimates between the individual and composite models suggests that there are two sub-groups within the Pacific Basin NICs: (1) Singapore and Korea, with more mercantilistic tendencies; and (2) Hong Kong and Taiwan, with more dependency overtones. Fourth, throughout the composite model analyses, the most consistent determinant is the NIC’s comparative advantage, confirming the conventional wisdom of international trade. Each of the Pacific Basin NICs, however, seems to diverge from this common foundation and has a distinct set of determinants for their success in penetrating the U.S. 133 markets. The analyses demonstrate the fact that the NICs’ import penetration can be explained by a number of determinants, including a lower labor cost, an expanding U.S. economy, and a NICs’ vigorous export policy. Apparently, it is difficult to generalize the Pacific Basin NICs’ import penetrations into the United States, without first considering their individual differences. CHAPTER VI CONCLUDING REMARKS Introduction The present study began with the theoretical review and the development of subsequent hypotheses concerned with international trade relations, specifically between the United States and the Pacific Basin NICs. Nith the completion of the analyses, this chapter provides concluding remarks regarding the major findings and features of the study. I will first review the findings of the individual and composite model analysis of the interdependence, dependence, and mercantilist perspectives. Then, I will consider a somewhat broader set of questions concerning the implications of this study and its place within the context of international political economy. At the end, I will elaborate on possible directions of future research to advance our understanding regarding international trade relations. 134 135 Suggestions pf the Analyses In their recent book Beyond Globa11s_, Vernon and Spar (1989:1) boldly state that, If Americans are confused about where their international economic policies are headed, they have a right to their confusion. The assumptions of the past four decades on which the country’s policies once rested have been swept away. The United States is no longer the unchallenged economic leader of the noncommunist world, leading other nations toward an ultimate goal of open global markets. And if that is no longer the objective of the United States, where is the American economy headed? A part of the confusion is unmistakably connected to the NICs’ success in penetrating the U.S. market. An increase in NICs’ import penetration into the United States (2.8 percent of the U.S. general imports in 1967 to 11.3 percent in 1985) has coincided with the decline in the U.S. hegemony in the international system.1 Based on the findings of the individual and composite model analyses, the following conclusions about the NICs’ import penetration into the U.S. market are presented. Overall, the interdependence model with its emphasis on comparative advantage has the highest explanatory power regarding the NICs’ import penetration. The dependency model which is based on macroeconomic conditions, and the mercantilist model which is a policy model, could do no better than a distant second and third, lRecently, considerable attention has been focused on this increasingly relevant theme. See, for example, Keohane (1984), Bergsten (1988a and 1988b), Cuomo (1988) and Vernon and Spar (1989). 136 respectively. On the average, the key determinants of the NICs’ import penetration are comparative advantage and the U.S. GNP growth rate. That is, a lower labor cost coupled with an improving U.S. macroeconomic condition are the most important causes of import penetration. The study basically confirms the conventional wisdom of international trade that price determines the exchange of commodities. More importantly, the study also confirms, as pointed out by Harris (1986:68), that the Pacific Basin NICs are four very different politico-economic systems.2 Clearly, all four NICs are economically dynamic and are very active trading economies. Although, the NICs share a common sensitivity to external demands and conditions (i.e., the macroeconomic conditions which include the dollar exchange rates and the U.S. GNP growth and unemployment rates), each NIC has a distinct set of determinants, profoundly displaying the divergences among the "Gang of Four." Singapore is the most mercantilistic of the four NICs. Singapore’s import penetration into the United States is conditioned by active governmental interventions, confirming the observation made by Harris (1986:60-61) that "Singapore is...the predominant state capitalism of mixed economics, replete with consistent Keynesian policies....The state intervenes in almost everything--from the long- term and the strategic, the regulation of currency and the shaping of a future industrial structure." In addition to government interventions, Singapore’s success can be directly attributed to external demands, especially that of the U.S. GNP growth rate. The 2The policies pursued by the Pacific Basin NICs were frequently different, as were the attitudes, endowments, histories and sizes. 137 bulk of Singapore’s export to the United State is composed of Level 6 commodities (74 percent of its total export to the United States which consists of: electrical machine (26 percent), radio and television set (15 percent), and office machine (14 percent)), and these commodities are capital and technology intensive commodities. As such, Singapore is very much dependent on international financial and technology systems for its sustained import penetration. Three remaining NICs (Korea, Hong Kong, and Taiwan) are apparently affected by a similar set of determinants-~comparative advantage, the U.S. GNP growth and unemployment rate, and the dollar exchange rate. These three NICs’ import penetrations are determined by internally generated cost factors, while at the same time, they are conditioned by external demands and conditions. The import penetrations are affected by their labor costs, but each seems to be conditioned by different external, dependency determinants. Over the years, their import penetrations into the United States have focused on Levels 4 and 6 commodities, these two levels have composed 81 percent of Korea’s, 83 percent of Hong Kong’s, and 79 percent of Taiwan’s export to the United States. Although similar, each of these three NICs has a distinct commodity emphasis. Korea began to focus on high-tech and capital intensive commodities, such as computers, electronic components, and automobiles. Hong Kong has the least diversified export economy, largely relying on a few select commodities (e.g., wearing apparel accounts for 40 percent of its total export to the United States). Taiwan has been more successful in exporting the commodities which require medium processing skills. For instance, among the NICs, it has the highest share of Level 5 138 commodities, which include metal products (tools), travel goods, and plastic materials. Among this group, Korea seems the least dependent on external demands and conditions. Although less dependent, Korea’s import penetration into the United States is caused primarily by its lower wages for labor: "...if it takes five hours to assemble a computer, it costs just $8 in Korea. In the United States the cost would run between $100 and $200 (Glitman, 1986:181)." In this vein, Harris (1986:43) claims that ...[The conventional wisdom] explains Korean development as the product of low wages. Since low wages are general in developing countries, South Korea’s advantage is limited. Korean wages were not the lowest in the world at the beginning of fast growth,...However, productivity gains more than offset the increase in wages, so that the labor cost of output fell; explaining the productivity gains,...0ther factors...which for much of the time reduced the price of exports and made imports relatively expensive. As a result, Korea’s comparative advantage could be clearly expressed. Hong Kong and Taiwan are the most similar pair among the Four. Both share the common determinants of comparative advantage, dollar exchange rate, and U.S. GNP growth rate. At the same time, Taiwan is closer in terms of its economic tendencies to Korea, while Hong Kong appears to be more sensitive to external demands and conditions. Contextually, Hong Kong and Taiwan have an economic pattern which is dominated by light, labor intensive operations, primarily owned by small and medium companies. Their primary exports are the commodities with a medium level (Levels 4 and 5) of processing, including wearing apparel, footwear, and chemicals. Yet, in a development sense, Hong Kong lacks a state, an agency which endeavors to change the economy to 139 achieve a target pattern of future output. Taiwan is similar to Korea in that both countries face similar domestic and international political climates. Domestically, both states have dominated their economic activities through active state capitalism. Internationally, both Taiwan and Korea have been close allies of the United States and are impelled by a similar context of military insecurities (i.e., mainland China and North Korea). Figure 6.1 summarizes the findings of the analyses emphasizing the dominant tendencies of the Pacific Basin NICs’ import penetration into the United States. Figure 6.1: The NICs’ Tendencies Toward Import Penetration Into the United States Dependency High Medium Low Interdependence Hong Kong Taiwan Korea__ Mercantilist Singapore As Figure 6.1 illustrates, Hong Kong and Singapore are similar in that they are the most sensitive to the dependency determinants. Nonetheless, they differ because Singapore has been more tightly controlled by the state and is thus the most mercantilistic, while Hong Kong lacks state involvement in the conventional sense and is, 140 therefore, the most interdependent or non-political among the NICs.3 Hong Kong and Taiwan share the similar factors of import penetration, labor cost and external conditions. However, Hong Kong is more dependent on external demand and conditions than is Singapore. Because of its relatively smaller capacity, limited labor force, and fewer natural resources, Hong Kong’s share of import penetration among the NICs has steadily decreased over the years. Realizing its limitations, Hong Kong is turning toward more highly specialized "non- trade" service sectors, such as banking and insurance. Taiwan and Korea are similar in their historical and political backgrounds--that is, both are former Japanese colonies that are now dominated by authoritarian regimes. Consequently, Taiwan and Korea share similar causes of import penetration into the United States. However, the present study suggests that Taiwan is more sensitive to external demands and conditions than is Korea. Taiwan is dominated by a mass of small companies, while Korea is dominated by giant companies (Jae- Bol).4 It is plausible that smaller companies are more sensitive and vulnerable to the changes in the international trade system than are larger corporations.5 A giant company in Korea, such as Hyundai, is less sensitive and vulnerable to changing external demands and 3Harris (1986:59) notes that Hong Kong has no economic strategy nor long-term plan, no great state investment wielded as instruments of public ambitions. 4Glitman (1986:181) notes that "The PCs leave the $500 million plant on trucks made by Hyundai and are loaded onto container ships built and owned by Hyundai. Their destination: the United States." 5Keohane and Nye (1989:11-19) define sensitivity as a degree of responsiveness within a policy framework, and vulnerability as an actor’s liability to suffer costs imposed by external events even after policies have been altered. 141 conditions; it is more capable of promoting independently its products during changing international circumstances. In summary, there are certain similarities and dissimilarities between the Pacific Basin NICs. It is incorrect, even absurd, to generalize about the Pacific Basin NICs as a single group of developing economies. Each NIC is very different from one another. - Nevertheless, pairs of the NICs are somewhat similar: Singapore and Hong Kong are the most dependent economies; Hong Kong and Taiwan are somewhat more interdependent; Taiwan and Korea share similar historical and political backgrounds; and, Singapore and Korea exhibit more mercantilistic tendencies. Policy Implications pf the Study Fpr the United States The present study reveals a number of policy implications for both the United States and the Pacific Basin NICs. As noted by Bergsten (1988b:1), the United States has, in the recent years, experienced a rapidly changing international economic situation evidenced by rising budget and trade deficits:6 ...[Bush administration and Congress] will confront a current account deficit that remains well over $100 billion, and that will probably never fall much below $100 billion on the basis of present policies and exchange rates....a United 50.s. News & Norld Report (11/28/1988) notes that "The Congressional Budget Office says the fiscal 1990 deficit will be $137 billion. Some analysts say it will be $150 billion or more." 142 States that is for the first time in modern history the world’s largest debtor, with a net foreign debt that could rise to $1 trillion in the early 1990s in the absence of corrective action. They will face growing concern, both at home and abroad, that the United States can no longer compete effectively in the world economy and is losing its leadership role and ability. Concurrently, Cuomo (1988:l) proclaims that America has failed to adjust to the new reality of a global economy, and we are beginning to pay the price. The price of failure will not be just economic; the consequences in the next decade will be more than a slower rate of economic growth or a stagnation in living standards and opportunities. Just as any debtor is at the mercy of its creditors, if the U.S. continues to sink into debt, our foreign creditors will eventually have undue influence over our future and the policies of our elected government. As noted above, the problem facing the United States is obvious and is critical in many ways. The United States, in order to restore its stability, must plan for a structural reduction in her trade and budget deficits.7 But questions remain as to what the United States needs to do and how the United States will accomplish those needs. The present study implies a number of specific recommendations to reduce the U.S. trade deficit. One obvious way is to decrease the import penetration from the NICs. Based on the present analysis, the most consistent cause of the NICs’ import penetration into the United States is their relatively low labor costs. It confirms the conventional wisdom that the United States must restore the comparative advantages in its commodity production, either in terms of competitive production costs and/or an increased productivity. In 7These two deficits are very closely linked. At a minimum, they are highly correlated from one another, the correlation coefficient is 0.82 for the period between 1967 to 1987. See, Appendix H for the actual values and trends of these deficits. 143 this regard, Cuomo (1988:17) writes that "The trade deficit is perhaps the most outstanding manifestation of the American loss of competitiveness," and suggests that To increase our productivity and produce what we are truly capable of producing: that is the great profit of participation...the success of a collective enterprise--be it a company or a nation--depends on the contribution of the individuals that are part of it. Each of us must be given a chance to participate. Once given that chance, we must do our share (Cuomo, 1988:183). At the same time, the United States must persuade the NICs, especially Singapore and Korea, to reduce those governmental controls over the labor market which have kept labor costs below international market wages. The Korean government in particular has successfully suppressed any expression of discontent aimed at its strict control of labor. The United States would benefit by pressuring Korea and Singapore to allow unionization of the labor forces. Since their industries are dominated by giant companies, employing thousands of laborers, unions would undoubtedly raise wages. The United States would benefit by pressuring Hong Kong and Taiwan to adopt and enforce more rigorous employment regulations and building codes, including health and sanitation standards for their factories, which are mostly of small and medium size. This would raise costs of production in Hong Kong and Taiwan. Another possible adjustment to reduce the trade deficit is further manipulation of the dollar exchange rate. This strategy, however, is not as convincing as that of building comparative advantage. Cuomo (1988:21) hints at this point by noting that "no single macroeconomic policy, such as a lower value of the dollar, will 144 miraculously solve all the problems." He also argues that "trying to buy competitiveness with a cheap dollar will not work. Instead we have to earn it by producing goods and services that our own people and the rest of the world want [at the prices that they are willing to pay] (Cuomo, 1988:96-97)." The dollar exchange rate does have an impact in reducing the U.S. trade deficit under certain conditions. The United States must maintain a lower value of the dollar against the NICs’ currencies while coordinating this effort with other relevant policies. More importantly, the United States must constantly strive to keep the value of the dollar from being too low, because if the dollar value becomes too low, it will prompt various negative consequences: inflation, a decline in the U.S. living standards, a global recession, and (possibly) an international market crash. The above points and arguments imply that, in order to reduce its trade deficit, the U.S. government must be aggressively involved in the process of international trade. In particular, the U.S. government must be an active partner in facilitating a competitive edge in the international market through more energetic export financing, more incentive structures, and continued technological innovations.8 Minimally, the United States must face the inter- national reality that it is no longer the hegemonic actor: it is the strongest single nation (or the strongest among the equals), no longer 8This could be promoted by assisting the Export-Import Bank of the United States (Eximbank) in financing small and medium-size U.S. manufacturers, who cannot finance sales with their own limited resources, are trying to get into the export game, and need commercial lending to support their deals (See, Stokes (1988) for a detailed suggestion). 145 dominating the system, but still with enough veto power to block the initiatives of others. For the Pacific Basin NICs Aside from reducing the NICs’ import penetration, the United States seeks to reduce its trade deficits by increasing its export shares to the NICs. Recent evidence suggests that the U.S. government is pressuring the Pacific Basin NICs to open the various markets for U.S. goods and services. In addition to the argument of reciprocity, United States has insisted on the theme that A hard landing of the American economy would have enormous effects on other countries, most of which are far more dependent on international trade and international financial stability than is the United States. Their major market would shrink abruptly. Their own interest rates would be driven up by the rise in American interest rates. Renewed global inflation would threaten,...The bulk of the American adjustment must therefore be targeted on the small number of countries that are running surpluses: Japan, Germany and a few smaller European countries, and the Asian NICs (notably Taiwan and Korea) (Bergsten, 1988:13). Simply put, the Pacific Basin NICs will be forced to accept the solution proposed by the United States. The NICs will be forced to open their closed markets to U.S. commodities, and they must do this soon. Examples include the United States’ push for the opening of various service sectors (e.g., banking, insurance, and stocks) from the NICs and urging the NICs to open agricultural markets (e.g., cigarette, wine, and fruit). The NICs should and must open their markets to U.S. goods and services, when their domestic industrial producers are sufficiently competitive in the international market. 146 However, the United States must not be too hasty in its demands, or it may produce negative effects on both the United States and the NICs. As shown in the analyses, the NICs’ import penetration tends to increase when the United States experiences a higher rate of overall growth. Quite clearly, the NICs should not bite the hands that feed them. By the same token, the analyses imply that U.S. goods and services are more likely to penetrate the NICs’ market when they experience higher rates of growth. Specifically, the United States would benefit by avoiding any overt protectionist policies toward the Pacific Basin NICs. They would need to sustain a higher rate of overall growth to facilitate an increase in their consumption of U.S. goods and services. In addition, the solution proposed by the United States would compel the NICs to diversify their export markets. Up to now, the United States has been viewed as the prime export market; but, from now on, the N105 must actively look to non-American markets to maintain their export-led economies. Cuomo (1988:80) points out that Agreements are needed to shift imports away from the U.S. to those countries which maintain large trade surpluses. Today, the NICs and developing countries are where Japan was decades ago, and once again the U.S. must work to see that Europe-~and Japan--accept more exports from these dynamic young economies. Currently, many developing countries have their eyes fixed almost exclusively on the American markets, and work with the other advanced countries to ensure that they will accept more exports from the Third Norld. Such an attempt made by the United States to diffuse the NICs’ export markets could be viewed as a part of burden sharing. Japan, Europe, and the N105 must share relatively equal responsibilities in maintaining the international system by providing their fair share of 147 global military security, and by being more actively involved in providing economic assistance. Along the way, the Pacific Basin NICs may be forced to divert their surplus from U.S. trade to assist other Third Norld countries in the region (e.g., the Southeast Asian countries). Theoretical Implications of the Study The present study empirically tested the models of import penetration based on the causal factors suggested by the three major theoretical perspectives. The interdependence model, which emphasizes comparative advantage among producers, is a microeconomic explanation of international trade. The dependency model focuses on the macroeconomic determinants of international trade, including U.S. dollar exchange, U.S. GNP growth, and U.S. unemployment rates. The mercantilist model, as a policy explanation, is concerned with the self-centered policies of trading nations as represented by trade disputes, NICs’ domestic capital formation and government expenditures, and U.S. economic and military assistance to the NICs. These models were tested on three levels of analysis: (1) a regional level using an aggregate sum for all the NICs; (2) a national level using a national total for each NIC; and, (3) a commodity level using the total for each of six clusters of commodities. In terms of theoretical clarity, the interdependence model is the most explicit of the three models, suggesting that the hypothesized relationship exists at the commodity level of analysis. The dependency and mercantilist models, on the other hand, are not 148 explicitly concerned with the micro or commodity level of analysis. The dependency and mercantilist models place emphasis on the aggregate or macro level of analysis. These models do not, however, preclude the possibility that relationship might exist at the commodity level of analysis. The results of the analyses indicate that the dependency and mercantilist focus on macro relationships is due more to the lack of attention to the micro level of analysis than to the nature of their theoretical frameworks (i.e., this is not an essential neglect, these models are just more concerned with the holistic approach). Throughout the analyses, the dependency and mercantilist determinants registered about equal percentages of estimates with the correct signs for the macro and micro levels of analysis. These results suggest no inherent limitations in the dependency and mercantilist models in dealing with the micro level of analysis. In addition to the three theoretical models, the present study included a fourth analysis, by pooling the determinants of the individual models. The results of the composite model analysis suggests that comparative advantage had (once again) the highest predictive ratio among the individual determinants. Additionally, U.S. unemployment rate, especially at the commodity level of analysis, became an important determinant of the NICs’ import penetration. The composite model analysis tells us that the interdependence model is a theoretically consistent explanation of the NICs’ import penetration-- consistent result in both the individual and composite model analyses were found. It seems that each individual model, to somewhat varying extent, is a necessary, but not sufficient condition for import 149 penetration from the NICs into the United States. Overall, to explore the theoretical implications, the interdependence model consistently had the highest explanatory power regarding the NICs’ import penetration. The interdependence model had the highest predictive ratio (63 percent), followed by the dependency model (48 percent) and the mercantilist model (44 percent). Among the individual determinants, the U.S. GNP growth rate and comparative advantage had the highest predictive ratios with 66 and 63 percents of estimates with the correct signs. A similar result was found at the commodity level of analysis--the interdependence model had the highest explanatory power, followed by the dependency and mercantilist models. These results from both the overall and the commodity level of analysis confirm the argument that import penetrations from the Pacific Basin NICs into the United States can be attributed to their competitive labor costs and improving U.S. macroeconomic conditions. The study supports the viewpoint that trade relations should be viewed from labor/production costs within a broader, macroeconomic setting. Although comparative advantage had the highest explanatory power, at the aggregated regional and national levels of analysis, the other determinants such as dollar exchange rate, GNP growth rate, NICs’ government expenditure, and U.S. total assistance had certain impacts on the NICs’ import penetration. As mentioned earlier, it is clear that the Pacific Basin NICs are very much different from one another. The Gang of Four share a common sensitivity to U.S. macroeconomic conditions (e.g., the U.S. GNP growth rate), but this is where the similarity ends. Singapore and Korea’s import penetrations into the United States are conditioned by active governmental interventions as 150 indicated by a higher government expenditure, while Hong Kong and Taiwan are affected by their competitive labor costs and dollar exchange rates. Minimally, it appears that the NICs’ import penetration is a complex event, caused by a number of (not just one or two) determinants. Theoretically, the present study suggests that no single model has the explanatory power to account for all the variations of the Pacific Basin NICs’ import penetration into the United States. The composite model analysis indicated that the individual model approach to analysis yields a somewhat insufficient rationale in explaining the phenomena of import penetration. Based on the findings and suggestions of the various analyses, the following recommendations which may serve as a basis for development of a new theory and/or theoretical synthesis of international trade relations are offered: (1) International trade relations should be thought of as a complex event. More precisely, in addition to the single equation model, some alternative models of international trade need to be developed (e.g., the simultaneous model and the dynamic model). It may be that a single equation approach (i.e., individual model analysis) is an inadequate way of analyzing international trade relations. As eluded to earlier (see Appendix B), it is plausible to conceptualize trade relations as a simultaneous event caused by commodity penetration and protectionist trade policy. Although some theorists have argued that import penetration and protectionist trade policy are essentially the same event, these events may be individual and independent outcomes; if so, they may be considered to be simultaneous events. In other 151 words, there may exist a feedback effect between import penetration on the one side and protectionism on the other side--as import penetration increases, the demand for protectionism increases, at the same time. For example, a simultaneous model of import penetration can be formalized as follows: Y1 = r12Y2 + B11X1 + B12X2 + B13x3 + B15X5 + U1t (6 1) Y2 - erYl + 822X2 + 323X3 + 324X4 + 825X5 + UZt (6.2) where Y1 . import penetration Y2 = trade dispute X1 a comparative advantage X2 = exchange rate X3 . GNP growth rate X4 = U.S. foreign aid X5 = NIC’s export policy Uts = error term This simultaneous model may provide a new insight in dealing with international trade and advance our understanding regarding the NICs’ import penetration into the United States to a single, linear equation approach. (2) It is necessary to synthesize the individual explanations to possibly provide a more plausible account of current circumstances. The analyses have suggested that the interdependence model, with its emphasis on comparative advantage, has the highest explanatory power regarding the NICs’ import penetration. Although the dependency model which is based on macroeconomic conditions, and the mercantilist model, a policy model, could do no better than a distant second and third, it is possible to develop: (a) the interdependence-dependency 152 model, and (b) the interdependence-mercantilist model. First because the key determinants of the NICs’ import penetration are comparative advantage and the U.S. GNP growth rate, it may be possible to synthesize the microeconomic and macroeconomic determinants of international trade into a single model to account for the variations of the Pacific Basin NICs’ import penetration into the United States. This would be an integrative economic model which might consist of comparative advantage, dollar exchange rate, and U.S. GNP growth rate.9 Second, the interdependence and mercantilist determinants could be combined to address the relationship between the NICs’ import penetration and comparative advantages, while controlling for the NICs’ mercantilistic policies. The present study has indicated that a lower labor cost is the most important cause of import penetration. The Heckscher-Ohlin comparative advantage argument is based on naturally endowed production factors, and assumes that flows of commodities are carried out in purely economic, market conditions-- completely divorced from political regulations or interventions. This is a questionable assumption and it is credible to ask whether NICs’ lower labor costs are naturally endowed factors or result from governmental controls. As such, this second synthesized model would consist of comparative advantage and the NICs’ domestic policy toward labor forces. If the NICs’ lower labor cost is a naturally endowed 9Although this may appear to be an overtly inductive approach, theoretically speaking, the dependency model of trade relations shares much common ground with the interdependence model. These models are more complimentary than competitive when comparing how each explains import penetrations. Both hold that trade relations are motivated by economic factors and are fueled by the notions of profitability. 153 factor, then we would expect to find a strong relationship between comparative advantage and import penetration while holding the NICs’ labor policy constant. And if not, we would expect to find a weak or significantly reduced relationship when the NICs’ labor policy is introduced into the analysis. In this way, this synthesis model may confirm or disconfirm the conventional wisdom of international trade that commodity penetration is simply determined by naturally endowed production factors, as defined by a lower labor cost. (3) The following factors and conditions need to be considered in developing a new theory of international trade relations: (a) There is a need to specify and enunciate the historical conditions and specific rules, domestic and inter-state, which govern the international trade system. It is necessary to identify the cultural differences and ethical factors of international trade; (b) More attention should be paid to the political side of the international trade system--that is, economic relations should be viewed as political activities; and, (c) More attention should be paid to the role played by the government and other domestic interests in contributing to the internal dynamics of the international system. Caveats and Prospects It is clear from the analysis and from the discussion of the study that the Pacific Basin NICs are important participants in international trade. In a final note, I would like to point out the plausibility of the present study’s external validity, discuss some of 154 its shortcomings in terms of construct validity, and explore possible directions of future research. The external validity of the present study concerns replicability over other places and times. Is it plausible that the current study can be replicated over other sets of countries for some other time periods? It is very likely that this question can be answered in the affirmative. For instance, the study could be duplicated over the Southeast Asian NICs (ASEAN, Association of South-Eastern Asian Nations) and the Latin American NICs, such as Brazil, Argentina, and Mexico. In this regard, Cuomo (1988:58) affirmed that "The ‘Four Tigers’ of East Asia--Korea, Taiwan, Singapore, and Hong Kong--got most of the publicity, but industrialization was becoming a brand phenomenon, and countries like Thailand, Malaysia, Bangladesh, and Pakistan were beginning to figure in U.S. trade statistics." Furthermore, a similar study could be undertaken for the Central/Latin American NICs to see if a similar set of determinants can be attributed as causes of their import penetration. In addition, for replicability over time, a similar analysis can be undertaken for longer or shorter periods of time for any given set of the NICs, or for any group of countries. Aside from the present study’s replicability, there are several methodological and theoretical issues which should be explored. There are certain questions concerning construct validities of the variables employed in the present study. That is, there are certain questions concerning the fit between the theoretical constructs and the precise indicators. The problems associated with construct validities may be most severe for the mercantilist determinants of import penetrations. 155 The present study utilizes the macro, broad measures of the NICs’ export policies to represent their mercantilist interests in trade relations. For instance, the present study employs U.S. total economic and military assistance to capture American security interests bestowed on a particular NIC. Although this indicator does not capture the U.S. economic interests in those NICs, it was not possible to reconstruct more reliable data. The data which summarizes U.S. economic and financial interest in the NICs were impossible to obtain on an annual basis--Hong Kong reports the percentage of the U.S. direct investment, but other NICs, especially Korea and Taiwan, do not. Furthermore, for the NICs’ mercantilistic policy, the present study employed the NICs’ domestic capital formation and government expenditure rather than other, more representative, indicators. Governmental assistance data for export productions are usually confidential and are very difficult to obtain. Because of the unavailability of this data, export policies of NICs were approximated from the rates of domestic capital formation and domestic governmental expenditure. In addition, there may be a similar problem of construct validities for the dependency determinants--whether these variables actually represent the theoretical constructs of the dependency perspective, especially for dollar exchange rates and U.S. GNP growth rate. Nith respect to the merits of the present study, one of the major contributions is its presentation of several, new variables of international trade. In particular, the present study introduces a new approach in conceptualizing trade disputes and comparative 156 advantage. Import penetration is categorized into six commodity clusters to allow a micro level of analysis, which is rarely done in cross-national analysis. Moreover, the data set is inspected and adjusted for autocorrelation and trend prior to estimation procedures. By introducing and undertaking these issues, the present study has explored some new possibilities in analyzing international trade. It is obvious that greater attention should be directed toward the formulation of new concepts and methodologies in the study of international trade. Attention needs to be focused on developing new variables of international trade. Examples are: a variable that describes governmental control over export production and labor wage, a variable that characterizes a country’s dependence on the international system, and a variable that captures domestic demand for protectionism. By following these caveats, prospects, and theoretical implications, a continuous process of thinking and re-thinking international trade relations in order to bring about an egalitarian and non-dominated discourse in the realm of international political economy becomes, perhaps, possible and attainable. APPENDIX A UNITED STATES IMPORTS BY LEVELS OF PROCESSING U.S. imports are coded by the Standard International Tarde Classification (SITC) and my categorization is based on the two-digit SITC commodi ties. level of processing There are total of 62 two-digit SITCs and each consisted of: Level 1--17 SITCs, Level 2--lO SITCs, Level 3--4 SITCs,. Level 4--13 SITCs, Level 5--7 SITCs, and Level 6--11 SITCs. Level SITC Commodity l 00 animal (live) 02 dairy and egg 03 fish , O4 cereals 05 vegetable and fruit 07 coffee and tea 08 feeding stuff 12 tobacco 21 hide, skin and furskin 22 oil seed and nuts 26 textile fiber»/ 27 crude fertilizer 28 metal ores 29 animal, vegetable material 32 coal 33 petroleum 34 natural gas 157 158 06 sugar and syrups 23 rubber (crude) 24 wood (log, lumber) 25 pulp 41 animal oil and fat 42 vegetable oil and fat 43 processed animal oil 61 leather u 11 beverage (wine, beer, spirit) 65 textile yarn, fabric, tile 9 66 nonmetallic mineral (lime, glass) 68 nonferrous metal (silver, copper) 51 organic chemical 52 inorganic chemical 53 dyeing, tanning material 55 perfume material 56 fertilizer 59 chemical product (pesticide) 63 wood manufactured 64 paper and paper manufactured 67 iron, steel (plate, pipe, pig iron) , 81 light fixture 82 furniture 84 wearing apparel 85 footwear 54 medical and pharmaceutical product 57 explosive and pyrotechnical product 58 synthetic and plastic material” 62 tire and tube (2 69 metal product (structure, nail, tool), 83 travel good 88 photographic apparatus 71 power generating machine 72 electrical machine 73 metal working machine 74 general industrial machine~' 75 office machine (typewriter, calculator) 76 telecommunication machine (TV, radio) 8* 77 electrical machine (circuit, transistor)vt 78 road vehicle 79 other transportation equipment 87 scientific and professional instrument~ 89 other manufactured product (piano) , APPENDIX B A SIMULTANEOUS MODEL OF INTERNATIONAL TRADE As mentioned earlier, there seems to be a gap between the economic analyses which focus on import penetration and the political analyses which highlight protectionism. Although most are unaware, some theorists have argued that because import penetration and protectionism are essentially the same event (i.e., different sides of same coin) they need not or cannot be view separately. However, my contention is that they must be treated as individual outcomes. If this is the case, these two events are indeed simultaneous. Consequently, it is only reasonable to assume that there is a feedback effect between import penetration on the one side and protectionism on the other side. As import penetration increases the demand for protectionism, at the same time, increases. A simultaneous model of import penetration can be formalized as follows: Y1 = r12Y2 + B11X1 + B12X2 + B13x3 + B15X5 + U1t (8 1) Y2 = rZIYl + Bszz + 823X3 + BZ4X4 + 325X5 + UZt (3.2) where Y1 import penetration Y2 a trade dispute X1 = comparative advantage X2 = exchange rate X3 = GNP growth rate X4 = U.S. foreign aid X5 = NIC’s export policy Uts = error term 159 160 For the simultaneous model, we must examine the questions concerning identification and estimation. Identification is a question of how much prior information is needed in specifying a model. It is a theoretical question as to whether a different form of the structural model could have generated the observed relationship. Specifically, for a simultaneous model to be identified, two conditions must be fulfilled: the order condition, which is a necessary condition and the rank condition which is a necessary and sufficient condition for identification. The order condition requires that the number of excluded pre- determined variables must be at least as great as the number of endogenous variables minus one.1 For the equations (4.1) and (4.2), we have excluded one predetermined variable (X4 and X1, respectively) and have included two endogenous variables (Y1 and Y2). For the rank condition, the equations (8.1) and (B.2) have the rank of one while both equations include two endogenous variables.2 Therefore, both equations are said to be "just identified,” fulfilling the requirements of the order and rank conditions. For identified simultaneous models, estimation techniques fall 1This condition can be stated as: [K2 2 M1 - l], where K2 = the number of excluded predetermined variables (i.e., exogenous and lagged endogenous variables) and M1 = the number of included endogenous variables. See, Judge, et al., (1985: 561-588) for more detailed illustration of this condition. 2The rank condition can be stated as: [rank (R1D) = M-l], where R1 = the excluded endogenous and exogenous variables; 0 . [matrix (r,B)]; and M = the number of endogenous variables. As such, for the equation (4.1), rank (R10 = [O B 4]) = 1; (M-1) = (2-1) - 1. And for the equation (4.2), rank (R10 = [B12 0]) = 1; (M-1) . (2-1) a 1. 161 into two classes: single equation technique and system wide technique. The coefficients can be estimated through the single equation techniques such as two stage least squares and through system wide techniques as LISREL, a full information system. An actual technique selected for the analysis hinges on the availability of data and is subjected to the budget constraints. APPENDIX C CORRELOGRAMS OF AUTOCORRELATION AND PARTIAL AUTOCORRELATION FUNCTIONS The correlograms of autocorrelation function (ACF) and partial autocorrelation function (PACF) are the useful diagnostic tools for detecting and specifying autoregressive processes. Although visual inspection of these correlograms is not an easy task, the following examples may provide some guidelines. Note that (*) represents autocorrelation coefficient and the boundary (.) indicates two standard error limits. I. The Correlograms of Autocorrelation Functions: 1. ACF showing non-stationary series. lag Box-Ljung prob. 1 ********.******* 2 *******.** 3 **** 4 * 5 *** 6 **** 7 . ***** 8 **** 9 . **** 10 . **** 11 . *** 12 . *** 13 * 14 . * 15 . ** 162 2. 18‘ K>G3\JO\U1#-0)h)hd 163 Box-Ljung prob. ******* ******* ******* .********** .****** .** *******. *** * * . *** . **** ****** **.***** ****.**** ****.**** ****.*** **.*** ACF indicating stationary series. _| a) U3 KDGJ‘JO\U1£-03h)hd Box-Ljung prob. . *** . ****** **** *** *** *** . * .***** .***** * ** * * ** * 164 II. The Correlograms of Partial Autocorrelation Function: 1. PACF showing the first autoregressive process. lag Box-Ljung prob. 1 . ******** . *********** 2 ****** 3 ** 4 i: 5 *** 6 * 7 '1: 8 *** 9 . * lo . **~k** 11 . *** 12 . ** 13 . * 14 . ** 15 . * 2. PACF showing no autocorrelation. ...I m 4.0 Box-Ljung prob. * **** *‘k‘k‘k **** 'k 'k * ** . *‘k 10 . ** 1 1 . ****** 12 . ** 13 . * 14 . * 15 . ** tomNChU'l-FWNH APPENDIX 0 RESULTS OF COEFFICIENT ESTIMATES FOR THE INDIVIDUAL MODEL This appendix presents the coefficient estimates of the individual models (i.e., equations) at the commodity level of analysis for Singapore, Korea, Hong Kong, and Taiwan. Table 0.1: Coefficient Estimates for the Interdependence Equations at the Commodity Level of Analysis SIG KOR HK TAN b1) bil b'1 b11 (Stand. rror) (Stand. Error) (Stand. Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) (t-ratio) CALI .022 .002 .018 .104 (.146) (.041) (.006) (.110) (.153) (.485) (2.936) (.945) CAL2 -.167 -.007 -.002 -.140** (.197) (.005) (.001) (.063) (-.847) (-.159) (-1.604) (-2.227) CAL3 .101 -.050 .005 .027 (.064) (.063) (.038) (.034) (1.574) (-.793) (.128) (.800) CAL4 .066 -.266 -.015 -.355** (.167) (.271) (.143) (.175) (.395) (-.982) (-.107) (-2.023) CAL5 -.005 -.012 -.080 .051 (.050) (.024) (.073) (.069) (-.094) (-.492) (-1.097) (.745) CAL6 -.638* -.384* .142 -.378 (.335) .218) (.162) (.422) (-1.904) (-1.764) (.880) (-.218) *Significantly different from 0.00 at .10 level. **Significantly different from 0.00 at .05 level. 165 166 Table 0.2: Coefficient Estimates of the Dependence Equations at the Commodity Level of Analysis for Singapore ER GNP UEP SIG b'z D13 5 4 (Stand? Error) (Stand. Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) L1 —0.119 -0.143 0.765 (0.244) (0.929) (1.845) (~0.487) (-0.153) (0.414) L2 -0.088 -3.317 -7.189** (0.335) (1.275) (2.533) (-0.262) (-2.601) (-2.839) L3 0.047 0.197 0.395 (0.078) (0.295) (0.587) (0.608) (0.668) (0.673) L4 -0.272 -0.582 -2.228 (0.208) (0.792) (1.573) (-1.307) (-0.734) (-1.417) L5 -0.039 0.154 0.001 (0.064) (0.245) (0.486) (-0.613) (0.629) (0.001) L6 0.604 3.895** 8.555 (0.477) (1.815) (3.604) (1.267) (2.146) (2.374) **Significantly different from 0.00 at .05 level. 167 Table 0.3: Coefficient Estimates of the Dependence Equations at the Commodity Level of Analysis for Koreaa ER GNP UEP KOR b1 013 0'4 (Stand.2Error) (Stand. Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) L1 -0.104 -0.414 -0.497 (0.033) (0.192) (0.367) (-3.169) (-2.160) (-1.356) L2 -0.004 0.014 0.051 (0.004) (0.024) (0.046) (-0.922) (0.570) (1.118) L3 0.058 0.080 0.087 (0.043) (0.249) (0.476) (1.364) (0.321) (0.182) L4 0.008 -0.006 0.261 (0.236) (1.374) (2.627) (0.036) (-0.005) (0.099) L5 0.014 0.214 0.191 (0.025) (0.148) (0.283) (0.574) (1.448) (0.674) L6 -0.127 -0.551 -l.274 (0.187) (1.093) (2.088) (-0.676) (-0.504) (-0.610) aAll coefficients not significantly different from 0.00 at .05 level. 168 Table 0.4: Coefficient Estimates of the Dependence Equations at the Commodity Level of Analysis for Hong Konga ER GNP UEP HK b b b (StandIZError) (Stand13Error) (Stand? Error) (t-ratio) (t-ratio) (t-ratio) L1 -0.011 -0.039 -0.74 (0.004) (0.029) (0.060) (-3.028) (-1.344) (—1.239) L2 0.001 -0.011 -0.020 (0.001) (0.007) (0.014) (1.265) (-1.617) (-1.472) L3 -0.015 -0.030 -0.431 (0.032) (0.218) (0.444) (-0.479) (-0.138) (-0.971) L4 -0.011 0.343 1.319 (0.094) (0.702) (1.432) (-0.120) (0.488) (0.921) L5 -0.007 0.185 -O.149 (0.055) (0.416) (0.849) (-0.123) (0.445) (-0.l75) L6 0.051 -0.394 -0.529 (0.102) (0.796) (1.622) (0.505) (-0.494) (-0.326) 0.00 at .05 level. aAll coefficients not significantly different from 169 Table 0.5: Coefficient Estimates of the Dependence Equations at the Commodity Level of Analysis for Taiwana ER GNP UEP TAN D12 b13 b 4 (Stand. Error) (Stand. Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) L1 0.018 -0.086 0.098 (0.284) (0.471) (0.930) (0.064) (-0.182) (0.105) L2 -O.186 0.182 1.047 (0.189) (0.314) (0.620) (-0.985) (0.581) (1.690) L3 -0.073 0.077 0.187 (0.075) (0.105) (0.200) (-O.971) (0.733) (0.938) L4 -0.039 0.090 1.096 (0.445) (0.739) (1.459) (-0.088) (0.122) (0.751) L5 -O.149 0.114 0.012 (0.176) (0.292) (0.576) (-0.847) (0.390) (0.022) L6 0.456 0.121 -1.734 (0.478) (0.793) (1.565) (0.954) (0.153) (-1.108) aAll coefficients not significantly different from 0.00 at .05 level. 170 Table 0.6: Coefficient Estimates of the Mercantilist Equations at the Commodity Level for Singapore T01 CF GE USTA SIG b1 01 51' ['11 (Stand. Error) (Stand. Error) (Stand. Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) (t-ratio) L1 -0.064 ~0.062 -0.575 -0.058 (0.106) (0.045) (0.267) (0.032) (-0.597) (-1.381) (-2.152) (-1.840) L2 0.192 -0.072 -0.469 -0.007 (1.110) (0.096) (0.570) (0.075) (0.173) (-0.756) (-0.823) (-0.090) L3 -0.062 0.027 0.104 -0.001 (0.056) (0.015) (0.090) (0.011) (-1.109) (1.822) (1.150) (-0.l33) L4 0.035 -0.099 -0.090 0.035 (0.018) (0.044) (0.263) (0.031) (1.902) (~2.234) (-O.342) (1.101) L5 0.035 -0.010 0.088 0.013 (0.046) (0.015) (0.089) (0.011) (0.766) (-O.659) (0.985) (1.140) L6 0.333 0.221** 0.935 0.024 (0.206) (0.084) (0.488) (0.059) (1.614) (2.630) (1.917) (0.409) **Significantly different from 0.00 at .05 level. 171 Table 0.7: Coefficient Estimates of the Mercantilist Equations at the Commodity Level for Korea T01 CF GE USTA KOR b1 b1 b1 01 (Stand. Error) (Stand. Error) (Stand. Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) (t-ratio) Ll 0.009 0.049 0.072 0.001 (0.030) (0.020) (0.111) (0.002) (0.301) (2.427) (0.648) (0.587) L2 -0.001 0.000 -0.011 0.000 (0.014) (0.002) (0.013) (0.000) (-0.040) (0.002) (-0.878) (1.196) L3 0.025 -0.013 0.008 -0.002 (0.066) (0.030) (0.134) (0.002) (0.383) (-0.449) (0.063) (-0.821) L4 -0.010 -0.086 -0.379 -0.002 (0.033) (0.130) (0.724) (0.011) (-0.320) (-0.663) (-0.524) (-0.157) L5 0.018 0.014 0.131 —0.001 (0.021) (0.014) (0.070) (0.001) (0.858) (0.973) (1.866) (-0.700) L6 0.153 0.202** 0.290 -0.002 (0.156) (0.103) (0.546) (0.008) (0.981) (1.963) (0.530) (-O.239) **Significantly different from 0.00 at .05 level. 172 Table 0.8: Coefficient Estimates of the Mercantilist Equations at the Commodity Level for Hong Konga T01 CF GE USTA SIG b1 b1 b1 01 (Stand. Error) (Stand. Error) (Stand. Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) (t-ratio) L1 -0.003 0.001 -0.003 0.002 (0.005) (0.003) (0.028) (0.006) (—0.490) (0.405) (-O.106) (0.304) L2 0.007 —0.001 -0.004 0.001 (0.004) (0.001) (0.005) (0.001) (1.579) (-1.022) (-0.784) (0.859) L3 0.079 0.006 0.135 -0.012 (0.045) (0.017) (0.155) (0.034) (1.767) (0.340) (0.872) (—0.358) L4 -0.006 -0.036 -0.557 0.140 (0.018) (0.061) (0.521) (0.114) (-O.353) (-0.590) (-1.069) (1.228) L5 0.090 0.028 0.246 —0.061 (0.067) (0.036) (0.302) (0.066) (1.333) (0.772) (0.815) (-0.922) L6 0.099 -0.009 0.023 -0.000 (0.137) (0.076) (0.638) (0.149) (0.718) (-0.115) (0.036) (-0.001) aAll coefficients not significantly different from 0.00 at .05 level. 173 Table 0.9: Coefficient Estimates of the Mercantilist Equations at the Commodity Level for Taiwan 1 T01 CF GE USTA SIG b1 01' 01' 01 (Stand. Error) (Stand. Error) (Stand. Error) (Stand. Error) (t—ratio) (t-ratio) (t-ratio) (t-ratio) L1 0.077 -0.004 -0.087 -0.003 (0.074) (0.047) (0.158) (0.009) (1.038) (-0.088) (-0.549) (-0.360) L2 -0.422** 0.049 0.182** -0.000 (0.131) (0.022) (0.073) (0.004) (-3.234) (2.198) (2.482) (-0.125) L3 0.021 -0.008 0.008 0.002 (0.037) (0.014) (0.058) (0.002) (0.552) (-0.570) (0.139) (0.827) L4 -0.008 0.069 0.332 -0.004 (0.021) (0.079) (0.252) (0.015) (-0.356) (0.878) (1.316) (-0.250) L5 0.013 -0.033 -0.016 0.004 (0.050) (0.032) (0.102) (0.005) (0.249) (—1.018) (-0.159) (0.765) L6 -0.039 -0.097 -0.251 0.002 (0.159) (0.101) (0.314) (0.018) (-0.247) (—0.953) (-0.800) (0.141) **Significantly different from 0.00 at .05 level. APPENDIX E COEFFICIENT ESTIMATES OF THE COMPOSITE MODEL ANALYSIS This appendix presents the coefficient estimates for the determinants of the individual models for the composite model analysis. The results of the regional and national level estimates are presented first, followed by the coefficient estimates for the levels of commodities. Table E.1: Coefficient Estimates of the Interdependence and Dependency Model Equations at Regional and National Level CA ER GNP UEP 01 b 0 01 (Stand. Error) (Stand.1Error) (Stand.iError) (Stand. Error) (t-ratio) (t-ratio) (t—ratio) (t-atio) NICs -0.208* 0.002 —0.056 -0.139 (0.120) (0.053) (0.198) (0.342) (-1.734) (0.045) (-0.285) (-0.407) SIG -0.004 -0.005 0.026 0.028 (0.005) (0.005) (0.020) (0.038) (-0.666) (-1.055) (1.299) (0.723) KOR -0.014 -0.016 -0.046 -0.095 (0.031) (0.018) (0.095) (0.173) (~0.459) (-0.886) (—0.482) (-0.547) HK -0.007 0.001 0.060 0.071 (0.031) (0.011) (0.061) (0.119) (-0.230) (0.076) (0.983) (0.595) TAN -0.040 0.016 0.088 0.112 (0.035) (0.082) (0.093) (0.115) (-1.138) (0.191) (0.953) (0.974) *Significantly different from 0.00 at .10 level. 174 175 Table E.2: Coefficient Estimates of the Mercantilist Model Equations at Regional and National Levela TDl CF GE USTA b b b b (Stand.iError) (Stand.iError) (Stand.iError) (Stand.iError) (t-ratio) (t-ratio) (t-ratio) (t-ratio) NICs 0.003 -0.042 0.086 0.005 (0.003) (0.024) (0.097) (0.004) (0.984) (-1.767) (0.889) (1.166) SIG 0.001 -0.000 -0.000 —0.000 (0.000) (0.001) (0.007) (0.001) (2.582) (-0.091) (-0.015) (-0.431) KOR 0.000 -0.007 0.028 0.000 (0.001) (0.008) (0.029) (0.000) (0.305) (-0.861) (0.982) (1.027) HK 0.000 -0.006 -0.047 0.012 (0.001) (0.005) (0.034) (0.010) (0.192) (-l.146) (-1.372) (1.293) TAN 0.001 -0.005 -0.015 -0.000 (0.001) (0.010) (0.022) (0.001) (1.287) (-0.487) (-0.704) (-0.224) aAll coefficients not significantly different from 0.00 at .05 level. 176 Table E.3: Coefficient Estimates of the Interdependence and Dependency Model Equations at Commodity Level for NICs CA ER GNP UEP NICs 01' b1 b1 b1 (Stand. Error) (Stand. Error) (Stand. Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) (t-atio) Ll 0.001 0.004 0.037 0.108 (0.007) (0.013) (0.053) (0.101) (0.146) (0.275) (0.699) (1.067) L2 -0.039* 0.088** -0.290 -0.554*** (0.020) (0.043) (0.101) (0.219) (-1.890) (2.042) (-2.884) (—2.527) L3 -0.069 0.028 -0.107 -0.358 (0.051) (0.049) (0.202) (0.342) (-l.352) (0.563) (-0.531) (-1.047) L4 0.068 -0.017 0.139 1.795 (0.158) (0.274) (0.910) (2.039) (0.428) (-0.062) (0.153) (0.880) L5 0.052 -0.152 0.262 1.241 (0.086) (0.131) (0.630) (1.226) (0.609) (+1.165) (0.415) (1.013) L6 -0.119 -0.163 —0.251 -0.168 (0.073) (0.078) (0.269) (0.476) (-1.634) (-2.091) (-0.931) (-0.352) *Significantly different from 0.00 at .10 level. **Significantly different from 0.00 at .05 level. ***Significantly different from 0.00 at .01 level. 177 Table E.4: Coefficient Estimates of the mercantilist Model Equations at Commodity level for NICs T01 CF GE USTA NICs b b b b1 . 1 . (Stand. Error) (Stand.1Error) (Stand. Error) (Stand.1Error) (t-ratio) (t-ratio) (t-ratio) (t-ratio) L1 -0.001 -0.010 0.011 0.000 (0.005) (0.006) (0.024) (0.001) (—0.238) (-1.730) (0.447) (0.234) L2 0.113 -0.011 -0.128 0.012*** (0.080) (0.011) (0.052) (0.004) (1.401) (-0.993) (-2.445) (3.125) L3 0.016 0.010 0.149 -0.000 (0.025) (0.021) (0.100) (0.003) (0.666) (0.470) (1.495) (-0.100) L4 -0.004 -0.086 -0.295 0.022 (0.017) (0.092) (0.432) (0.024) (-0.258) (-0.939) (-0.683) (0.950) L5 0.057 -0.036 -0.088 -0.002 (0.093) (0.062) (0.259) (0.015) (0.608) (-0.592) (-0.341) (-0.104) L6 0.041 -0.045 0.034 -0.015 (0.032) (0.027) (0.121) (0.007) (1.278) (-1.688) (0.281) (-2.050) ***Significantly different from 0.00 at .01 level. 178 Table E.5: Coefficient Estimates of the Interdependence and Dependency Model Equations at the Commodity Level for Singapore CA ER GNP UEP SIG b1 D1 01 01 (Stand. Error) (Stand. Error) (Stand. Error) (Stand. Error) (t-ratio) (t-ratio) (t—ratio) (t-atio) L1 0.104 0.228 0.469 3.498 (0.097) (0.193) (0.704) (1.421) (1.070) (1.183) (0.666) (2.462) L2 -0.265 0.221 -4.367 -7.674** (0.265) (0.468) (1.852) (3.427) (-1.000) (0.472) (-2.358) (-2.239) L3 0.038 0.037 0.466 0.318 (0.025) (0.089) (0.311) (0.621) (1.521) (0.414) (1.497) (0.511) L4 0.090 -0.270 -0.036 -0.787 (0.251) (0.301) (1.224) (2.284) (0.358) (-0.899) (-0.030) (-0.345) L5 0.005 -0.143 -0.029 -0.238 (0.059) (0.095) (0.330) (0.654) (0.087) (-1.515) (-0.088) (-0.364) L6 -0.266 0.155 3.182* 3.388 (0.565) (0.458) (1.726) (3.826) (-0.471) (0.338) (1.843) (0.886) *Significantly different from 0.00 at .10 level. **Significantly different from 0.00 at .05 level. 179 Table E.6: Coefficient Estimates of the Mercantilist Model Equations at the Commodity Level for Singapore T01 CF GE USTA SIG b Tb b 1 1 1 b1 (Stand. Error) (Stand. Error) (Stand. Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) (t-ratio) L1 -0.098 -0.l66 -0.661 -0.013 (0.080) (0.045) (0.244) (0.028) (—1.220) (-3.717) (-2.712) (-0.456) L2 -0.805 -0.131 -0.312 0.006 (1.317) (0.144) (0.596) (0.087) (-0.611) (-0.908) (-0.522) (0.071) L3 -0.087 0.058*** 0.004 -0.009 (0.055) (0.019) (0.112) (0.012) (-1.573) (3.028) (0.033) (-0.797) L4 0.031 -0.058 0.083 0.029 (0.031) (0.070) (0.463) (0.047) (0.999) (-0.829) (0.178) (0.625) L5 -0.006 0.002 0.187 0.006 (0.055) (0.021) (0.123) (0.014) (-0.115) (0.075) (1.512) (0.448) L6 0.174 0.310** 0.631 0.005 (0.213) 0.125) (0.651) (0.068) (0.818) 2.483) (0.969) (0.071) **Significantly different from 0.00 at .05 level. ***Significantly different from 0.00 at .01 level. 180 Table E.7: Coefficient Estimates of the Interdependence and dependency Model Equations at the Commodity Level for Korea CA ER GNP UEP KOR 01 b1 b1 01 (Stand. 1rror) (Stand. Error) (Stand. Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) (t-atio) L1 0.003 -0.203 -0.500 -0.636 (0.055) (0.075) (0.278) (0.557) (0.061) (-2.710) (-1.802) (-1.143) L2 0.001 0.005 0.054 0.124 (0.012) (0.013) (0.087) (0.162) (0.075) (0.407) (0.619) (0.767) L3 -0.082 -0.006 -0.100 -0.234 (0.116) (0.099) (0.466) (0.844) (-0.711) (-0.062) (-0.214) (-0.277) L4 -1.413** -0.721 -2.023 -2.682 (0.676) (0.475) (2.113) (3.707) (—2.091) (-1.516) (-0.957) (-0.724) L5 -0.024 0.058 0.467* 0.645 (0.032) (0.048) (0.263) (0.469) (-0.744) (1.212) (1.774) (1.374) L6 -0.500 -0.188 -0.l94 -0.313 (0.341) (0.375) (1.729) (3.092) (-1.469) (-0.502) (-0.112) (-0.101) *Significantly different from 0.00 at .10 level. **Significantly different from 0.00 at .05 level. 181 Table E.8: Coefficient Estimates of the Mercantilist Model Equations at Commodity Level for Korea T01 CF GE USTA KOR b1 b1 b1 01 (Stand. Error) (Stand. Error) (Stand. Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) (t-ratio) L1 0.081 0.000 0.269** -0.001 (0.041) (0.028) (0.127) (0.002) (1.969) (0.001) (2.115) (-0.449) L2 0.014 0.004 -0.016 0.000 (0.023) (0.006) (0.015) (0.000) (0.608) (0.642) (-1.051) (1.071) L3 0.011 -0.020 -0.009 -0.001 (0.086) (0.047) (0.169) (0.003) (0.130) (-0.421) (-0.051) (-0.430) L4 0.018 -0.224 —0.273 0.003 (0.043) (0.201) (0.754) (0.013) (0.412) (-1.110) (-0.361) (0.242) L5 -0.030 0.031 0.060 0.000 (0.038) (0.020) (0.079) (0.002) (-0.785) (1.536) (0.754) (0.044) L6 0.189 0.133 -0.147 -0.006 (0.186) (0.166) (0.738) (0.010) (1.015) (0.802) (-0.l99) (-0.543) **Significantly different from 0.00 at .05 level. 182 Table E.9: Coefficient Estimates of the Interdependence and Dependency Model Equations at Commodity Level for Hong Kong CA ER GNP UEP HK 51 b1 b1 b1 (Stand. Error) (Stand. Error) (Stand. Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) (t-atio) L1 0.030 0.004 0.012 0.004 (0.009) (0.007) (0.030) (0.054) (3.402) (0.549) (0.399) (0.074) L2 -0.006*** -0.002 -0.005 -0.025 (0.002) (0.001) (0.008) (0.016) (-2.688) (-l.214) (-0.575) (-l.548) L3 -0.041 -0.025 -0.221 -l.058** (0.064) (0.048) (0.246) (0.485) (-0.645) (-0.533) (-0.899) (-2.180) L4 -0.047 -0.032 0.629 2.110 (0.251) (0.142) (0.967) (1.770) (-0.185) (-0.225) (0.650) (1.192) L5 -0.126 -0.006 -0.017 -0.658 (0.133) (0.074) (0.693) (1.282) (-0.948) (-0.080) (-0.025) (-0.513) L6 0.298 0.021 -l.029 -2.665 (0.344) (0.189) (1.075) (2.243) (0.868) (0.111) (-0.958) (~1.188) **Significantly different from 0.00 at .05 level. ***Significantly different from 0.00 at .01 level. 183 Table E.10: Coefficient Estimates of the Mercantilist Model Equations at Commodity Level for Hong Konga TDl CF GE USTA HK bi bi bi bi (Stand. Error) (Stand. Error) (Stand. Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) (t-ratio) Ll -0.004 0.005 -0.009 -0.001 (0.004) (0.003) (0.019) (0.005) (-0.925) (1.572) (-0.452) (-0.165) L2 0.010 -0.001 -0.014 0.004 (0.005) (0.001) (0.007) (0.002) (1.902) (-l.557) (-2.047) (1.925) L3 0.110 0.021 0.032 0.039 (0.046) (0.019) (0.165) (0.041) (2.394) (1.077) (0.194) (0.942) L4 -0.009 -0.069 -0.420 0.089 (0.029) (0.075) (0.775) (0.223) (-0.307) (-0.928) (-0.542) (0.398) L5 0.004 0.055 0.394 -0.068 (0.132) (0.040) (0.452) (0.100) (0.032) (1.376) (0.873) (-0.688) L6 0.235 -0.025 -0.153 0.209 (0.207) (0.112) (0.789) (0.247) (1.136) (-0.223) (-0.l94) (0.845) aAll coefficients not significantly different from 0.00 at .05 level. 184 Table E.11: Coefficient Estimates of the Interdependence and Dependency Model Equations at Commodity Level for Taiwana CA ER GNP UEP TAN b1 b° b1 b1 (Stand. Error) (Stand.1érror) (Stand. Error) (Stand. Error) (t-ratio) (t-ratio) (t-ratio) (t-atio) L1 0.013 -0.844 -0.734 0.981 (0.221) (0.779) (0.961) (1.224) (0.058) (-1.084) (-0.764) (0.802) L2 -0.162 0.020 0.002 -0.535 (0.110) (0.392) (0.441) (0.632) (-l.473) (0.052) (0.005) (-0.848) L3 0.066 -0.086 0.030 0.223 (0.062) (0.182) (0.218) (0.256) (1.072) (-0.473) (0.138) (0.873) L4 -0.313 0.150 0.051 -0.072 (0.314) (1.218) (1.324) (1.956) (-0.996) (0.123) (0.039) (-0.037) L5 -0.092 -0.797 -0.681 -0.204 (0.121) (0.552) (0.650) (0.854) (-0.762) (-1.444) (-1.048) (-0.238) L6 -0.607 1.504 1.290 -2.985 (0.486) (1.359) (1.450) (2.126) (-1.248) (1.107) (0.889) (-1.404) aAll coefficients not significantly different from 0.00 at .05 level. 185 Table E.12: Coefficient Estimates of the Mercantilist Model Equations at Commodity Level for Taiwana TDl CF GE USTA TAN b- b- b- b' (Stand.IError) (Stand.]grror) (Stand.1Error) (Stand.1Error) (t-ratio) (t-ratio) (t-ratio) (t-ratio) Ll 0.087 -0.145 -0.155 -0.004 (0.097) (0.108) (0.209) (0.012) (0.897) (-1.338) (~0.740) (-0.350) L2 —0.368 0.061 0.099 0.001 (0.157) (0.050) (0.105) (0.005) (-2.341) (1.222) (0.942) (0.177) L3 0.025 -0.023 0.058 0.002 (0.028) (0.026) (0.074) (0.003) (0.903) (-0.921) (0.784) (0.855) L4 -0.017 0.080 0.226 -0.007 (0.028) (0.170) (0.310) (0.020) (-0.626) (0.469) (0.727) (-0.367) L5 0.039 -0.082 -0.086 -0.002 (0.063) (0.073) (0.125) (0.007) (0.609) (-1.130) (-0.686) (-0.297) L6 0.057 0.243 0.046 -0.003 (0.241) (0.201) (0.332) (0.030) (0.236) (1.212) (0.138) (-0.092) aAll coefficients not significantly different from 0. .05 level. 00 at APPENDIX F CHANGES IN COEFFICIENT ESTIMATES BETWEEN THE INDIVIDUAL AND COMPOSITE MODELS This appendix presents the differences between the individual and composite model coefficient estimates for the commodity level of analysis. The differences are computed as the percentage of the differences in absolute values. Note that ”I" denotes the individual model estimates and "C" signifies the composite model estimates. Table F.1: Percentage Changes in Coefficient Estimates between the Individual and Composite Estimates CA-I CA-C DIFF ER-I ER-C DIFF NICLl 0.000 0.001 0.0% 0.020 0. . NICLZ -0.001 -0.039 3800.0% -0.018 0.088 588.9% NICL3 -0.023 -0.069 200.0% 0.045 0. . NICL4 -0.025 0.068 372.0% 0.049 -0.017 134.7% NICL5 -0.035 0.052 248.6% -0.044 -0.152 245.5% NICL6 -0.039 -0.119 205.1% 0.007 -0.163 2428.6% SIGLl 0.022 0.104 372.7% -0.119 0.228 291.6% SIGLZ -0.167 -0.265 58.7% -0.088 0.221 351.1% SIGL3 0.101 0.038 62.4% 0.047 0.037 21.3% SIGL4 0.066 0.090 36.4% -0.272 -0.270 0.7% SIGLS -0.005 0.005 200.0% -0.039 -0.143 266.7% SIGL6 -0.638 -0.266 58.3% 0.604 0.155 74.3% KORLl 0.020 0.003 85.0% —0.104 -0.203 95.2% KORLZ -0.001 0.001 200.0% -0.004 0.005 225.0% KORL3 -0.050 -0.082 64.0% 0.058 -0.006 110.3% KORL4 -0.266 -1.413 431.2% 0.008 -0.721 9112.5% KORLS -0.012 -0.024 100.0% 0.014 0.058 314.3% KORL6 -0.384 -0.500 30.2% -0.127 -0.188 48.0% HKLl 0.018 0.030 66.7% -0.011 0.004 136.4% HKL2 -0.002 -0.006 200.0% 0.001 -0.002 300.0% HKL3 0.005 -0.041 920.0% -0.015 -0.025 66.7% HKL4 -0.015 -0.047 213.3% -0.011 -0.032 190.9% HKLS -0.080 -0.126 57.5% -0.007 -0.006 14.3% HKL6 0.142 0.298 109.9% 0.051 0.021 58.8% TANLl 0.104 0.013 87.5% 0.018 -0.844 4788.9% TAHLZ -0.140 -0.162 15.7% -0.186 0.020 110.8% TANL3 0.027 0.066 144.4% -0.073 -0.086 17.8% TANL4 -0.355 -0.313 11.8% -0.039 0.150 484.6% TANLS 0.051 -0.092 280.4% -0.149 -0.797 434.9% TANL6 -0.378 -0.607 60.6% 0.456 1.504 229.8% GNP-I GNP-C DIFF UEP-I UEP-C DIFF NICLl 0.064 0.037 42.2% 0.068 0.108 58.8% NICLZ -0.135 -0.290 114.8% -0.080 -0.554 592.5% NICL3 0.044 -0.107 343.2% 0.058 -0.358 717.2% NICL4 0.295 0.139 52.9% 1.455 1.795 23.4% NICLS 0.286 0.262 8.4% 0.782 1.241 58.7% NICL6 0.142 -0.251 276.8% 0.139 -0.168 220.9% SIGLl -0.143 0.469 428.0% 0.765 3.498 357.3% SIGLZ -3.317 -4.367 31.7% -7.189 -7.674 6.7% SIGL3 0.197 0.466 136.5% 0.395 0.318 19.5% SIGL4 -0.582 -0.036 93.8% -2.228 -0.787 64.7% SIGLS 0.154 -0.029 118.8% 0.001 -0.238 23900% SIGL6 3.895 3.182 18.3% 8.555 3.388 60.4% KORLl -0.414 -0.500 20.8% -0.497 -0.636 28.0% KORLZ 0.014 0.054 285.7% 0.051 0.124 143.1% KORL3 0.080 -0.100 225.0% 0.087 -0.234 369.0% KORL4 -0.006 ~2.023 33617% 0.261 -2.682 1127.6% KORLS 0.214 0.467 118.2% 0.191 0.645 237.7% KORL6 -0.551 -0.194 64.8% -l.274 -0.313 75.4% HKLl -0.039 0.012 130.8% -0.074 0.004 105.4% HKL2 -0.011 -0.005 54.5% -0.020 -0.025 25.0% HKL3 -0.030 -0.221 636.7% -0.431 -1.058 145.5% HKL4 0.343 0.629 83.4% 1.319 2.110 60.0% HKLS 0.185 -0.017 109.2% -0.149 -0.658 341.6% HKL6 -0.394 -1.029 161.2% -0.529 -2.665 403.8% TANLl -0.086 -0.734 753.5% 0.098 0.981 901.0% TAWLZ 0.182 0.002 98.9% 1.047 -0.535 151.1% TANL3 0.077 0.030 61.0% 0.187 0.223 19.3% TANL4 0.090 0.051 43.3% 1.096 -0.072 106 6% TAWLS 0.114 -0.681 697.4% 0.012 -0.204 1800 0% TANL6 0.121 1.290 966.1% -1.734 -2.985 72 1% 188 TD-I TD-C DIFF CF-I CF-C DIFF NICLl 0.000 -0.001 0.0% -0.008 -0.010 25.0% NICLZ 0.068 0.113 66.2%, 0.017 -0.011 164.7% NICL3 0.016 0.016 0.0% 0.003 0.010 233.3% NICL4 0.016 -0.004 125.0% -0.002 -0.086 4200.0% NICL5 0.051 0.057 11.8% -0.028 -0.036 28.6% NICL6 0.034 0.041 20.6% -0.019 —0.045 136.8% SIGLl -0.064 -0.098 53.1% -0.062 -0.160 158.1% SIGL2 0.192 -0.805 519.3% -0.072 -0.131 81.9% SIGL3 -0.062 -0.087 40.3% 0.027 0.058 114.8% SIGL4 0.035 0.031 11.4% -0.099 -0.058 41.4% SIGLS 0.035 -0.006 117.1% -0.010 0.002 120.0% SIGL6 0.333 0.174 47 7% 0.221 0.310 40.3% KORLl 0.009 0.081 800.0% 0.049 0.000 100.0% KORLZ -0.001 0.014 1500.0% 0.000 0.004 0.0% KORL3 0.025 0.011 56.0% ~0.013 -0.020 53.8% KORL4 -0.010 0.018 280.0% -0.086 -0.224 160.5% KORLS 0.018 -0.030 266.7% 0.014 0.031 121 4% KORL6 0.153 0.189 23.5% 0.202 0.133 34 2% HKLl -0.003 -0.004 33.3% 0.001 0.005 400.0% HKL2 0.007 0.010 42.9% -0.001 -0.001 0.0% HKL3 0.079 0.110 39.2% 0.006 0.021 250.0% HKL4 -0.006 -0.009 50.0% -0.036 -0.069 91.7% HKL5 0.090 0.004 95.6% 0.028 0.055 96.4% HKL6 0.099 0.235 137.4% -0.009 -0.025 177.8% TANLI 0.077 0.087 13.0% -0.004 -0.145 3525.0% TANLZ -0.422 -0.368 12.8% 0.049 0.061 24.5% TANL3 0.021 0.025 19.0% -0.008 -0.023 187.5% TAHL4 -0.008 -0.017 112.5% 0.069 0.080 15.9% TAHLS 0.013 0.039 200.0% -0.033 -0.082 148.5% TANL6 -0.039 0.057 246.2% -0.097 0.243 350 5% 189 GE-I GE-C DIFF TA-I TA-C DIFF NICLl 0.028 0.011 60.7% 0.001 0.000 100.0% NICLZ 0.049 -0.128 361.2% 0.002 0.012 500.0% NICL3 0.074 0.149 101.4% -0.003 0.000 100.0% NICL4 -0.181 -0.295 63.0% 0.040 0.022 45.0% NICLS -0.212 -0.088 58.5% 0.017 -0.002 111.8% NICL6 -0.025 0.034 236.0% -0.007 -0.015 114.3% SIGLl -0.575 -0.661 15.0% -0.058 -0.013 77.6% SIGLZ -0.469 -0.312 33.5% -0.007 0.006 185.7% SIGL3 0.104 0.004 96.2% -0.001 -0.009 800.0% SIGL4 -0.090 0.083 192.2% 0.035 0.029 17.1% SIGLS 0.088 0.187 112 5% 0.013 0.006 53.8% SIGL6 0.935 0.631 32 5% 0.024 0.005 79.2% KORLl 0.072 0.269 273.6% 0.001 -0.001 200.0% KORL2 -0.011 -0.016 45.5% 0.000 0.000 0.0% KORL3 0.008 -0.009 212.5% -0.002 -0.001 50.0% KORL4 -0.379 -0.273 28.0% -0.002 0.003 250.0% KORLS 0.131 0.060 54.2% -0.001 0.000 100.0% KORL6 0.290 -0.147 150.7% -0.002 -0.006 200.0% HKLl -0.003 -0.009 200.0% 0.002 -0.001 150.0% HKL2 ~0.004 -0.014 250.0% 0.001 0.004 300.0% HKL3 0.135 0.032 76.3% -0.012 0.039 425.0% HKL4 -0.557 -0.420 24.6% 0.140 0.089 36.4% HKL5 0.246 0.394 60.2% -0.061 -0.068 11.5% HKL6 0.023 -0.153 765.2% 0.000 0.209 0.0% TANLI -0.087 -0.155 78.2% ~0.003 -0.004 33.3% TAWLZ 0.182 0.099 45.6% 0.000 0.001 0.0% TANLB 0.008 0.058 625.0% 0.002 0.002 0.0% TAHL4 0.332 0.226 31.9% -0.004 -0.007 75.0% TANLS -0.016 -0.086 437 5% 0.004 -0.002 150.0% TAHL6 -0.251 0.046 118 3% 0.002 -0.003 250.0% APPENDIX G RESULTS OF F-TESTS BETWEEN THE INDIVIDUAL AND COMPOSITE MODELS The results of F-tests between the individual and composite models are presented for the aggregated regional, the aggregated national, and the commodity level of analysis for Singapore, Korea, Hong Kong, and Taiwan. As the standard means of evaluation between the models (e.g., whether the additional variable in the composite model produce a significant changes over the individual model), F- ratios are computed based on the R25 and the number of parameter estimated in the individual and composite model. Table G.1: F-ratio Between Individual and Composite Model for Regional and National level of Analysisa Type of Analy. INT DEP MER COMP I&C D&C M&C NICs 0.080 0.377 0.553 0.946 18.201 16.731 14.460 SIG 0.007 0.336 0.280 0.985 85.680 79.620 108.072 KOR 0.002 0.072 0.315 0.960 31.020 40.227 36.537 HK 0.023 0.206 0.196 0.761 3.534 3.724 4.737 TAN 0.253 0.649 0.474 0.966 27.354 17.040 33.014 COMP-Composite, I&C=between Interdependence and Composite, D&C=Dependency and Composite, and M&C=Mercantilist and Composite Model. 190 aNote that INTsInterdependence, DEP=Dependence, MERsMercantilist, 191 Table 6.2: F-ratio Between Individual and Composite Model for Commodity Clusters at Regional and National Level of Analysis Type of Analy. INT DEP MER COMP I&C D&C M&C NICLl 0.000 0.241 0.479 0.610 1.786 1.514 0.668 NICL2 0.001 0.335 0.248 0.851 6.527 5.545 8.103 NICL3 0.047 0.187 0.144 0.952 21.527 25.471 33.646 NICL4 0.006 0.398 0.496 0.979 53.069 44.359 46.050 NICL5 0.022 0.237 0.371 0.634 1.913 1.738 1.437 NICL6 0.016 0.046 0.341 0.814 4.897 6.599 5.082 SIGLl 0.001 0.096 0.436 0.699 2.981 3.610 1.968 SIGLZ 0.043 0.374 0.119 0.885 9.403 7.991 14.970 SIGL3 0.127 0.045 0.335 0.686 2.284 3.665 2.508 SIGL4 0.010 0.265 0.404 0.918 14.150 14.231 14.003 SIGL5 0.001 0.171 0.224 0.721 3.318 3.549 4.006 SIGL6 0.185 0.306 0.603 0.969 32.656 38.666 26.672 KORLl 0.014 0.440 0.362 0.427 0.924 -0.041 0.253 KORL2 0.002 0.173 0.166 0.886 9.987 11.275 14.225 KORL3 0.040 0.124 0.119 0.279 0.426 0.387 0.500 KORL4 0.057 0.004 0.058 0.815 5.278 7.904 9.223 KORL5 0.016 0.229 0.295 0.451 1.019 0.729 0.640 KORL6 0.163 0.052 0.246 0.789 3.809 6.274 5.775 HKLI 0.365 0.414 0.047 0.389 0.045 -0.065 1.120 HKL2 0.146 0.330 0.213 0.496 0.792 0.527 1.123 HKL3 0.001 0.209 0.294 0.773 3.876 3.968 4.212 HKL4 0.001 0.101 0.120 0.749 3.414 4.137 5.022 HKL5 0.074 0.124 0.199 0.880 7.707 11.784 11.406 HKL6 0.049 0.064 0.049 0.780 3.789 5.195 6.636 TANLl 0.053 0.028 0.099 0.713 2.965 4.304 4.825 TANL2 0.237 0.272 0.670 0.491 0.641 0.772 -0.794 TANL3 0.038 0.144 0.078 0.636 2.111 2.435 3.449 TANL4 0.204 0.107 0.208 0.695 2.071 3.471 3.594 TANL5 0.034 0.152 0.105 0.963 31.932 39.021 51.580 TANL6 0.048 0.270 0.153 0.795 4.696 4.619 7.058 APPENDIX H UNITED STATES BUDGET AND TRADE DEFICITS This appendix provides the magnitude and trend of the United States budget and trade deficits. Specifically, the trade deficit originating from the Pacific Basin NICs and Japan are presented to demonstrate the problem facing the United States. Table H.l: United States Budget and Trade Deficits, 1967-87 (5 Millions) Year Budget Trade U.S. Trade Balance with Deficit Deficit SIG KOR HK TAN Japan 1967 -8700 4714 50 297 -243 167 -304 1968 -25200 1299 73 312 -333 117 -1107 1969 3200 1963 97 408 -451 5 —1398 1970 -2800 3272 159 273 -538 -22 -1223 1971 -23000 -1465 179 219 -567 -307 -3206 1972 -23200 -5784 120 27 -760 -665 -4101 1973 -14300 1863 224 272 -704 -605 -1363 1974 -3500 -2466 435 86 -755 -680 -l776 1975 -43600 10690 460 320 -765 -287 -1862 1976 -73700 -5680 270 -389 -1296 -l354 -5360 1977 -53600 ~26459 289 -512 -1602 -1873 -8021 1978 -59200 -28315 394 -586 —l849 -2828 -11573 1979 -40200 -24525 863 -143 -1923 -2630 -8664 1980 -73800 -24088 1112 538 -2053 -2517 -9924 1981 -78900 -27600 889 -111 -2793 -3744 -18081 1982 -127900 -31677 1019 -108 -3087 -4526 —18965 1983 ~207800 -57510 891 -1223 ~3830 ~6537 -21665 1984 -185300 —107838 -304 -3370 -5204 -9765 -36796 1985 -212300 -148493 -937 -4756 -6208 -13061 -49749 1986 -221200 -169789 -1504 ~7142 -6443 -15727 -58575 1987 -150400 -171230 -2342 -9892 -6507 —18994 -59825 Source: U.S. Department of Commerce, U.S= Foreign Trade Highlights and StatiscalgAbstract of the Untied Stetes. 192 193 Table H.2: The NICs and Japan’ Trade Surplus as Percent of U.S. Trade Deficit, 1967-87a Year SIG KOR HK TAN NIC JAP 1967 NA NA NA NA NA NA 1968 NA NA NA NA NA NA 1969 NA NA NA NA NA NA 1970 NA NA NA NA NA NA 1971 NA NA 38.70% 20.96% 32.49% 218.84% 1972 NA NA 13.14% 11.50% 22.10% 70.90% 1973 NA NA NA NA NA NA 1974 NA NA 30.62% 27.58% 37.06% 72.02% 1975 NA NA NA NA NA NA 1976 NA 6.85% 22.82% 23 84% 48.75% 94 37% 1977 NA 1.94% 6.05% 7 08% 13.98% 30 31% 1978 NA 2.07% 6.53% 9 99% 17.20% 40 87% 1979 NA 0.58% 7.84% 10 72% 15.63% 35 33% 1980 NA NA 8.52% 10 45% 12.12% 41 20% 1981 NA 0.40% 10.12% 13 57% 20.87% 65 51% 1982 NA 0.34% 9.75% 14 29% 21.16% 59 87% 1983 NA 2.13% 6.66% 11 37% 18.60% 37 67% 1984 0.28% 3.13% 4.83% 9 06% 17.29% 34 12% 1985 0.63% 3.20% 4.18% 8 80% 16.81% 33 50% 1986 0.89% 4.21% 3.79% 9.26% 18.15% 34.50% 1987 1.37% 5.78% 3.80% 11.09% 22.04% 34.94% aNA is assigned when the United States has trade surplus or when exporting countries do not have trade surplus with the U.S. BIBLIOGRAPHY Balassa, Bela and John Nilliamson (1987) Adjusting to Success: Balance of Payments Policy insthe East Asian NICs. Nashington, DC: Institute for International Economics. Baldwin, Robert E. (1985) The Political Ecgnomy of U.S. Import Policy. Cambridge, MA: MIT Press. Baran, Paul (1952) "On the Political Economy of Backwardness," Ih_ Manchgster Sshool, January: 66-84. Bauer, Raymond A., 1. Pool, and L. Dexter. (1964) American Business and Publis Pglicy; The Politiss of Foreign Trade. New York: Atherson Press. Bergsten, C. Fred (1988a) “US International Macroeconomic Policy," in Thomas Bayard and Soo-Gil Young, ed. Econgmic Relations between the United States and Korea: Conflict or Cooperation. Nashington, DC: Institute for International Economics. Bergsten, C. Fred (1988b) America in tha Norld Economy: A Strategy for tha 19905. Nashington, DC: Institute for International Economics. Bergsten, C. Fred and Nilliam R. Cline (1987) The United States--Jaoan Economig Problem. Nashington, DC: Institute for International Economics. Bergsten, C. Fred and John Nilliamson (1983) "Exchange Rates and Trade Policy," in Nilliam Cline, ed. Trade Policy in the 19805. Nashington, DC: Institute for International Economics. Bhagwati, Jagdish N. (1982) ed. Import Competition and Response. Chicago: University of Chicago Press. Brander, James A. (1986) "Rationales for Strategic Trade and Industrial Policy," in Paul Krugman, ed. Strategic Trade Policy and New International Esonomics. Cambridge, MA: MIT Press. Brewer, Anthony (1980) Marxist Theorjgs of Imparialism: A Critical Smrvay. London: Routledege & Kegan Paul. Caporaso, James A. (1981) "Industrialization in the Periphery," International Studies Quarterly, 25: 347-384. 194 195 Cardoso, Fernando H. (1973) "Associated-Dependent Development." in Alfred Stephan, ed. Agthoritarian Braajl, New Haven: Yale University Press. Chatfield, C. (1985) The Analysis of Time Series; An Introduction 3rd edition). London: Chapman and Hall. Cline, Nilliam R. (1987) The Future of Norld Trada jm laxtiles and Appamal. Nashington, DC: Institute for International Economics. Cline, Nilliam R. (1984) Exoorts of Manpfactprgd Frgm Devaloping Cpuntrigs. Nashington, DC: Brookings Institution. Cline, Nilliam R. (1983) International Dab; and the Stability of the Norld Ecpnomy. Nashington, DC: Institute for International Economics. Cline, Nilliam R. (1982) Reciprocity: A New Approach to Norld Tramp Policy2. Nashington, DC: Institute for International Economics. Cooper, Richard (1986) Economic Policy in an Interdependent Norld: Essays in Nprld Esonomics. Cambridge, MA: MIT Press. Cooper, Richard (1968) The Economfics of Interdgpengenca. New York: McGraw Hill. Cumings, Bruce (1984) "The Origins and Development of Northeast Asian Political Economy: Industrial Sectors, Product Cycles, and Political Consequences," Intarnational Organization, 38: 1-40. Cgomg Commission Report (1988) by Cuomo Commission on Trade and Competitiveness, New York: Touchstone Book. Destler, I.M. and John Odell (1987) AnLj-Progegtion: Changing Forces in United States Trade Politics. Nashington, DC: Institute for International Economics. Destler, I.M. (1986) America rade ' : st nd S r 55. Nashington, DC: Institute for International Economics. Dickson, Bruce and Harding, Harry (1987) Egomomjs Ralations in the Asian Pasifjg Region. Nashington, DC: Brookings Institution. Dornbusch, Rudiger and Jefferey A. Frankel (1987) "Macroeconomics and Protection,” in Robert H. Stern, ed. U.S. Tradg Eoljgjes in a Changing Norld Egonomy. Cambridge, MA: MIT Press. Dos Santos, Theotonio (1985) "Historical Perspectives on Political Economy," in R. Art and R. Jervis, ed. International Politics. Boston: Little Brown and Company. Carr, E. H. (1939) 1mg Twenty Yeams’ Crisis, 1919-1939. London: St. Martin’s Press. 196 Chase-Dunn, Christopher K. (1981) ”Interstate System and Capitalist Norld Economy: One Logic or Two?" Internatjnnal Studies Quarterly, 25: 19-42. Engle, Robert F. (1974) "Specification of the Disturbances for Efficient Estimation," Econometrita, 41:225—38. Evans, Peter (1979) Dependant Qevalopment: The Alliance of Multinatjnnal, Stats and Local Capital in Brazil. Princeton: Princeton University Press. Feldstein, Martin (1988) ed. Internatjnnal Egnnnmig Conperation. Chicago: University of Chicago Press. Feldstein, Martin (1980) ed. The American Esnnomv in Transition. Chicago: University of Chicago Press. Firebaugh, Glenn and Bradley P. Bullock (1986) "Level of Processing of Exports: Estimates for Developing Nations," International Studies Quarterly, 30: 333-350. Finlay, Ronald (1981) ”Factor Proportions and Comparative Advantage in the Long Run,” in Jagdish N. Bhagwati, ed. Intgrnatjonal Trade: Salgtt Rgadings, Cambridege, MA: MIT Press. Frank, Andre Gunder (1966) ”The Development of Underdevelopment,“ Monthly Raview, 18: 17-31. Frey, Bruno S. (1984) International Eoljtital Esnnomics. New York: Basil Blackwell. Gilpin, Robert (1985) "The Nature of Political Economy." In Rober Art and Robert Jervis, ed. International Poljtigs. Boston: Little, Brown and Company. Gilpin, Robert (1975b) "Three Models of the Future," in C. Fred Bergsten and Lawrence Krause, ed. Norld Politics and International Egonnmits. Nashington, D.C.: Brookings Institution. Gilpin, Robert (1975a) US Pouer and the Multinational Corporation: The P litical conomv of Foreign Direct Investment. New York: Basic Books. Glitman, Russell (1986) "Korean Firms Chalk Up Surprising Sales Success In the U.S. Micro Market," EC Neek, 9 September: 181-83. Goldstein, Judith (1986) "The Political Economy of Trade," ri n Pnlitita! Stience Raviaw, 80: 161-184 Graham, Norman A. (1985) Thg Impact of Enrgign Industrial Practices on tng U.S. Computar Industry. Glastonbury, CT: The Futures Group. 197 Haggard, Stephan and Chung In Moon (1983) ”The South Korean State in the International Economy: Liberal, Dependent, or Mercantile?” in John Ruggie, ed., The Antinomies of Interdependence. New York: Columbia University Press. Haggard, Stephan (1986) "The Newly Industrializing Countries in the International System," orld Eolities, 38: 343-70. Hanushek, Eric A. and John E. Jackson (1977) Statistical Methods for Sneial Seientists. New York: Academic Press. Harris, Niegel (1986) [he End at the Third Nnrld: Newly Industrializing Qountries and the Decline of an Ideology. New York: Penguin Books. Hirschman, Albert (1980) ti n l ow n e Str c re of Foreign lrade. Berkeley: University of California Press. Huntington, Samuel P. (1973) "Transnational Organization in Norld Politics, Norld Politics, 25: 333-368. Jackman, Robert N. (1982) ”Dependence on Foreign Investment and Economic Growth in the Third Norld," Norld Polities, 34: 175-96. Johnston, J. (1984) Econometric Methods (3rd edition). New York: McGraw-Hill. Jones, Ronald N. (1981) ”The Structure of Simple General Equilibrium Model,” in Jagdish N. Bhagwati, ed. International Trade; Seleet Readings, Cambridege, MA: MIT Press. Judge, George, Griffiths, N.E., Hill, R. Carter, H. Lutkepohl, and T.C.Lee (1985) The Thedry and Practise of Eeonometrics (2nd edition). New York: John Niley and Sons. Kennedy, Peter (1985) A Guide tn Econometrics (2nd edition). Cambridege, MA: MIT Press. Keohane, Rober 0. (1984) After Hegemony: Cooperation and Discord in the Norld Political Economy. Princeton: Princeton University Pess. Keohane, Robert O. and Joseph S. Nye (1989) w r d ter ndence: Norld Politics in Transition (2nd edition). Boston: Little, Brown and Company. Kindleberger, Charles P. and Peter H. Lindert (1978) Internatidnal Ecnnomies (6th edition). Homewood, IL: Richard Irwin, Inc. Kmenta, Jan (1986) Elements of Eeonnmetries (2nd edition). New York: Macmillan Publishing. Knorr, Klaus, (1975) The Power of Nations: The Political Eeonomy of International Relations. New York: Basic Books. 198 Krasner, Stephen D. (1976) "State Power and the Structure of International Trade," Norld Polities, 28: 317-348. Krasner, Stephen D. (1985) Structural Cdnflict: The Third Nprld Against filopal Liberalism. Berkeley: University of Calif. Press. Krasner, Stephen D. (1978) Defending the National Interest: Raw Materials Investment and U.S. Foreign Pelicy. Princeton: Princeton University Press. Krause, Lawrence B. (1986) ”The Developing Countries and American Interests," in James Morley, ed. The Pacific Basin: New Challenges for the United States. New York: Academy of Political Science. Kurth, James R. (1979) "The Political Consequences of the Product Cycle," Internatinnal Organization, 33: 1-34. Krugman, Paul R. (1986) "Introduction: New Thinking about Trade Policy," in Paul Krugman, ed. Strategic Trade Policy and Neu Internatinnal Economics. Cambridge, MA: MIT Press. Leamer, Edward E. (1984) Sources of International Comparative Advantage: Theory and Evidence. Cambridge, MA: MIT Press. Lehman, Jean-Pierre (1985) "Dictatorship and Development in Pacific Asia: Nider Implication," International Affairs, 61: 591-606. Linder, Staffan B. (1986) The Eaeifie Century; Eeenomie and Enlitical Qonseguenees of Asian Paejfic Qynamism. Stanford: Stanford University Press. Lundestad, Geir (1986) East, Nest, Nurth, Snuth; Major Qeyeldpments in International Politics 1945-1986. Oslo: Norwigian University Press. Morse, Edward L. (1976) "Interdependence in Norld Affairs." in James Rosneau, et al., Norld Politics. New York: Free Press. Nam, Chong-Hyun (1987) "Export-Promoting Subsidies, Countervailing Threats, and the General Agreement on Tariffs and Trade, Th d Bank Econdmie Review. 1:727-43. Nivola, Pietro S. (1988) "Process Protectionism: Blame Avoidence Politics in U.S. Trade Policy,” paper presented at the American Political Science Association, Nashington, DC. Nivola, Pietro S. (1986) "The New Protectionism: U.S. Trade Policy in Historical Perspective," Pnlitical Scince Quarterly, 101: 577-600. Odell, John S. (1985) "The Outcomes of International Trade Conflicts: The US and South Korea, 1960-1981," International Studies Quarterly, 29: 263-286. Odell, John S. (1980) "Latin American Trade Negotiation with the United States," International Organizatinns, 34: 207-28. 199 Ohlin, Bertil (1967) Interregional and International Trade (revised edition). Cambridge, MA: Haravard University Press. Okita, Saburo (1986) ”Pacific Development and Its Implications for the Norld Economy," in James Morley, ed. The Eaeifie Basin; New Challenges fer the United States. New York: Academy of Political Science. Paik, Non K. (1987) "Dependent Development and International Trading System: A Poststructural Analysis,” Ihe Korean Journal of International Studies, 19: 517-36. Patrick, Hugh (1986) "The Burgeoning American Economic Stake in the Pacific Basin," in James Morley, ed. The Pacifie Basin; New Challenges fer the United States. New York: Academy of Political Science. Pirages, Dennis (1989) Global Technopnlitiss; The Internatinnal Politics of Technology and Resources. Pacific Grove, CA: Brooks/Cole Publishing. Rhee, Hak-Yong (1984) "Protection Structures of the Developing Countries in Pacific Asia," sia-Yungu, 27: 419-40. Schattschneider, E. E. (1935) Politics. Pressures and Tariff. New York: Prentice Hall. Schott, Jeffrey J. (1988) "US Trade Policy: Implications for US-Korea Relations," in Thomas Bayard and Soo-Gil Young, ed. Eennomic Relations etw e Uni ed S ates nd K e : Confli or Coo ration. Nashington, DC: Institute for International Economics. Smith, Adam (1976) An Inquiry into the Nature and Causes of The Health at Nations. Chicago: University of Chicago Press. Smith, Tony (1981) The Patterns of Imperialism. Cambridge, London: Cambridge University Press. Sneider, Richard (1986) "United States Security Interest," in James Morley, ed. The Paeific Basin: New Challenges for the United States. New York: Academy of Political Science. Spero, Joan Edelman (1985) The Politics of Internatinnal Eeonomic Relations (3rd edition). New York: St. Martin’s Press. Strange, Susan (1985) ”Protectionsim and Norld Politics," International Qrganization, 39: 233-259. Strange, Susan (1979) ”The management of Surplus Capacity: Or How Does Theroy Stand Up to Protectionism 19705 Style?" International Urganizatign, 33: 303-333. Stokes, Bruce (1988) "A Thorny Agenda Facing Eximbank," National Qdurnal, 45: 2800-01. 200 Sylvan, David J. (1981) "The Newest Mercantilism," Internatidnal Organization, 35: 375-393. United Nations (1967-85) Industri l t t's i Ye rboo . Vol.1: General Industrial Statistics. New York: United Nations. United Nations (1970-85) Yearhook df National Accnunt Statistics. New York: United Nations. U.S. Department of Commerce. (1988) U.S. Merchandise lrade Position. Nahsington, D.C.: U.S. Government Printing Office. U.S. Department of Commerce. (1967-1985) U,S, General Imports: Norld Area by Commndity Groupings (FT 155). Nahsington, D.C.: U.S. Government Printing Office. U.S. Department of Commerce. (1987) U.S. Foreign Trade Highlights. Nahsington, D.C.: U.S. Government Printing Office. U.S. Bureau of Census (1967-1987) Statistical Abstract of the United States. Nashington, D.C.: U.S. Government Printing Office. U.S. International Trade Commission (1967-1988). Annual Report. Nashington, D.C.: International Trade Commission. Vernon, Raymond and Debora Spar (1989) Beyond Globalism: Remaking American Foreign Economic Policy. New York: Free Press. Vernon, Raymon (1982) "International Trade Policy in the 19805," International Studies Quarterly, 26: 483-510. Vernon, Raymon (1971) Sovereignty At Bay: The Multinational Spread of US Enterprises. New York: Basic Books. Viner, Jacob (1948) "Power Versus Plenty as Objectives of Foreign Policy in the Seventeenth and Eighteenth Centuries," Norld Politics 1: 1-29. Nalleri, R. Dan (1978) "The Political Economy Literature of North-South Relations," nternational Studies Quarterly, 22: 587-624. Nallerstein, Immanuel (1979) lhe Capitalist Nnrld Economy. London: Cambridge University Press. Nallerstein, Immanuel (1974) The Modern Ndrld System. New York: Academic Press. Naltz, Kenneth (1954) Man, the State, and Nar; A Theoretical Analysis. New York: Columbia University Press. Norld Bank (1988) EQEQQ; Managing the Industrial lransitidn. Nashington, DC: The Norld Bank. 201 Yoffi, David B. (1983) Power and Proteetionism. Princeton: Princeton University Press. Young, Soo-Gil (1988) "Korean Trade Policy: Implications for Korean-US Cooperation," in Thomas Bayard and Soo-Gil Young, ed. Eeondmic Relations hetween the United States and Korea: Conflict or Quoperation. Nashington, DC: Institute for International Economics. HICHIGRN STRTE UN 1 llllmlllllllllllllllll IV. LIBRARIES V IMIWHIWINWHI 312930 7841384