PRICE ELASTICITIES AND THE EFFECTS OF TRADE LIBERALIZATION FOR THE UNITED STATES, THE EUROPEAN ECONOMIC COMMUNITY AND JAPAN Dissertation for the Degree of Ph. D. ' MICHIGAN STATE UNIVERSITY JOE ALLAN STONE 1977 This is to certify that the thesis entitled PRICE ELASTICITIES AND THE EFFECTS OF TRADE LIBERALIZATION FOR THE UNITED STATES, THE EUROPEAN ECONOMIC COMMUNITY AND JAPAN presented by Joe Allan Stone has been accepted towards fulfillment of the requirements for Ph. D.Ajmgyeh1Economics Major professor Date W7 0-7639 GICHCIogw PRIC Th Improved e second, to the POtent‘ The COmmod. cateSlom'es Conflunity, Est 1mIIY‘OVed In and Specifi eIaStICltieg Same sAmple the EIaStIci One Unit the derived 0” t estimameS 1's ABSTRACT PRICE ELASTICITIES AND THE EFFECTS OF TRADE LIBERALIZATION FOR THE UNITED STATES, THE EUROPEAN ECONOMIC COMMUNITY AND JAPAN By Joe Allan Stone The purpose of this study is two-fold--first, to develop improved estimates of price elasticities of import demand; and second, to use these estimates to predict, at "industry" levels, the potential effects of alternative trade liberalization schemes. 1 The commodity categories included in the study are manufacturing categories for the United States, the expanded European Economic Community, and Japan. Estimates of price elasticities of import demand are improved in four areas: comparability, applicability, consistency, and specification. First, comparability is improved because the elasticities are estimated for comparable categories using the same sample period and a generally similar methodology. Second, the elasticities presented are more applicable to the E.E.C. as one unit than are previous estimates, since the estimates are derived on that basis. Third, the statistical consistency of the estimates is improved by attempting to correct for errors in unit value try tional us specifics ll presented percent r five per c original t on textile alternativ ties relat elasticiti, Joe Allan Stone value trade data and for "simultaneity" bias. Fourth, the tradi- tional use of relative price is reconsidered, and a more flexible specification is generally used. The estimates of the effects of trade liberalization are presented for three modalities of liberalization: first, a sixty per cent reduction in all tariffs and elimination of tariffs of five per cent orless; second, percentage reductions equal to the original height of each tariff; and third, the elimination of quotas on textiles and steel. Rising supply prices are considered as an alternative to constant supply prices, and the traditional identi- ties relating a trade elasticity to the domestic demand and supply elasticities are also reconsidered. PRI PRICE ELASTICITIES AND THE EFFECTS OF TRADE LIBERALIZATION FOR THE UNITED STATES. THE EUROPEAN ECONOMIC COMMUNITY AND JAPAN By Joe Allan Stone A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Economics I977 ‘9 Copyright I. JOE ALLAN 5 I977 Copyright by JOE ALLAN STONE 1977 Hordecha fessors L recogniti more, Pro State Uni overcome r acknowledg In the De; In the com deserve thr Provided dL ACKNOWLEDGMENTS The chairman of my dissertation committee, Professor Mordechai Kreinin, and the other members of the committee, Pro- fessors Lawrence Officer, Anthony K00, and Robert Rasche deserve recognition for their invaluable suggestions and advice. Further- more, Professor Daniel Hamermesh and other friends at Michigan State University provided much of the encouragement required to overcome many of the obstacles encountered along the way. A special acknowledgment is also due the Bureau of International Labor Affairs in the Department of Labor for their partial financial assistance in the completion of this study. Finally, the members of my family deserve the most sincere acknowledgment for the special help each provided during the long course of this research. My appreciation goes first and foremost to my wife, Crystal. ii TABLE OF CONTENTS ACKNOWLEDGMENTS . LIST OF TABLES LIST OF FIGURES . CHAPTER I: INTRODUCTION . l. 1 Background and Purpose . . 1.2 Previous Studies and Methodology . l. 3 The Format . CHAPTER 2: ANALYTICAL FRAMEWORK 2.l Import Demand and Export Supply Importance and Derivation Zero Homogeneity in Prices Functional Form . . . Tariffs and Quotas . Effects of Uncertainty Institutional Parameters . 2.2 Static Effects of Trade Liberalization : Nature of Analysis . . Prices and Trade Volumes . . Domestic Consumption and Production Domestic Employment Welfare Gains CHAPTER 3: METHODOLOGY AND DATA 3.1 Treatment of Variables ' Commodity Categories . Price and Quantity Variables Alternative Price Variables . Income-Activity Variables Tariffs . . Trade Volumes . . Labor- -0utput Coefficients 3.2 Choice of Estimator . Import Demand Equations Page ii viii d Export Supply Equations 3.3 Use of Relative Price CHAPTER 4: ESTIMATES OF PRICE ELASTICITIES 4.l Empirical Estimates 4.2 Arbitrary Estimates 4. 3 Autocorrelation . 4. 4 Evaluating the Results CHAPTER 5: STATIC EFFECTS OF TRADE LIBERALIZATION 5.1 Introduction . . 5.2 Tariffs and Tariff Changes . . Computation of Tariff Changes Scheme A Versus Scheme B . 5.3 Changes in Trade Volumes Percentage Changes . Absolute Changes Aggregate Changes 5.4 Welfare Changes . 5.5 Employment Effects in Othe U. S. . . 5.6 Liberalization of Textile and Steel Quotas CHAPTER 6: SUMMARY AND CONCLUSIONS BIBLIOGRAPHY . iv Page Table l-l. 4-T. 4-2. 4-4. 4-5. 4-6. 4-8. 4-9. 4-10. 4-11. 4-12. 4-13. 4-14. 4-15. 4-16. 4-17. 4-18. 4-19. LIST OF TABLES Annotated List of Selected Trade Studies . Commodity Categories Leather Manufactures Rubber Manufactures . Mood Manufactures Paper Manufactures Textile Semi-Manufactures . Textile Articles . Clothing Mineral Manufactures Glass Manufactures Iron and Steel, Unworked Iron and Steel, Semi-Manufactures Aluminum Manufactures Other Metals Metal Manufactures Petroleum Manufactures . Organic Chemicals Inorganic Chemicals . DTC Materials . Plastic Manufactures Table 4-20. 4-21. 4-22. 4-23. 4-24. 4-25. 4-26. 4-27. 4-28. 4-29. 4-30. 4-31. 4-32. 4-33. 4-34. 4-35. 4-36. 4-37. 4-38. 5-1. 5-3. 5-4. 5-5. Oils, Perfumes Other Chemical Products Power Machinery Agricultural Machinery . Office Machinery . Metalworking Machinery . Textile Machinery Other Machinery Electrical Machinery Telecommunications Apparatus . Motor Vehicles Miscellaneous Transportation Equipment Precision Instruments Footwear, Travel Goods and Handbags Photographic Manufactures . Furniture Sound Manufactures Toys . Comparison of Estimated Price Elasticities Tariffs and Tariff Changes Import- or Export-Bias of Scheme 8 Relative to Scheme A . . . . Percentage Changes in Trade Volumes Absolute Changes in Trade Volumes Summary of the Absolute Changes in Trade Volumes vi Page 76 77 77 78 78 79 79 80 80 BT 8l 82 82 83 83 84 85 85 9T 95 109 ITO 122 T34 Table 5-6. 5-7. 5-8. Michange Estimates of Scheme A Welfare Effects Summary of Mid-Range Estimates of Scheme A Welfare Effects . . . . Mid-Range Estimates of the Employment Effects in the United States . Summary of Mid- -Range Estimates of the Employment Effects in the United States . Changes in Trade Due to the Elimination of Textile and Steel Quotas . . vii Page l39 I50 152 I56 I58 LIST OF FIGURES Figure Page 2-l. Tariff-Induced Welfare Changes . . . . . . . . 4l viii CHAPTER I INTRODUCTION l.l Background and Purpose Meeting in Tokyo on l4 September l973, the Ministers of the General Agreement on Tariffs and Trade (GATT) officially opened a new round of negotiations aimed at dismantling restrictions to international trade. The "Tokyo Round" begins with a legacy of success established by a series of trade negotiations held since World War II. These post-war rounds of international trade negoti- ations have contributed to three decades of economic growth among industrialized countries. Unlike its predecessors, however, the Tokyo Round is staged against a backdrop of world-wide recession, chronic inflation, and drastically higher petroleum prices. The purpose of this study is two-fold--first, to develop improved estimates of the required price elasticities of import demand and, second, to use these estimates in a model designed to predict, at "industry" levels, the potential effects of alternative trade liberalization schemes. The focus is on trade in manufactures of the United States, the recently expanded European Economic Com- munity, and Japan. 1.2 Previous Studies and Methodology, There have been many studies of trade flows and trade liberalization in the post-war period--too many to describe in detail here. Leamer and Stern (63) provide an extensive survey of this literature and Magee (68) provides a recent discussion of research issues in this field. Table l-l at the end of this chapter provides an annotated list of selected studies beginning with the benchmark Orcutt (8l) article, tabulated by author, year, scope, disaggregation level, sample period, functional form, and lag tech- nique. Specific studies will be singled out when pertinent to the development of a particular topic. The distinguishing characteristics of this study are best discussed in relation to the dual purposes established above. In regard to the first objective, estimates of the required price elasticities of import demand are improved in four areas: compara- bility, applicability, consistency, and specification. First, comparability is improved because the elasticities for the U.S., the E.E.C., and Japan are directly estimated for comparable com- modity categories based on the same sample period and on a generally similar methodology. Previous work has largely been on a piecemeal basis--estimating price elasticities for one country or for several countries but for noncomparable categories. The amalgamation of such disparate studies is a potential source of significant bias in policy studies such as this where one is concerned with one country's estimates relative to another's. Second, the elasticity estimates presented here are more applicable to the E.E.C. as an entire unit. Previous research treating the E.E.C. as one unit has been at the aggregate level or has concentrated on only a few commodities. The remaining studies related to the E.E.C. have concentrated on individual member coun- tries. Unfortunately, the price elasticities for one or more component countries may not provide adequate information about the price elasticities for the E.E.C. as one unit. The sample data for each category in this study are E.E.C. trade, net of intra-E.E.C. trade. Where both possible and essential, data for the three new E.E.C. entrants (the United Kingdom, Ireland, and Denmark) were included in the sample to assure the applicability of the results to the expanded E.E.C. Third, the statistical consistency of the estimates is improved by giving due consideration to the potential bias resulting from the use of unit value trade data and from the simultaneity problem. Trade data are published only after some aggregation of commodities has taken place. Unit value and quantity statistics taken from this data are correct only if the composition of the category remains unchanged or if the changes within the category cancel one another. Otherwise, the statistics will be in error. In this case the ordinary least squares estimator of the price I coefficient is biased toward minus one. This potential bias and the procedures used to minimize it are discussed in more detail in 1See Shinkai (97, p. 272). Chapter 3. The simultaneity problem refers to the fact that quantity and price are usually determined simultaneously by the interaction of the forces of demand and supply. The ordinary least squares estimator in this case is biased toward the price coefficient of the corresponding demand or supply function. Like the unit value problem, this issue is discussed in more detail in Chapter 3. Fourth, the traditional speCification of trade functions Using a relative price variable is used only when this approxi- mation is required to obtain reliable estimates. Basic theory suggests that if all prices change proportionately and all real explanatory variables remain unchanged, then the quantity demanded or supplied remains unchanged. This absence of "money illusion" is the justification for using a relative price variable which imposes the assumption that the own price of the traded commodity and the price of the domestic substitute have equal but opposite effects on trade in the commodity. There are a variety of reasons for believing that in practice this assumption is a poor approxi— mation. Chapter 2 discusses these reasons in detail, and Chapter 3 discusses the restrictions placed on the use of relative price in this study. The second major objective is to build the analytical frame- work needed to predict the effects of trade liberalization. Fortun- ately, most of this framework has been constructed and used previously by others.2 Three distinguishing characteristics of this particular 2For both theoretical and applied examples, see Johnson (45) endi(46), Balassa and Kreinin (5), Leamer and Stern (63), and Magee 69 . study are worth mentioning, however. First, three methods or modali- ties of trade liberalization are considered. The first two involve explicit tariffs and the third concerns non-tariff restrictions. There are two basic approaches to tariff reductions. One is across- the-board reductions in all tariffs, and the other is reductions in tariffs which are proportional to the original height of the tariff. The objective of the latter approach is the harmonization of tariffs across both countries and commodities. For the purposes of this study these two approaches are formulated as follows: Scheme A: The maximum authority under the Trade Reform Act of 1974--a sixty per cent reduction in all tariffs and elimination of tariffs of five per cent or less. Scheme 8: Percentage reductions equal to the original height of the tariff. There is an inherent scale problem in comparing across-the- board schemes with harmonization schemes. One objective, therefore, is to compare the relative effects of the two approaches given any arbitrary scale of average tariff reductions. The computed effects for Scheme 8 may be adjusted to any scale of trade liberalization by a simple multiplicative factor.3 Elimination of non-tariff barriers may be done separately or in conjunction with reductions in explicit tariffs. There is an almost infinite list of non-tariff barriers to trade. Work has been done recently on the American Selling Price system of evaluation 3This is true of all the effects except deadweight loss (gain) calculations which depend upon the square of the scale. See Chapter 2, Section 2.2. in chemicals by Jadlow (42), on the barriers to trade in iron and steel by MacPhee (67), and on a variety of quotas by Magee (69). Consequently, only the major quotas on steel and textiles will be considered here. The primary effects of these quotas will be sum- marized in light of the relevant price elasticities estimated in this study. A second distinguishing characteristic of this study is that the possibility of rising supply prices is considered. Most studies examining the effects of eliminating or imposing trade barriers have claimed that the full burden of the barrier is borne on the import side by assuming that export supply is perfectly elastic. This study follows the precedent of Balassa and Kreinin (5) in considering the possibility of rising supply prices for large traders. Chapter 3 examines the estimation of price elasticities of export supply in detail. A third distinguishing characteristic is a critique of the traditional use of elasticity identities which relate import (export) price elasticities to the domestic demand and supply elasticities. These identities have been used extensively for a variety of purposes. Unfortunately, the statistics available for use in these formulas are not the same as those dictated by the theory underlying the formulas. Chapter 2 discusses this issue in conjunction with the problem of allocating the domestic effects between producers and consumers. l.3 The Format Chapter 2 deals with the analytical framework required to predict the effects of trade liberalization on prices, trade volumes, employment, and welfare. Chapter 3 discusses the treat- ment of variables, the choice of estimators, and the use of relative price. Chapter 4 presents the econometric results, an explanation of the estimation process, and a summary of the results. Chapter 5 details the effects of the various liberalization schemes on tariffs, trade, and welfare and also includes a special section which estimates the domestic employment effects in the United States. Chapter 6 includes a summary of results and recommendations for future research. .mmvuaum :uon cw Empnoga a mcmrm =mcog3= .mp—ammg Loon saw: creme .Lmuusn ncm .mumpq -meum .ummzz Low mcowumaam apnaam use ucaeoo op nowpaaa osz .mszmmL Loon saw: some axons? zgaczou xmm mcu op cmTFQam m4o .mpm>m_ wwwgmh .m.: amo— mnp cw mmcmgu xom a mo mcowuomnoga commm_e -Eou mewgmh mem— cog: ummmn vcmemc axons? mo mmwueuwummpm =$wwgmu= owow_asa co mooaaaomm .mmwgommpmo magmp Low mam—noga zxmucw= ucm .mmmp woven .copum>gwmao we mgoggm .cowums uwumw comamzcm mpmcwm x5 ngmzczou ummmwn ago: mmwuwuwumopm woven mo mmumswumm apgum was“ umpmmmmzm A.¥.=v coupsn new .Ang ooapa mepm .Amcwucmm lk< aoNoz an? rpmgpm=wp zuwcoeeou mmmpuommp use an m_muoe cmmprommp chm>mm .c_mwgo mmm_-omm_ we mcowmmg N P-“ ecu tease. epm_-ommp so mowoaooaeoo mo_to poo_eou comma_== -mo_ aom_-ommp N, o» N eoea -czoo 9 _mm_ a cameo: -- -- -- moaowuoesoo o“. m: _mm_ moat>oe -- -- -- -- r- ommF opzoto mxgmemm mmm4 cowuuczm nowgmm cowpmmmgmmmmwa wqoom me> Lozua< .mowezom ovate noooopom to down aooooo:=<--.p-_ mom<~ .mema mewumeumgew gesrumee use em>wgme mmwppu_pmm_m eeeEme mmgeuuemeeee washes?» .eooc axons? we mmumemumm u- Lemcwp mmm_-m~mp .mpewgmpes 3mm m: mmmp gmmmwmz .eceeee ugeesw we Nmmprmemp maeeemmmeN mmpeewume emew>eee Lem» wee Lemewp mmm—ummmp m_ewgmpms zem m: mmm_ a ame>e4 .meaeze ece mewweeu Ce eewmcmemam Peempep twee .muepeeeo e on one wagons? mep cw monomeucw we Aeewc_ee =pemexm: :ee: ewmeev mezegm mooaswomo noew>ota -- -- -- someoeeoo eNN m: mmmp oasewa .Aaemp .Lmewem an emmsv moo renew zawueeae pgeexm :wmwee meeeeepmmz eee meFe> are: 3m: ee meewmme m an ee> w .me saw: meeFm emuzeeeuew meeegm mmooese rewmm_;om coPoPooaeoo Po=o_mom -- tamer, AmmP-mmmp -co-omaom a m: NmmF .to_e< .mmwuwuwpmepm afieezm pgeexm ece tensoe axons? e>wgme e» eem: mmvupupumepm apeeem FeegmucP zgegaweee ecu Asmaaae ecu werewowpmepm eeeewe Leeewp .uemgz .Peezv a? peeemucp ewueEwpmm .. -me_ wmmprommp mmwuweeeseu m rpeeum=< Nmmp emcee: mxeeemm woe; eewuuceu eewgma cewuemmgmmemwo meeum me> Lezpz< .mmmm_o=oe --.F-_ memmF mesmermee -mwm pee .eewue_mggeo reuse 3e; .mAO on Low; renew xpemmpu me: 42“; .meeuew eee .mmmez. .xpeeem .eeesme mew teem—axe mugeexm .N.z Lemewp coupes .eEeF nee—emN we Fmees :ewuezem um < :1 -mep mmm—rwmmF .muuaeeee xgwmo 3oz mmmp Eeeummgem .mugeexm Le mucous? emsbwm Lem mmweuezeu 3mm m xFee Lem eeaew muemwowwem -eu more; accommwemwm »_Faowomwoapm .eopmxm me agape: m>_mgzewe >e mmwepczeu Lemme emvwwumzn mAO .eweeemw as» Low mmweem mcewpweeeu x_ee=m new: -mueu m>wm ep mseucw eee .mugeexe oz“ gov: mggee .mpgeesw mew:_sgmume new new wagon meme» meewpezem we smumzm < Pup Leeewp mmmprmemP -xm muemmemo< rezeu mm mmmp xepee .eee:_uew mpumwmm =emwpewapaez munch mpeeaxm .mzeemmexm meowge Amweeueem gee e sum: appeaew>weew races eee .eeem .mueeeew empeswpmm Amugeexm ece .mmegmpeE Lem FF wagons? speev mcewu Le: zeev mewgem .Fpeem>e wcempmweez renew Peeauozgpm mm 1- LemeWF remueH rmumo mmgep m mezeem m mmm— a Lemmyez mxgesmm mum; :ewpueeu eewgmm cewpemmemmmmmo meeom gem> seguz< FIE! | .eoacaocoe--.~-_ memeeme mmmcesu eeeeep xeaeeeh age eo eomae em-memev muemeem eeueaeumu .mmepememeeee eee moeuPUwpmepm meme» op meeeeaeeaa emme op emm. Eoee nonmeenaa mepeeuee eee mxeee me me xmeee use xm>eem < .meeeem sue: meeep emcee Eeee meeem mzu mumseumm ea xeepm pee: muemmPF upmuee umeseeeum esp me wees meeeuuwheee zeeceeuepeumee= emm: .eeesme ueeexw eee emeeseumm wee meeueoeu -mepm meeuee ece mueee .ee>e3e: .mmepeuee -meFm ueeexm m>eeme e» emm: mmeueuepmeFm oepmmeee zeeeueee< enmeep umep =eou=emecee= ego =eoo=eoe= ooze eoee>ee .Ampewempee 3ee ece eeee emme-mma_ amoev ooamoemm< mezeem .. zueeeeaeu m_ mueeexw wee :eee use mmmprmpmp epeeeeez .ueeee mmmprommF -mzm: .pemez m: mmmp mmamex -- mmme ozone x: mmmp cemeeee eveceu nmmp coupe: mxeesmm l‘ Ill Fl mmee seepeeee eeeeme eeeuemmemmemea meeem eem> eeep=< .eoaceoaoe--.e-e memee I2 .appeeee>weee ewuee repmw meeeueeem .pweee ueumwaee ewueemwuee no me ween we ewewseumm .eeeeeu =ueeepeweom= ee eeepee we» meeeepw>we cw ewm: mppzmwx .muuweew weeep we msewu ace -peeeeeeuee ece mw>e=o =eewpwepe> peeem news: we mw>eeo eeeewe meema mwePF emeeeesm nee: meeFe meeeeep +e muuweew weeepwz we“ ee Hewsueweu poweee < .weeeu we meewu accumeew pm Peeemeez.mfl.m eweemewe m:_em wewepwz .mw:e_ eewep compuzewe zpmee= use =wea= meeooee ea meeepezewe eeeeep eesewu mmuommp wee ee muoweew ewuweewmm meeem -we m ea mueee -xw meeewwewmew eee .eeee .wuem owemm< .mueeesw meweeuweeecee eee .mpeeewp emmprmemp nee zee .eeee mweeuxwu .mweeuueeeees -wswm .mpweeeee mmm—rommp reew Pemepmeeee weeeee> a .eeeszN 1m: .Fpem x: Pomp .cww—x -- come cemceoc ewe heeEeww comp upwempwswz mxemewm eeewweee eeeewe eeeuemwemmmmea weewm eew> eecu=< amwmmflocoe--.e-_ wwmwu rum .wuwee w>wpwwwe we eewuueew e we eewe -we peeesw ewumewumm .eweepeew wme upw> “we: we wmewwwo .mwwewwew> wpeeeewm we eweeaw use woeea new: mewuwe ueeesw wees -wumw e» ewm: uwzeeee we meme eewpuwmummeeu .ewumwmmam :ewpwmwemmemwe new wane apewpewae we wma .ewmeweee: ecw upeueo we Emewpweu apeew we agessem eee mwweeum weeep we xw>e=m w>wpeewe>w e< .ewewewmeew xuwuwpmewwrmmeeo e» meeaee ewZew new ewee: .eaeem Feeueeu =ewu=eweeec= w eeew ew>wewe eeeem zewoee -we: we» ee mwmewee wwwemu we muuwwwm wueuwum team we wowee eew era Apewuewae Fee: eaocee owm_-amme eoeaeaa 3am mmL3HUM$=c ME eawcee mmme-eame wo maaoeo N mweepwwwzeee we mwweemwpwu =ewe=eweeee= --. mmm_-mmme use geooaeoe= xwme—ee w: mom. a cemsmewe m: Nomp wmzmex .- meow mweem m: —omp :wcwwex mxewswm I'.UUIT‘T mam; cewpwcee eeweww cewpwmwemmwmwo weeum emw» eoeo=< .vwacwucouur.Fu— m4me mewuwwem ucwwwwwemwm ecaee .eegmeaz E. w xuwwwp: pcwpmeee eee wemeu we mEewe uewum -eeu we eweemwws uwwwww weewwwz .ewewewmceu emwe mueeexw ewweweeu ee uwwwwm .mpwzmwe wueev w>wewe op ewm: mwowee eweex eew owumweee :wwzpwe wwewewwwwe .mwwuwwwumeww aweeem new eewswe wwumwsee xeweuweew we :ewpecew e we newswe peeeew new: ewepe we» .xeeam emme m.eo=ewe can: ewmee weeureewmewe imam wueae new wwweeu we mauwwww we» eew mwueewemw we mpwm ezw .ewuwswumw wewz mcewpwwwe eeeewe aweo .mweepowwecws twewm eew wweewew> m we ewm: meueum To ewmwee we mcewmwe xwm we mueeesw :2: 82-32 5.5:: meeeem zu -weeEEee uHm mm mewmmwueee -werwmwum awewuewee xe mwweemwuwo eewcww comprmemw m eee mew>ewm m: comw ewewwm m: womw cewpm gazewz m: Nomw a Ppwm mxeeewm mow; eewuweew eewewe :ewpemwemmemwc weeum emw> eeeu=< .Ieosczcoe--. Te 39: I5 .nwewnwm -eeu =xeewwwwwwem1wwwm= we umeu .mww>ww wwwewu meewew> eew nwpseeeo mumew eewwueneee new eewueszm -eeu .meewueeew eewu -wEeewmewep new xpwwwu: wwwwwexw Eeew awuuwe -wn eewpwwpeee we amen we» weemwwe op peewww e< .ewwwen we» we eewpweww> 1ew>e wpwawemw e» nwm: newemwm .hweeem new newewn we mwwu rwwwumwww wweewuew we mwpwewumw =eewee= Eeew nwpoeewmeew mueeexw new meeeeew we aweeem w newewn we wwwuwwwumwwm .mom—reomp eew wnwe meewpuwneew .meewmwe wweep we“ eew wwnee eewpwzcw mm w we uewe mw mueeeew new mpeeexw wawswumw e» nwm: mwo -1 r- -1 a- momw eemeeen mueee wmwe new new mueee -- mmwe-emme -xo wwwmoewme m: meme weoee eewmwe nweez xe mweeexw new we umwm uwzwwe mueeeew wmwn weeezm .3 reemmwem ewweww Nomwumnm— neweuewE quow m: nomw a meweeozm mxewewm mow; eewpweew neweww eewpwmwemmwmwo weeem eww> eeep=< .woaeewwmu--.e-e wwwew I6 .mwwpwuwpmwww wwe -ewuew Fwwwpewnw mew -Ezmmw we muwzmwe m: Eeew nw>wewn meewmwe eewwe m eew mwuws mewmmwueee uweumwmwpm e3» we mwweum neew twpmm .ewwwewm mewamwe wuu new ewweww awewpewev new mweepwwwe nwmmwwe: new nwmmww nwmmwwez -mew enmwuwmmw news .wuwmwemm< m: womw ewewwex .eewuw—weeeweuew meewmw>wn nwwwnwe newep wewu we mew ume mm .mwweeu wm: .mwwwewuwe wneeu -wwmewm Awewpewne -wuww xawneeeew eew nwm: uuwwww nwww w-» .N+p ewweww momwrmmmw w .wpwmwemm< newwweH womw ewmww .muuwwww =zewneeuwm= wwewmmee ewnwmeeu e» wwnes wwewewm w euew mwuw>ewm new nweweew meewuwezw Flu new awewuewec mnoem we mueee eemewnew: peeeaw nwpwswumm nwmmwwen ewweww womwnmmmw -Ew wuwmwemm< x: wmmw a eepmeeen .xwewwwewm nwseew rewe mseew ewweww -mew new ewweww .mewn -ee emwweew ewwwexw op nwm: eewuwsneee ewweww eeuwm peweeeu eu mewnee Pup new -mew xwewpewea a .wwwm nwwwwwee we ewuwm nwmmwwee ewweww Nomwrnmmw mweeu wewewwz x: comp .ewzwum Eewweepwe new .mepew .ueeem ewweew mmww .ewmam .wwwwew uwewwe wewu peeeew .mewumww awewuewee mewe mewmmwweee reemmwem -weew eew nwemwmwo xoxex ewweww Pomwrmemw weumwmwum m ed m: mom_ a meweeeem mxewswm mmwn eewuweww neweww eewuwmwemmwmwo mqoom E El. wJEWHCOU'F o pip u4m “we: on wen w- newzep xuwuwpmw—w wwwee meo ew wwwm .mueeexw wmwewewn eew newEwn we xnepm eewpwwmrmmeeu .mwwuwo -wpmwww wnweu nwpwewpmw xwmeew>wee Eeew nwuzeeeu eewpwzww>wn uwuwwe new new ewwm wewwwwz .mwwwewu m: ew eewueenwe wwewuwwwe: w we mpuwwww wewwwwz wee ew mpuwwww wnweu we meewu wnewu new ea nwemwmwn wwnez .nwnnw puwwww wnwep we msewu eew: wwwemewz .mH.m nwezmwwa mewwm wewwwwz .nwewnwmeew moon Eeew mpuwwww :xuwenwwez .zwwewep tweew nwueneepew xuww -wpmwww aweezm ueowxw .mpwemwe wee w>wewn ea wwwpwuewnw xuwwwumwww new eewawEeewew eewee mewm: enaum wwuuwwum newew-oze Lowc: m .mmmmmpu -moe mwm_-awme wwmww-weo e eweww wwme eweewem wmwe -- Nwm_-wmme owwmwemme we wwme e>omwm mwenpuwwnewe new mwwwewuws wwwep -menew .mwmmwww mwweueeew mewmmwueee wwwep ewewwex r- wmwe onmw rwerwmwpm m -menew n womw a wmmwwwm mxewswm mom.— cowuucsm flowewm I :otmmmemummwc maeum gwmx kg “2‘ / IN wmmmmmmmrs..-_ wwwee I8 .mweww> awe: we wuwwe ew nwm: mwuwnew wowee m: wwepw< .wnwewu eew e» meeeem neeex xnweewx we» we weeexw FN mpwwwww wwwee wpwswpmw mwmmwwe op nwm: mweeweeuwu we awewaewee ueeeew wmweeueez 1euwenww: new wee euem eeew< ewweww meow-mmmw wme-new m wnwewu meme w ewewwwo we eew .mnepm Eeew nwpque mwmmwwu mueeexw meewwoeew Aweeem mewmmwweee w m eew .awewwewee mwmmwww -we-wmwem mueeesw xewneesew .xwwwee m we mpeee .mwwep 1ew nwpwewumw xeueaew -sw new mueeexw 1e=ew ehe wwwpew .newswn eemw-wwmw .mwwepeeeu ww m— eew we xpwewpmwww wsewew ewweww xwwweeew we mpeeexw new mueeeew wemwz a can: wwowee mwmweQEw -- -moe wwme-emm_ meeoaee m: ewwow wmweoexw meme ewxxweezo: .eewuwewumw ew nwme memw new meo wwwwewuws zwe euem .wwnee mpewsxwe awewpewze .mneem wwpweww we wuewwwe w mw e: wwm xuxex ewweww «new-mmmw .mneem ew53meeu m: mnmw xezeeuwee .peeem ewweew mmww news 1wn ueeexw new ueeeEH .nwpuwse meewueeew eewpwewumwn he aweeem .mwwewwew> pew awewuewze mueeexw .meeeem 1newewn mw nwm: mewuwm xwxex ewweww meme-mmmw wmz-new o_ m: meow eemewem .x: wee mewNwmweesw .wnwep ew =mmwew>wu 1wuweseu= ee mewewwe euewwmwe we xw>ezm 1- 1- 1- 11 11 weep ewewwoez mxewewm. mew; eewueeee newewe eewuwmwemmwmwe weewm eww> eeeu=< r .nweewpceu11._1w m4m weepw 1ewwexw meesw mwezmwws auw>weww new .Awwwewp paeeewzv wowee .mewn -ee nwwwwwe: .mwweeu 1ew>eH .nwuewmwee wew meewpweew mewpmwuweew new wwezuweepm euem P1p eewum - - -- -- oeme w ewewww wwweueeew ewweww mp ea mueeexw -moe meme-meme weee=wowescwe m: weme ewewe eewueeewewm 13N1ewxwz awewpewec mpeeexw a mueee a .weezmm ewaee meme-mmme -ee oomaoemwe wwwew ee meme .mane meewewm mew; eewpeeew newewe eewuwmwemmwmwe weeem eww> eeeue< .nwecwHCOU11._1F m4m xeeuwewwexw mw nwm: mewpwe wwwee new .wewp mwwewezneee .mewpwe weepew>ew ew53meee .mwwe nwmmww .wswu mewuwwz 1ween ew53meee .eewpuezw xuwwwp: .mwweween ewe mmu Eeew nw>wewn ewweww awewpewze 1eneee .mpeee meoewoeae weweow weoeee N.“ ._-w -wo_ meme-meme -ee owwmoemme m: eeme eeomoew eewuwe 1wumwn new ewe 1wee we meewmwe .nwewnwm wwew>wm we 1eew emww meeweww xpwwsme-1mepew wowee1eee ewepe new .uewEewaew ueee euwwweo .m>w>e:m wwwee 1meweu .zeweweu wwzuww eew: nwmwe 1we wwwweuuwww mwewnew wowee .m: wee .mwenuewweewe ee mwmweeew eew: zmmwe ewweww comp-Pomp wwpwe .mwwuwe 1w>wuwuweeew wwwee= 1mew wmmw meeeewweee memo wwmewe mmwmmw ow weewwwe - ewwewe meme .wwwwm w eeee m: eeme w me>we¥ .mneeeem wwuwpweeweu ewwww ee nwww>ew wew mpwem 1we eewpwwweeew flewmwv ew_w< wee ewe» “awesow - - - - - eewe eweaeoe mxewswm mew; eewpueew newewe eewpwmwemmwmwe weewm eww> eeeue< .nwzewuceu11.w1p m4mewmwe wweewuwe 1ewpew new mewew new muewem m: .nwm: He: =wwwemewz= wew mweeow 1ew>ew mmwewmze Eeww -eez .newswn ueeeew new peeexw nwpwewumm .mwweueeee nweeww>wn new mwezw 1uwwaews ee mwmweesm .eewuwewemw ew nwm: mweww> awe: new meo .ueeem xew> ewweew mmww weww .mwwp -wpewnw Seew nw>wewn mwwuwuwumwww peee 15w new ueeexm .wnwewu new we even eew Eeew 11 ewweww xuw>wpuw \wEeoew eew .epe w .mwwwee eew wmwew>w ewweww .epe m 1mew muueneee zewwn new uwwe .ew -e=m .mwwwuxwu .wwwpm .e=w_eeu 1we .wpwmwemm< Awewuewee meeeee ueeeew m neew-oomp xwewuewne mmmw-eomw meseem peeexw N meewmwe ww we mwweemwuwe mm m: meow meow: m: «Now xewzx mumpncgozh m: Nwmw w wowee meewewm mew; eewpueew woeewe eewuwewemmwmwe weeum eww» eoewae tmwxeeeeow-.e-e wwwew 22 .wuewwewexw m: ee mwmweQEm .mwuese ueeexw xewueawe> memew> mwueee ueeeew we meewuwewwesw wwwwuwwee new wwseeeew eeweowwew one we ewaem -- -- 1- - - «ewe eoemmeom ewmwee .wmeez we meewewe xwpemwwm eewpwweswm ew ewwwwem m .meeeem wm: .ewppwe appemwwm mwp new ewweww xwewuewee 1new w .mpeee ewwwwem eewuwewumw eH nwmmwwe: 1mew weep-mmm— 15w wuwmwemm< w: «New eemwwz .mpeeexw ee uewwww Emweweows wwwee-eee w mwe newewn uwumween «we» mwmweueeae wee nwuwwnwm .eeuezew ewww mewewe eew eew meewuweew newswn w1u new ewweww xwewuewee w ea v .m: we» mwwepezeo ueeeew wweewm nwuwewpmm nwmmwwe: 1mew neew-weep eew meeeem om ow meow ewewwex wwwum .mwueee new .Eewwoeuwe ueeeew .m.: we mpmee .uwws .mpueneee wee we eewuwowwwmmwww eewwn .mwwwu new eewpwueeeeu 1- -1 1- 1xww .ewmem m: meow ~eewz .nwewnwmeeU wwwe wxpewr1wxezm we pewwww .nwezuewu weew>we wwwewe waerEw we eewueeeeee new wwwwee nweez new mxewewm mew; eewpueew newewe eewuwmwemmwmwe weeum eww> eeew=< .woseeweow--.e-_ wwwee 23 w we wneewemwe wee wwweeeaew eew: wwwew> eewaemwa eeEw< wwwee wwwee ewee wuewnw>w xuxex ewwewp aweweewee meeeesw 1m=new ewex new weeeee wcsoe ewzem womaweez -moe meme-wmwe oewmwewme Ne weme eewemweow .wsewew nweez ew eezeem seewwez eew: wewee 1wewewn nweez wwwewwwe wnwee g: new we ewee mwmwee eeeeeam meeemwm mwweeeeew .wEewew we meowwww xw—weeew wwwee ewwewwm new wwwwwezw ewweww 1wewm meeeeew 1m=new mmem wewewewm ee eeewee< 11 1me_ Nwmw-oomw wewmwemm< we meme new ewex =.»ewwwemwww eeweeeweemwnz ee wen mwwewwwemwww eew wwwe ewweww. meeewwm m eeweweweemw 3eem mewemwm 11 -mew owmw-mmmw a wewmwemm< wwwzewew> meme eweg .meoe eweE eew nwwwmwewm we eewewneee eweewe 1wwwemewz ewee emwmmzm mewamwe wew .eewewwweeew 1ee3w we eewawwwwew an nweemwwe mweeee we ewweee .eeweweem newewn eeee awwwewewm wwnes Eewee xexex a ewwewp -xw new eeee -w_w=em .nwma mew N nwmmwwee 1mew meme-wmmw -Ew wewmwemm< meow me «new ewex .eewewenwe ewwee we mwweeeeew eeweew wee new mewweewe mewweneee wwwewe-eee we wnee wwwem 3:32 as. .5 82.5mm .. -- -1 :83 3.8: :2 8&8: meewswm mmwe eeweeeew newewe eewewmwemmwmwe weeum eww> eeeee< .wwzeweeoe-.e-e womee 24 .mwewewemw meew>wee ewee ewewwem wwwwewewm wwwewuwemwww wwwee .meeeexw we aweeem new eew newewn wewe -wemw ee nwme neeewwweww Ezewst eewewEeewew-wwee .mewemwe w>wewn ee nwme mwewewemw wewuwemwww meew>wew .eewsxewesw new .wewwwwz ,wE:Fe> wnwee .mwewee ee mewweewe wwwewe1eee new wwwewe eeee mewwanwe we euweew wee nwewwaN .eewewzew m: wee eew nwewwnwe wwwee ew eewwewm -eee: eewN .eewewwwe 1mwm meewwwwwwwew wowee ew wwewewwwwo .mwowee wewewewm eew; eweee wee new wwwee w>wewwwe eew: wee-1wnwewu eew nwewewemw meeweweew eew .Mmeweewee m ee wv eeeem we wowee .wmewew wowee xeaex w weee ewweww zweweewee -moe weee-mmme meeoexo eweoe eemew-wweee e w meeeem ewoeeweo weem nemew-esow mm xweweewwe ewweee meme-ewme -moe weee-omme meeeeEw wwweeeeew .new ewex new m mum? :wwemnwew zoewww m: away we ewe=ew wewmweem< a wnwewo meme new zweezz mxewewm mew; eeweweze newewe eewewmwemmwmwe weewm ewwe eeee=< '! .wazceecow--._-_ weeee CHAPTER 2 ANALYTICAL FRAMEWORK 2.l Import Demand and Export Supply Importance and Derivation Any attempt to predict changes in trade patterns must begin with the conditions of import demand and export supply. Relying upon traditional assumptions of consumer behavior,1 one can derive h th consumer's demand for the j the it good as: de = de (Pij-Pikayi) where p is the money price, y is money income, and k is a vector exclusive of j. If consumer decisions are independent of decisions made by others, the market demand function is the sum of the indi- vidual demand functions. If in addition, consumers face the same prices and the distribution of income remains constant, the market demand function may be expressed as: DJ = DJ (Pjspkav) where Y is the sum of individual incomes. 1For a general survey of consumption and production theory, see Quirk and Saposnik (89) and Ferguson (25), respectively. 25 26 In the absence of a domestically produced substitute, the market demand function is also the import demand function. Other- wise, import demand is the difference between domestic consumption and production at various prices. The existence of a domestically produced substitute is a difficult issue to resolve. This is espe— cially true when dealing with a category of commodities rather than a more precisely defined commodity. The Lancaster approach to con- sumer theory, emphasizing characteristics rather than goods, would differentiate between almost identical goods; an aggregate approach would treat all goods alike. In any case, the presence of domestic production means that the import demand function is now an excess demand function. Conse- quently, a theory of sgpply is necessary to explain import demand. Replying upon traditional assumptions of firm behavior, one obtains the firm supply function by substituting the profit-maximizing input demand functions into the firm's production function. If the output 'price is exogenous to the firm, the supply function of the 9th firm may be expressed implicitly as: Sod = 593 (pgj’pok) The aggregation of firm supply functions is subject to difficulty because of the possibility of non-price interdependence among the firms. These intra-industry externalities mean that the market supply function is not simply the sum of the individual firm supply functions. In the absence of external effects from outside the industry, however, the market supply function remains a function 27 of the output price and other prices. On the other hand, the presence of externalities from outside the industry means that market supply is also a function of the activity of other relevant industries. Consequently, the import demand function may be expressed in general as: M. = M. (P.,P ,Y,A. .1 J .1 k J) where Aj is a vector of activity variables relevant to the partic- ular industry. To extend the analysis to intermediate goods one need only begin with a firm's derived demand for an input rather than an individual's demand for a final good--no new assumptions are required, and import demand will remain a function of own price, other prices, and income-activity variables. Now consider the derivation of export supply. Following the same procedure used in deriving an import demand function, one may express export supply as: x. =xj’(PJ.,P J Y’Aj) k, In the presence of domestic production and consumption, the assumptions required to establish import demand and export supply schedules are identical, and the functions themselves are con- tingent upon conditions of demand ang_supply. 28 Zero Homogeneity_in Prices Under competitive conditions traditional economic theory suggests that neither demand nor supply curves exhibit money illusion. In other words, a proportionate change in all prices would leave all real variables unchanged. In this case demand and supply functions are homogeneous of degree zero in prices. Most previous studies have interpreted this as a justification for using only one price variable (the ratio of own price to that of the closest domestic substitute) in the estimating equation. There are at least three reasons for believing that the use of a relative price variable may significantly bias the estimated price coefficient. The most obvious objection is that the weights used in constructing domestic price indices are generally quite different from the trade weights. A second objection is that domestic indices usually include the prices of imported commodities. In this instance, the own price appears in both the numerator and denominator of a relative price variable. A third objection is related to the use of unit value trade data. The contracted price of traded goods when recorded at customs may refer to a completely different time period than the current domestic price index of that same category. Because of these objections, Murray and Ginman (76) recently tested the relative prize hypothesis at an aggregate level and rejected its general validity. These objections, however, do not necessarily disqualify relative price as a useful approximation. Its power in limited 29 samples may exceed the bias it introduces. Chapter 3 discusses this issue and the use of relative price variables in the estimation process. Functional Form The traditional choice of the mathematical functional form for trade functions is the double-log form. Its advantages include - the fact that the elasticities are themselves the estimated coeffi- cients. There are more flexible alternatives (e.g., the transcendental-log function) which allow the functional form to be dependent upon the sample.2 This approach, however, is best used when the other elements of the specification are relatively precise and the sample is fairly reliable. In light of the general volatility of trade functions and the uncertainty surrounding the sample data, the traditional choice of the double-log form appears to be the best alternative. Tariffs and Quotas The presence of a tariff means that the price paid by the consumer will differ from the price received by the supplier in a systematic fashion. This difference is easily accounted for in the case of ag_valorem tariff rates. Assuming that t equals the import tariff rate, the system of equations for a particular commodity category becomes: 2See Sargan (94, pp. 145-204). 3O Mj = Mj (Pj’Pk’Y’Aj) Xi = X'j (P'j,P'k,Y"A'j) Pj = P'j (l + t) Mj = X'j where primed variables refer to the rest of the world. Excluding transportation costs and other factors, the foreign and domestic prices are separated by an amount equal to tP'j. I The introduction of quantitative restrictions in addition to tariffs requires further modifications. Under certain condi- tions quotas and tariffs may be treated as static equivalents in terms of the final demand price and the quantity traded.3 One sufficient set of assumptions is that foreign producers, quota holders, and domestic producers all be subject to competitive conditions. If the quota is filled, this either proves that quota holders are competitive or eliminates the need for the assumption, since the only power a quota holder can exert is to fail to fully exploit the quota share. The significance of the other two assumptions is less clear. The two major manufacturing categories presently affected on a wide scale by quantitative restrictions are textiles and iron and steel. The assumption of competition 3The development of this issue may be traced through Bhagwati (l3) and (T4), Shibata (96), Holzman (36), and Ophir (80). For a brief treatment of the dynamic non-equivalence of tariffs and quotas see Kreinin (60). 31 is more applicable to the former than to the latter. For this reason the estimates must be viewed as crude approximations-~the degree determined largely by the lack of competitiveness. In the absence of monopoly or effective collusion, however, the estimates should provide a good first approximation. The full tariff is the sum of the explicit tariff and the tariff implicit in the quota, measured in per cent by the divergence 1". .27} of the domestic and world prices (ignoring transportation and other factors). The position of the import price, net of explicit tariffs, between the domestic and foreign prices will determine the allocation of tariff revenues. If the import price is equal to the foreign price, all tariff revenues are captured by domestics. On the other hand, if the import price differs from the domestic price only by the explicit tariff, the tariff revenue implicit in the quota is captured by foreigners. Due to the manner in which quotas are typically enforced, the implicit revenue tends to be captured primarily by the country controlling the administration of the quota.4 If quota allocations are controlled domestically, the revenue largely accrues to domestic sources, whether private or governmental; if controlled by foreign agents, the revenue accrues largely to foreigners. The system of equations for a commodity category restricted by a quota is: 4For a discussion of this assumption see Mintz (73) and Bergsten (ll). X'J=X' jj(P', P'k,Y ,A .) Pj = P (I + q) (I + t) MJ = X'j = 00 where q is the implicit tariff rate and Q0 is the quota restriction. Effects of Uncertainty The analysis thus far has been based upon perfect knowledge and foresight. Because this is not actually the case, when unfore- seen events occur complete adjustment of demand and supply to market conditions takes time. There are two fundamental approaches used to explain this delayed response. One approach draws upon the relationship between stocks and flows to account for the adjustment problems caused by uncertainty. One might, for example, draw upon the explanatory power of changes in the stocks of final goods and inputs. A variety of such variables have been used to explain trade flows.5 These include inventories of final goods and inputs, industry capacity utilization, and order backlogs. The other method of explaining the adjustment process is the use of lagged variables. These have been entered in estimating equations separately or as weighted combinations of variables. Table l-l provides a survey of the use of various lag techniques such as the Koyck, Almon, and Shiller lags. 5See Kwack (62), Gregory (34), Steuer, et al. (100), and Adams, et al. (1). . 1. 1..., 33 As for the present study, there are three reasons why none of these disequilibrium approaches are used. First, unit values already contain a weighted average of present and past prices, since the value reported is the contracted price rather than the current price. The difference between the two is a function of the contract- ing horizon in the particular industry. Traditional lag techniques are not precisely applicable in this case. Second, most previous 6 Since studies have indicated that price lags are relatively short. monthly or quarterly data are not used in this study, an equilibrium model does not appear unreasonable. Finally, an equilibrium model provides a conformity which simplifies comparisons of the relative results for the U.S., the E.E.C., and Japan. Institutional Parameters In addition to the primarily economic relationships discussed thus far, institutional parameters play a key role in affecting trade flows. The system of international exchange, for example, affects the volume and pattern of trade. Although greater flexibility is possible under current exchange conditions, this study will project estimates of the effects of trade liberalization based upon constant exchange rates. This is done for several reasons. One is that since the model is not all-inclusive, some approximating assumption is required in any case. A second is that since universal trade liber- alization is being considered rather than unilateral, the exchange rate effects should not be as large. 6See for example Rhomberg and Boissnneault (93) and Branson (IS). 34 In addition, there are many highly irregular exogenous variables which affect trade. Among these are changes in trade barriers, strikes, wars, changes in market structure, and the like. Where appropriate, these are noted and included as explanatory variables in dummy form. Often there are changes in the composition of a category which introduce errors in the computed quantity and unit value indices. This may be the result of changes within the population or changes within the indices. Where possible these changes are accounted for in the form of dummy variables. 2.2 Static Effects of Trade Liberalization Nature of Analysis It is important to note at the start of this section that the effects computed here are static price effects derived from partial equilibrium analysis. Dynamic elements such as technologi- cal growth, improved market structures, and changing tastes are not considered. With regard to partial equilibrium analysis, changes in the relative prices of non-tradables and the existence of non-zero cross-elasticities present innumerable unknowns. Johnson (45, p. 333) has indicated, however, that these may well be ignored for two major reasons. One is that the number of non-tradables is relatively small if one excludes services; the other is that the consequences of cross-price effects will tend to cancel. This is especially true in the case of multilateral trade liberalization where both imports and exports are expected to expand. 35 Prices and Trade Volumes The conversion of a tariff change to a percentage change in price depends upon the relevant import demand and export supply elasticities. A variety of formulas have been used for this purpose, and none of them are completely satisfactory. The formulas employed in the recent study by Jadlow (42) are used in this study. The percentage change in the import price in this case is: 6m (At) n + (I + t) m EX where Em and ”m are the prize elasticities of export supply and import demand, respectively. The percentage change in the export price is: nm (At) nm + ex (I + t) These formulas are used for three reasons. First, because the percentage change in import price is adjusted to account for the previous influence of a tariff on the price. Second, the per- centage changes in the import price and the export price are equal if the elasticities are equal--a requirement of constant elasticities. Third, these formulas represent a compromise among other alternatives. 7Starting at pre-tariff equilibrium the formulas would be ex/(nm + ex) and nm/(nm + ex), respectively. In the case of constant elast1cities, the general formulas derived by partial differentiation 36 Once the relevant percentage change in price is computed, the change in trade may be computed by multiplying the percentage change in price by the corresponding price elasticity and the original volume of trade. The result, of course, is the same whether one uses the import or export side of the market. Domestic Consumption and Production A change in trade volume implies a change in both domestic consumption and domestic production. With income constant, for example, an increase in imports implies both an increase in domestic consumption and a decrease in domestic production. The problem of estimating these changes has traditionally been ap- proached in terms of relative elasticities. It is easily proven that, by definition:8 D S n = -n + --e m M M where nm, n, and 6, represent price elasticities of import demand, demand, and supply respectively, and where M, D, and S represent import demand, demand, and supply, respectively. Under the as- sumption that two countries share common domestic price elasticities, these domestic elasticities can be obtained by solving the set of two equations. Balassa and Kreinin (5), for example, used this approach in estimating the effects of the Kennedy Round. 8nm - ‘Eg‘ (dLD-SI/dp) D n + 3'6 ZICJ 37 Unfortunately, the true D and S are unknown because composite categories inevitably include both imports and exports. Using the total D and S for the category or just netting out exports will lead to biased results. It is possible, however, to modify the tradi- tional identity so that under certain assumptions the bias is eliminated. First assume that all goods within a category are homogeneous. The identity in this case is: P om = - -——————— (d[D-S+X]/dP) (D-S+X) D. S X = -n + —-€ - —-€x M M M Dropping the assumption of homogeneity, assume that goods within the category can be categorized (on the basis of cross-price elastici- ties) as either importables or exportables so that where the subscripts m and x refer to importables and exportables, respectively. The true identity for the elasticity of import demand may now be expressed as 38 The difficulty, of course, is that Dm and Sm are unknown. The 'question now is under what assumptions is the modified identity equivalent to the true identity? This is the condition that If nm = nx = n and Em = Ex = e, the condition holds.9 The original, unmodified identity holds only if exports in the category equal zero. Subtracting exports from S, a correction sometimes made, will yield results equivalent to the true identity only if none of the exportables are consumed at home (Sx = X and Dx = O). In a recent study, for example, Magee (69, p. 665) complains that the domestic demand and supply elasticities derived from using the traditional identity are unrealistically low. This is precisely the point. To be consistent with a given import elasticity, the domestic elasticities must be unrealistically low because the con- sumption/import ratio and the production/import ratio are much higher than the true values. The modified identity presented here will help in removing this bias. This identity should, however, be viewed purely as an accounting relationship and not causal in any way. 9Subtract the true identity from both sides and divide through by M to obtain m m Dxn + Sxe - Xex - 0 Substituting and rearranging, one obtains x x _ Dxn + Sxe - Xex which is the true identity for export supply. 39 Unfortunately, knowing the internal elasticities still does not enable one to allocate a change in imports or exports between domestic consumption and production. Traditionally, n/ (n + 6) has been used as the domestic consumption share in the total change and 10 These formulas are e/ (n + e) as the domestic production share. valid only if one begins the analysis at a no-trade equilibrium I where D = S. In general the true formulas are Dll/ (Dn + Se) and 1] Again, one must know the true 55/ (On + Se), respectively. values of D and S, which are unavailable. In light of these difficulties, the simplest procedure appears to be the use of arbitrary shares. The upper limit to the value of the supply share is when the domestic supply is infinite and all adjustment takes place on the supply side. Wemelsfelder's study (llO, pp. 94-l04) in Germany found that liberalization in the late l950's resulted in a greater contraction of production than in an increase in consumption. This implies a lower limit to the supply share of one-half. The mid-point of this range is three— fourths for the production share and one-fourth for the consumption share. This is compatible with Magee's (69, p. 665) assumptions in his recent study. 10See, for example, MacPhee (67, p. 39). II = D Wé- Dn + Se n1n M” M6 "on+ssI"m) I Dn/ (Dn + $5) + 58/ (Dn + Se) 40 Domestic Employment The change in domestic production caused by the net change in the trade sector implies a change for domestic employment in that industry. To estimate the magnitude of this change without specific information concerning the industry one must make some assumptions regarding the relationship between output and employ- ment. if one assumes that production in the industry is subject to constant returns to scale and that wages move proportionately with the prices of competing inputs, the labor-output ratio will remain constant. The first assumption makes the factor-use ratio dependent only upon the factor-price ratio; the second assumption means that the factor-price ratio is constant. The change in production multiplied by the labor-output ratio will give the implied change in employment under these assumptions. Welfare Gains The basic framework of the Marshallian approach to measuring welfare changes, utilizing the net change in consumer and producer surplus and tariff revenues, is well defined and will not be presented 12 in detail here. However, it may be useful to underscore some of the major assumptions of this approach. i. The presence of some form of social utility structure must be assumed. ii. Production must take place on the production frontier. 12For‘a detailed presentation of this approach as it applies to international trade see Johnson (45). 41 iii. The trade functions used must either be "income—compensated" or exhibit zero income elasticity. iv. No goods disappear from the market.13 The welfare change in an importing country caused by a tariff reduction is measured by subtracting the fiimporting“ surplusM before the tariff reduction from that after the reduction and adding the net change in tariff revenues. This clearly requires the assumption that dollar values have the same welfare weight in all sectors. This procedure is easily demonstrated below in Figure 2-l where Q and P are quantity and price, respectively. a 3 P2_— XI Q1 Q2 Figure 2-l.--Tariff-Induced Welfare Changes. 13See Leamer and Stern (63, pp. l96-197). 14"Importing surplus" is used for lack of a better term. Consumer surplus is misleading because an import function implies botdi consumption and production effects domestically. 42 X] and M refer to export supply and import demand, respectively. The subscript 1 refers to trade positions prior to the tariff reduc- tion; subscript 2 refers to those after the tariff reduction. Based upon a linear approximation, the net change in importing surplus may be expressed as: (P1 ' P2) 01 + a (P1 ' P2) (02 ' 0]) = (a + b) The second expression (b) is the familiar deadweight loss triangle. The net change in tariff revenues is measured by: (e+c)-(a+e+d) (c - a — d) (P2 - Pé) QZ - (P1 ' Pi) Q] The total change in welfare, therefore, is: 3: ll (Pl-P2) Q] + 5(P1‘P2)(02’Q]) + (Pz'P'2)Qz ’ (PI'PI])Q] (b + c - d) where w represents the welfare of the importing country. The welfare changes for the exporting country are derived in like fashion except that no tariff revenues are involved. Based upon linear approximation, the change in welfare of the exporting country may be expressed as: N. = (P'z'P'])Q] + %(P'2'Pl])(Qz“Q]) : (d + f) 43 It is important to note for future reference that calcula- tions of the deadweight loss triangles involve the square of the change in price. More specifically, I — 1 C" 2 éAPAQ - 6(bAP) nm(P)(Q) This is significant partly because the square of an average tariff is not necessarily equal to the average of the squared component tariffs. Chapter 3 discusses the factors used to correct for this aggregation bias. Evaluating the overall gains over time is subject to diffi- culty. Ordinarily, the appropriate overall measure of welfare changes is their net present value. This is the measure used by Magee (69) in assessing the aggregate benefits of moving to free trade. Un- fortunately, this measure may be more misleading than informative. The volatile nature of trade functions15 means that the calculation of present values through an infinite future may be subject to great error. As a result, it may be fruitful to look at net present values for shorter time spans. The most appealing measure along this line is the per period welfare effect after adjustment is complete. This provides a meaningful measure which may be implicitly evaluated by the reader. 15As an example, see Wilson (112, pp. 50-1, 105, 109, 114). CHAPTER 3 METHODOLOGY AND DATA 3.l Treatment of Variables Commodity Categories International trade statistics are compiled and published by the United Nations (l9) according to the Standard International Trade Classification (SITC). A recent GATT tariff study (9) has combined three and four digit SITC groups into "industry" level classifications and computed tariff averages for each. These industry classifications, with some modifications, form the basis for the categories used in this study. In general, categories for which the tariff is less than five per cent for all three principals were omitted. This is also true of the raw material portions of some categories. The resulting categories are presented in Table 3-l along with a category number and description, a convenient abbreviated description, and the SITC coverage. These thirty-seven categories account for almost all of the trade in manufactures that is sig- nificantly affected by tariffs or quotas. The most significant categories not covered are airplanes and ships. These were excluded because the data are inadequate and because the markets are so heavily influenced by government policies and purchases. 44 TABLE 3-l.--Commodity Categories. 45 Cat. # Category Description Abbreviation S.I.T.C. 1 Leather articles and Leather mfgs 6ll-l3, semi-manufactures 841.3,842 2 Rubber articles and Rubber mfgs 62l,629, semi-manufactures 84l.6 3 Wood and cork Wood mfgs 631-33 manufactures 4 Paper manufactures Paper mfgs 64l-42 5 Textile semi- Tex semi-mfg 261-67, manufactures 65l-53 6 Textile articles Tex articles 654-57 7 Clothing and Clothing 84l excl accessories 84l.3 & 84l.6 8 Mineral manufactures Mineral mfgs 66l-63, 666 9 Class and glassware Glass mfgs 664-65 lO Iron, steel and I & S, unworked 67l ferro-alloys, unworked ll Iron and steel I & S 672-79 semi-manufactures semi-mfgs l2 Aluminum and Aluminum 684 aluminum products l3 Other non-ferrous Other metals 682-689 metals and products l4 Metal manufactures Metal mfgs 691-698 l5 Petroleum manu- Petrol mfgs 332 factures 16 Organic chemicals Org chem 5l2 l7 Inorganic chemicals Inorg chem 513-15 TABLE 3-1.--Continued. 46 Cat. # Category Description Abbreviation S.I.T.C. l8 Dyeing, tanning and DTC mat 531-33 coloring materials 19 Plastic materials Plastics 581,893 and articles 20 Essential oils, Oils, perf 551, perfumes, toilet 553-54 preparations, soaps, cleaning compounds 21 Other chemicals Other chem 541,571, 599 22 Power-generating Power mach 711 machinery, non- electric 23 Agricultural Ag mach 712 machinery, non-electric 24 Office machines, Office mach 714 non-electric 25 Metalworking Metal mach 715 machinery, non-electric 26 Textile and leather Tex mach 717 machinery, non- electric 27 Other machinery, Other mach 718-19 parts and accessories, non-electric 28 Electrical machinery, Elect mach 722-23, tools and parts 725,729 29 Telecommunications Telecom 724 apparatus TABLE 3-1.--Continued. 47 Cat. # Category Description Abbreviation S.I.T.C. 30 31 32 33 34 35 36 37 Motor vehicles and parts Miscellaneous transport equipment and parts Precision instru- ments-- professional, scientific and controlling instruments, photographic apparatus, clocks and watches Footwear, travel goods and handbags Photographic and cinematographic supplies Furniture Musical instru- ments, sound recording or reproduction apparatus, and sound recordings Toys and sporting goods Motor veh Misc trans Prec instr Shoes, bags Photo mfgs Furniture Sound mfgs Toy mfgs 732 excl Can- adian trade for U.S. 731,733 726,861, 864 831,851 862-63 821 891 894.2 & .4 48 Price and Quantity Variables Published data according to the SITC begins at the beginning of the last decade. The first few years, however, are probably subject to errors of classification due to the adjustment to the SITC system. The sample period for this study begins in 1963 and ends in 1972, and the chosen observation period is semi-annual. This minimizes the problem of coordinating the data for the three principals. The statistics actually published are value and quantity figures. In many cases, however, quantity figures are not reported. In a few cases this necessitated the use of partner trade data. Other countries' data for exports to the U.S., for example, was substituted for unavailable U.S. import data. Care was taken in these instances to include a representative sample of trading partners. If one divides value by quantity, a unit value is obtained. For each category, price indices were constructed from component sub-categories using these unit values. Paasche price indices were used because of their computational simplicity and because they maintain a unique relationship to the “true" price index.1 Laspeyres quantity indices were computed by dividing value by the corresponding Paasche price index. A Laspeyres quantity index was used because it maintains a unique relationship to the "true" quantity index. These indices were computed at the most disaggregate 1For an excellent discussion of price indices and the unique quality of Paasche price indices and Laspeyres quantity indices, see Fisher and Shell (26). 49 level generally available throughout the sample period. For the U.S. this is the seven digit level of the Schedule A for imports (107) and the Schedule B for exports (106). For the E.E.C. and Japan, the most disaggregate level is generally the three or four digit level. Data collection for the E.E.C. was particularly difficult because the data must generally be collected on an individual country basis. This was further complicated by the fact that intra-E.E.C. trade must be subtracted from the total to obtain E.E.C. trade with the rest of the world. A special effort was made to include data for the three new entrants into the E.E.C. If one of these three accounted for ten per cent or more of E.E.C. trade in a category in 1972, data for that country was added to the E.E.C. total. When such data were unavailable in international sources, a special effort to collect the data from national sources was made if the country's share exceeded 25 percent in 1972. Reference to national sources was complicated by their use of different classification systems and national currencies. In practice, these procedures resulted in the inclusion of U.K. data in most categories and exclusion of Danish and Irish data in almost all. These procedures, however, should make the resulting estimates fairly close approximations of the true elasticities for the total E.E.C. Unfortunately, it is too soon to account for any shifts in the structural parameters that may have been caused by the reduction of trade barriers among the original E.E.C. members and the new entrants. 50 Alternative Price Variables The prices of domestic substitutes were collected for each category and principal. For the U.S., these prices are the whole- sale price index for the appropriate category. In most instances, the domestic category is reasonably compatible with the trade category. For the E.E.C. alternative prices were computed from the intra-E.E.C. trade flows previously collected in the process of obtaining the extra-E.E.C. trade statistics. The Japanese alter- native prices were collected from national sources and converted from yen into dollars. These prices generally refer to a slightly broader category than the trade category itself and are not as compatible as the U.S. and E.E.C. statistics. The "rest of world" alternative prices present a more difficult problem. Since such statistics are not available for a sizeable number of countries, the "rest of world" price used in a particular function is a weighted average of the alternative prices of the other two principals. The weights in each category are equal to each principal's average relative share in the value of trade.2 2The price and income weights used here are the weights traditionally used: destination (buyer) weights for exports and origin (seller) weights for imports. In general these weights seem quite appropriate, but the weighting issues involved are far from settled. Clearly, an exporting country's share in the imports of another country is important to the weight its price and income variables should have, but its import share in total world exports may also be important. Take the extreme example of petroleum as a case in point. The traditional weights in this case would be each exporter's share in the imports of a particular country. Equally important, perhaps, are the import shares of other industrial countries in total world trade in petroleum. ll ll'll.‘l. ' I1 "II' ‘III I. all. Ill 11. '11.-..111 1| 51 Income-Activity Variables Income-activity variables are available on an annual basis for a wide range of countries. However, these variables are not generally available for LDC's for periods shorter than a year. The one variable which is available on this basis for LDC's is industrial production. Therefore, this is the income-activity variable used in each equation. The weights for the "rest of world" variable in each category are the average relative shares in the value of trade. Tariffs The tariff rates used in this study are taken from the recent GATT tariff study (9). The tariff averages derived from world trade weights were used to avoid the bias inherent in using a country's own import weights. World weights are also biased to the extent that they do not reflect what the free trade composition of trade would be for the particular country. Recent changes however, in the composition of trade due to exchange rate realignments and higher petroleum prices make the world weights a preferred alternative to "own-import" weights, since the latter have probably changed since the 1970 base year in the GATT study. The tariff rates for chemicals in the GATT study were computed on the assumption that the Kennedy Round ASP package would be imple- mented. Since this has not been the case, these rates were adjusted using U.S. Tariff Commission data (102). The rates presented for the U.S. do not account for the fact that the base for some tariffs is the "American Selling Price." This is likely to be a separate issue, and the recent study by Jadlow (42) examines this issue exclusively. 52 U.S. tariff rates are generally applied on free-on-board (f.o.b.) value, while E.E.C. and Japanese rates are applied on coast-insurance-freight (c.i.f.) value. 1974 U.S. data (35) on the ratios of these two values were used to adjust the potential U.S. tariff changes to a comparable c.i.f. basis. The “world" tariff rates computed by the GATT study include all the major industrial market economies. For each principal, however, it was necessary to extract the impact of its own tariff on the world tariff to obtain the average tariff levied on its exports. The computation of the tariff changes implied by Scheme A and Scheme B (p. 5) requires some knowledge of the individual tariff rates within each category. The change in tariff for each case is: Scheme A: (.6 Ewiti + .4 Ewitz) 2 Scheme B: Z wit, where ti refers to all_the component tariffs, tZ refers to all those equal to or less than five per cent, and wi refers to the weight assigned to each tariff. To avoid using the thousands of tariff lines in each princi- pal's tariff schedules, this study employed the technique used recently by Magee (69). This technique requires only a summary of the distribution of tariffs within each category. The mid-point of a small range of tariffs was substituted for the t1 and t2, and 53 the value share of that range of tariffs was substituted for the original wi. The tariff distribution data are furnished in a study by the U.S. Tariff Commission (102). With very few exceptions, the distribution tables are generally compatible with the categories used in this study. Where this was not the case, further research was undertaken in an attempt to modify the table. The tariff data in the Tariff Commission report are weighted using own-trade weights. To make the results compatible with the GATT tariff averages, the following conversion factors were computed using the distribution data: .6Zwit1. + .42witZ Scheme A: .62wit1. 22w.t? 1 1 Scheme 8: —————-——- (Zwiti)2 The relevant change in tariff was obtained by multiplying these conversion factors by the expression in the demoninator as computed from the GATT data. In this way very good approximations to the Scheme A and Scheme B tariff changes were obtained without having to refer to the thousands of individual tariff lines. These con- version factors are also used in Chapter 4 to correct the aggregation bias in the welfare effects discussed in Chapter 2. 54 Trade Volumes The base trade volumes used to compute the changes in trade were taken from 1974 U.N. data (19). It is important to use the most recent year available to minimize the distortion due to composition changes caused by currency realignments and the quad- rupling of petroleum prices earlier in this decade. Since only E.E.C. imports and Japanese imports are expressed as c.i.f. value, these were converted to f.o.b. value to be compatible with the other trade values. The U.S. ratios mentioned earlier were used for this purpose, even though they are not exactly applicable. However, most of the variable factors such as composition of trade by distance and method of transportation tend to be offsetting. In summary, all value statistics are reported as f.o.b. value in 1974 U.S. dollars. There are several issues relating to exclusions from the trade totals in each category. Trade with centrally planned economies was excluded because of their varying MFN status with the different principals and because of the government control over trade. E.E.C. trade with the remaining members of the European Free Trade Association (E.F.T.A.) was excluded because agreements eliminating tariffs on mutual trade have been made by the E.E.C. with most of the members of the E.F.T.A. and negotiations are underway with the remainder. Trade with LDC's is a more difficult issue to resolve. 0n the import side the "Generalized System of Preference" (GSP) schemes provide preferential treatment to LDC exports into the 55 U.S., the E.E.C., and Japan. A reduction in the MFN tariff rate, therefore, may result in a reduction rather than an increase in imports from LDC's by reducing the preference margin. There are two reasons for believing that the LDC share in the imports of the U.S., the E.E.C., and Japan will remain relatively constant. First, the GSP schemes are not really effective in sensitive (i.e., high tariff, high price-elasticity) manufacturing categories.3 Second, LDC's have contended that an attempt should be made to maintain preference margins by reducing preferential rates to zero and liberalizing non-tariff barriers on products included in the GSP schemes.4 It is assumed here that some effort to maintain preference margins will be made on the part of industrialized countries. For these reasons, LDC exports were included in the import totals for each principal. 0n the export side, LDC's are not required to make "fully reciprocal" reductions in their own tariffs in order to enjoy reduced MFN rates on their exports. However, there is reason to believe that LDC imports will expand roughly in proportion to the general expansion in the exports of industrialized countries. The fact that LDC's will not be held responsible for implementing a reduction formula generally applied by the industrialized countries does not mean that some reciprocity will not be expected and obtained. Because of the generally high tariff barriers among 3For a general survey of preference schemes, see Iqbal (41, pp. 34-39). 4See Anjaria (4, pp. 25-28). 56 LDC's, even a small degree of liberalization is likely to put the LDC reductions on a par with the general level of reductions. Even if this does not occur, however, an expansion in LDC exports to industrialized countries as a result of trade liberalization will eventually necessitate an expansion in imports of roughly the same total magnitude. The composition of this trade balance effect is difficult to determine, but the effect is clearly in the direction of expanding LDC imports from industrialized countries. Therefore, LDC imports were included in the export totals for each principal. Labor-Output Coefficients Chapter 2 described the procedure for obtaining the U.S. employment effects presented in Chapter 4. This procedure requires the use of a labor-output coefficient for each category. These coefficients were obtained from 1974 U.S. data (35). The number of employees per million dollars of output was adjusted by an average work week factor to derive a labor-output coefficient which expresses the number of "forty-hour-equivalent" workers per million dollars of output. All data refer to the 1974 period to make the coefficients compatible with the 1974 trade data. These coefficients, of course, are only averages and will not be exact for imports or exports. However, there is little presumption as to the direction of any possible bias.5 In any case, the net industry effects are probably more reliable than the individual import and export effects. 5See TC Publication 473 (20. pp. 145-152). 57 3.2 Choice of Estimator Import Demand Equations Criteria for choosing an estimator are always a critical concern in empirical work. Johnston (47, pp. 408-420) presents a general survey in his text of the characteristics of the major estimators, and Mikhail (72) offers recent Monte Carlo comparisons of these estimators. In addition, Sawa (95) has developed criteria for choosing the optimal k-class estimator and for establishing the mean square error (MSE) dominance of ordinary least squares (OLS) over two-stage least squares (ZSLS). The limited sample (20 observations) constructed for this study and the importance of isolating possible specification errors in the export supply equations should eliminate full—information methods from consideration. The most promising of the remaining alternatives are OLS and ZSLS. Were it not for two problems, OLS would be the clear choice between the two. First, there is the potential for a special sort of measurement error (p. 3) in which errors in the right hand unit value variable are correlated with errors in the left hand quantity variable. Second, there is also the possibility of simultaneity bias. In these two cases the OLS estimates are inconsistent. The trade-off between the two estimators is between the expected smaller variance of the OLS estimator in small samples and the expected smaller bias of the ZSLS estimator. The Sawa criteria are indecisive under the special circumstances here, and l 11 111 111‘ '11 1111||ll||l|11j|lll1lll|lll|llulil 1|Illll1 l7 58 a test developed by Feldstein (23) is inapplicable because the test depends upon the true value of the parameter when the measure- ment errors on the left and right hand sides are correlated. Since a more precise criterion for choosing between OLS and 2SLS is not available, a crude set of criteria is presented below. These are based on the trade-off between bias and standard error and on the expected directions of bias. The first test is a simple sign test of the two price coefficients. The price elasticity of import demand for manu- facturing categories is almost certainly negative for the three principals. Therefore, if the coefficient using one estimator is positive and the coefficient using the other is negative, the latter was selected. In most cases, however, both price coeffi- cients are negative. In the few cases when both are positive, neither was selected. The second test is a more complicated "counter-expectations" test. The unit value problem will tend to bias the OLS coefficient toward minus one, and the simultaneity problem will bias the coefficient toward the positive value of the export supply coeffi- cient. When the OLS coefficient is separated from the direction of these biases by the ZSLS coefficient (i.e., OLS < -1 and OLS < ZSLS), one may reasonably conclude that the bias in the OLS coeffi- cient is probably not greater than the bias in the ZSLS coefficient. OLS was chosen in this instance because of its smaller standard error. 59 '- The third test is an “absolute mean error" (AME) test to be used when the first two are inapplicable.. Ordinarily, one might use a mean square error test of the bias and variance of the two estimated coefficients. This test, however, is severely biased in favor of OLS. In such a small sample one cannot really assume that ZSLS has removed all of the bias in the price coefficient. Use of this coefficient as the true value in a relative mean square error test will bias the test in favor of OLS because the bias is squared. This is easily demonstrated by the fact that (b- B)2> (b- B)2 when 5 lies between b and B. b, 5, and B refer to the OLS, ZSLS, and true coefficient, respectively. Since the OLS standard error is understated when unit value errors are present, the use of squared standard errors (variance) further biases the test in favor of OLS. If one uses absolute differences rather than squared differences, the bias toward choosing OLS is reduced. When the first two tests are inapplicable, the OLS, ZSLS, and true coeffi- cients will all be the same sign. If one also assumes that the ZSLS coefficient lies between the OLS coefficient and the true coefficient, the following relationship holds: lb-BI-IE-0|=|b-E| 60 This means that: ~ AME(b) - AME(b) = lb-Bl+ob-lb-Bl+og= lb - bl + ob - Ob One chooses OLS when this difference is negative and ZSLS when it is positive. This test is less biased in favor of OLS because it is independent of the true value of the parameter and because the downward biased OLS standard error is not squared.6 The procedures above were followed in all but about four cases. In these exceptions the ZSLS estimator usually appeared superior on the basis of these criteria but was clearly outside the range of feasibility, given the estimates of the other equa- tions. The estimate in each case was outside this range by a factor of about two to four. Export Supply Equations A representative sample of about a third of the export supply equations was estimated using a variety of estimators, variables, and specifications.7 Out of this sample of about 70 6No particular statistical properties are claimed for the AME other than its being a combined measure of bias and standard error which is less biased toward OLS than an MSE criterion. It is clearly not, for example, a measure of the second moment, the MSE. 7Among these are FIML, 3SLS, ML, ZSLS, IV, OLS, industry- specific activity variables, normalization on price, and linear functions. 61 equations only about two estimated equations could be taken seriously as true export supply functions. It appears that given the sample data and available variables, successful estimation of relatively disaggregate export supply functions is virtually impossible. One positive aspect of these results is that the difficulty in identifying export supply equations probably means that the simultaneity bias is small in the import demand equations. This may be simply the result of the volatility of export supply relative to import demand or the result of a recursive system in which price is determined exogenously in the export supply function. The consequence of these disappointing estimates of export supply is that arbitrary assumptions about the various price elasticities of export supply must be employed. The traditional assumption has been that export supply is infinitely elastic. While export supply is probably more elastic than import demand, there is some evidence that export supply is less than infinitely elastic whether one is concerned with export supply to or from each princi- pal. The control group studies by Kreinin (58) and Krause (55), for instance, indicate that export supply for large traders is far from infinitely elastic. By comparing a tariff-reduced group with a non-reduced group, Kreinin concluded that: It appears plausible that close to half of the benefit from tariff concessions granted by the United States accrued to foreign exporters in the form of increased prices (58, p. 317 . 62 This short-run result probably represents a lower limit to the export supply elasticity to the U.S. The supply elasticity for U.S. exports is likely to be higher because exports comprise a smaller share of the U.S. market. The two extremes above are used as the limits to the price elasticity of export supply in each instance. The upper limit is infinite elasticity, and the lower limit is an elasticity equal to the corresponding price elasticity of import demand. The mid- point of this range is consistent with the assumption made by Balassa and Kreinin (5, p. 130) in their study of the Kennedy Round. 3.3 Use of Relative Price The objections to the use of relative price have already been discussed in Chapter 2. All equations are estimated using separate price variables as a result of these objections. In such a limited sample, however, the use of relative price as an approximation may reduce the variance of the price coefficient by more than the bias it introduces. Consequently, in cases where relative price might be critical, estimates have been made using relative price. Unless the choice is clear on the basis of sign, the decision to substitute these estimates for the original esti- mates requires the use of a non-central F test as developed by Wallace and Toro-Vizcarrondo-(lOB). The use of relative price implies that the two price coeffi- cients are equal but opposite in sign. For the restricted estimator b' a test of the hypothesis 53 Ho: MSE(b') §_MSE(b) can be based on the critical points in the non-central distribution F(m,T-K,k) where m equals one, T is the number of observations, and K is the number of variables. The hypothesis was tested at the five per cent level of significance. The test was used for both OLS and ZSLS, although it is not strictly appropriate to the latter.8 1 The two-step process of first estimating without relative price and then using relative price as an approximation is clearly a regression strategy. This affects the distribution of the sampling statistics in the cases in which relative price was actually used. This is not too serious in practical terms for two reasons: First, the use of relative price represents a return to the traditional method of estimation; and second, the instances in which relative price was used are clearly indicated so that the appropriate degree of skepticism about the test statistics can be shown. 81f OLS and ZSLS estimates are affected differently by the use of relative price. the choiCe between OLS and ZSLS may be affected. In practice, this possibility never occurred. CHAPTER 4 ESTIMATES OF PRICE ELASTICITIES 4.1 Empirical Estimates Empirical estimates of the price elasticities of demand for the imports and exports of the U.S., the E.E.C., and Japan 1 are presented for each category in Tables 4-1 through 4-37. A representative estimating equation is: lnMj = a + b11nPj + bzlnPk + b3lnY + b4S + e where Mj = the import quantity index for the jth category P. = the import price index including tariff for the jth 3 category P = the alternative price index for the jth category = the appropriate index of industrial production S = the semi-annual dummy variable e = the error term 1Although price elasticities are the primary concern of this study, the income elasticity estimates are also presented. These are generally positive and significant, as expected. These elasticities, however, need not be positive because of the differing effects of activity on domestic demand and supply. Magee (68. PP. 188-192) provides an excellent survey of this issue. Also, any time trend effects will impact primarily upon the activity coefficient since activity is highly correlated with time. No attempt has been made to extract these possible trend effects because the price elasticities are the primary concern. 64 65 The estimator selected for each equation (OLS or ZSLS) is listed beside each set of estimates in the tables. The elasticity estimates are accompanied by their t-statistics. These t-values are in parentheses beneath the coefficient to which they pertain. The number in parentheses in the column headed "t P DIFF" is the t-statistic for the difference between the own price coefficient and the alternative price coefficient. This t-statistic is equal to:2 b1 + b2 2 2 1/051 + 062 + 2 Est. Cov. (6],b2) If no statistic appears in this column, a relative price variable was used. In this case, the two price parameters are restricted to be equal but opposite in sign. The R2 column presents the coefficient of determination after correction for the degrees of freedom. The computed value of the Durbin-Watson autocorrelation statistic and the degrees of freedom are reported in the next two columns. In a typical case where relative price was not used, where no special dummy variables were used, and where a Cochrane-Orcutt transformation was not made, there are fifteen degrees of freedom (20 observations less five explanatory variables). 2kmenta (54, p. 372). 66 4.2 Arbitrary Estimates Arbitrary estimates of price elasticities are also presented in the tables. When imports are less than ten per cent of the exports (or vice versa) in a category, no empirical estimates were obtained for that equation and an arbitrary price elasticity was assumed. Arbitrary elasticities were also assumed when a reliable estimate of the price coefficient was not obtained by empirical estimation. These arbitrary elasticities were imposed on the basis of seven assumptions. These assumptions are: 1. An estimate from another source is applicable. 2. An estimate from another category is applicable. 3. An estimate from another category is applicable after scale adjustment using another principal's estimates. 4. An estimate using the average of a group of categories is applicable. 5. An estimate using the average of a group of categories is applicable after a scale adjustment based on another country's average. 6. The E.E.C. elasticity is the average of the U.S. and Japanese estimates. 7. The U.S. elasticity is the average of the E.E.C. and Japanese estimates. These assumptions are referred to by number each time an arbitrary elasticity is used. The selection of the appropriate assumption depends on the circumstances of each case and is basically a subjective process. The logic behind each assumption is straight- forward except, perhaps, for assumptions (6) and (7). One normally expects that, ceteris paribus, the U.S. import price elasticity will 67 TABLE 4-1.--Leather Manufactures. EQN EST Price t P DIFF Income 82 0w DF USM OLS -1.33 (1.93) 2.39 .944 2.23 15 (l 81) (9.77) usx OLS -1.57 ( .34) —- .41 .711 1.94 15 (2.83) ( .73) EM OLS - .29a 2.05 .944 1.53 15 1.09 (3.21) EX 2SLS -1.23 ( .75) 1.05 .955 1.20 15 (2.55) (8.87) JM zsLs - .94 (7.97) 1.00 .959 1.14 15 (4.91) (9.21) 0x OLS -1 09 (2.51) 3.57 .953 1.15 15 (2.14) (5.49) aSubstitute -l.l4 based on assumption (6). TABLE 4-2.--Rubber Manufactures. EON EST Price t P DIFF Income R? on DF USM zsLs -2.22 (3.41) 2.40 .854 1.73 15 (1.99) (2.99) usx OLS -1.31 (1.74) .84 .757 2.52 15 (4.13) (4.71) EM OLS - .70 (1.05) 1.54 .945 1.18 15 (1.44) (3.67) EX OLS - .03a 1.01 .920 . 1.53 15 ( .05( (10.58) JM b JX ZSLS -1.44 ( .78) 2.19 .930 2.21 15 (1.92) (4.83) aSubstitute -l.38 based on assumption (6). bTota1 is less than 10% of exports. Substitute -1 97 based on assumption (4) and categories 1, 3 and 4. 68 TABLE 4-3.--Wood Manufactures. EON EST Price t P DIFF Income 82 DW DF USM OLS - .05a (1.40) .18 .351 1.81 15 ( .14) ( .59) usx OLS -2.14 ( .52) .80 .881 1.79 15 (4.13) (1.50) EM ZSLS -1.57 ( .54) 1.43 .910 2.02 15 (1.94) (5.25) EX OLS - .05b .23 .585 1.91 15 ( .25) (5.56) JM zsLs -2.95 ( .01) 3.15 .851 1.58 15 (3.18) (3.52) 0x zsLs -1.53 ( .14) .35 .501 1.85 15 (4.25) ( .99) aSubstitute -2.26 based on assumption (7). bSubstitute -l.84 based on assumption (6). TABLE 4-4.--Paper Manufactures. EON EST Price t P DIFF Income R2 DM DF USM OLS - .82 .53 .729 1.29 15 (1.55) (5.17) usx ZSLS -4.01 (3.95) .81 .977 2.27 15 (10.91) ( .99) EM zsLs - .80 .77 .953 2.32 15 .91) (l 40) Ex zsLs -1.09 (8.44) .31 .951 1.59 15 (3.41) (2.04) JM OLS -2.01a (1.55) .94 .975 1.39 14 (18.38) (4.81) JX OLS -1 50 (2.07) .92 .957 1.34 15 (4.92) (2 50) aTransformed variables (Cochrane-Orcutt). 69 TABLE 4-5.--Textile Semi-Manufactures. EON EST Price t P DIFF Income PZ 0w DF USM ZSLS -1.51 (2.12) .55 .422 1.79 15 (2.37) usx OLS .97a (3.35) .45 .522 1.53 15 (1.82) (3.15) EM OLS - .84b 1.04 .830 1.24 15 (1.35) (3.58) EX 2SLS - .51 ( .57) 1.37 .950 2.01 15 (2.84) (10.53) . JM OLS -1.50 (2.90) 1.75 .959 1.38 15 (2.97) (5.23) 0x zsLS -2 30 (4.11) 3.40 .989 1.88 15 (2.55) (35.10) aSubstitute -l.46 based on assumption (7). bSubstitute -l 51 based on assumption (6). TABLE 4-6.--Textile Articles. EQM EST Price t P DIFF Income 82 0w DF USM OLS - .73 ( .22) .85 .783 1.58 15 (3.14) (1.34) usx OLS - .79 (1.49) .57 .944 1.50 15 (2 26) (1.25) EM \ OLS - .74 (1 25) .77 .740 1.53 15 (2.33) (l 63) Ex OLS - .51 ( .55) .75 .955 2.05 15 (2.59) (9.03) JM ZSLS -l.06 (1.32) 1.44 .985 2.79 15 (2.97) (6.35) ax zsLS - .20a 1.01 .905 1.48 15 ( .62) (7.25) aTransformed variables (Cochrane-Orcutt). based on assumption (7). Substitute -.97 TABLE 4-7.--C10thing. 7O EQM EST Price t P DIFF Income 82 0w DF USM zsLs -1 24 (7.94) 1.22 .978 1.75 15 (3.34) (1.45) usx ZSLS -2.44 ( .35) .74 .954 2.19 15 (3.55) (l 20) EM ZSLS .83a (2.31) 3.42 .957 1.98 15 ( .78) (4.42) EX OLS - .73 (2.13) .05 .871 1.32 15 (1.99) ( .16) JM OLS -1.25 ( .97) 2.08 .945 1.07 15 (3.59) (4.28) 0x zsLs -2 01 1.52 .505 1.83 15 (2.55) (3.50) aSubstitute -1.25 based on assumption (5). TABLE 4-8.--Mineral Manufactures. EQN EST Price t P DIFF Income R2 DW DF USM OLS 2.05 (4.55) .47 .852 1.43 15 (3.95) (1.70) usx OLS - .84 (4.17) 1.52 .525 1.22 14 (3.15) (3.72) EM OLS -1.27a (4.15) 2.05 .951 2.37 14 (8.01) (5.78) EX OLS - .98 (2.35) .58 .949 1.01 15 (3.85) (2.32) JM b 0x ZSLS -1.52 (1.90) .81 .894 1.88 15 (3.11) (3.08) aTransformed variables (Cochrane-Orcutt)- b on assumption (2) and category 13. Total is less than 10% of exports. Substitute -l.22 based 71 TABLE 4-9.--G1ass Manufactures. EQN EST Price t P DIFF Income R2 on DF USM ZSLS -l.38 ( .53) .73 .852 1.14 15 (2.14) (2.14) usx OLS -1.05 ( .45) .14 .775 1.31 15 (3 95) ( .55) EM 2SLS -1 31 (1.42) 1.41 .877 1.29 15 (3 95) (3.28) EX OLS -1.19 (1.19) .74 .899 2.27 15 (4.48) (3.14) JM OLS -1 22 ( .53) 1.75 .975 1.45 15 (5.02) (13.12) 0x ZSLS -1.49 ( .79) 3.92 .932 1.03 15 (5.53) (4.97) TABLE 4-10.--Iron and Steel, Unworked. EQN EST Price t P DIFF Income R2 DM DF USM OLS - .51a (2.22) .13 .514 1.08 14 ( .72) ( .31) usx b EM 2SLS - .88 (1 58) 1.32 .755 1.85 15 (3.97) (4.55) EX 25LS -1 22 ( .05) .07 .837 1.00 14 (4.06) ( .35) JM OLS -2.43 (3.25) .72 .823 1.18 14 (5.51) (3.55) JX ZSLS -4.90 ( .24) .21 .597 1.17 15 (3 61) ( .30) aSubstitute -2.83 based on assumption (2) for category 11. b Total is less than 10% of imports. on assumption (2) for category 11. Substitute -l.20 based 72 TABLE 4-ll.--Iron and Steel, Semi-Manufactures. EQN EST Price t P DIFF Income 82 0w DF USM 2SLS -2.83 (1.58) 1.79 .757 1.55 15 (1.99) (3.13) usx OLS -1.20 - .14 .458 1.50 15 (3.51) ( .54) EM OLS -1.55a (1.54) 1.17 .945 1.83 14 (5.90) (4.55) EX zsLS - .53 ( .59) 1.20 .855 1.52 15 (1 59) (5.47) JM b JX OLS -1 72 ( .55) 2.40 .988 .77 14 (4.94) (16.05) aTransformed variables (Cochrane-Orcutt)- bTotal is less than 10% of exports. Substitute -2.43 based on assumption (2) and category 10. TABLE 4-12.--Aluminum Manufactures. EON EST Price t P DIFF Income R2 DW DF USM OLS -2.51 (1.00) 1.11 .591 1.59 14 (1.10) (2.33) usx OLS 4.09a (1.08) 1.09 .505 1.03 15 (3.89) (2.24) EM OLS .52b (1.22) .98 .787 .55 15 ( .47) (2.13) EX OLS -1.47 (4.33) 1.41 .881 1.75 15 (3.87) (8.04) JM ZSLS 3.57C ( .59) 2.40 .878 .95 15 (1.18) (7.05) 0x OLS -1.70 (1.05) 1.29 .753 1.75 14 (2.42) (1 38) aSubstitute -l.24 based on assumption (6). bSubstitute -l.21 based on assumption (2). cSubstitute -.91 based on assumption (2). 73 TABLE 4-13.--Other Metals. EON EST . Price t P DIFF Income R2 DW DF USM ZSLS -3.49 (3.75) 1.91 .859 2.19 14 (4.53) (3.25) usx OLS -1.53 ( .35) 1.22 .828 1.49 15 (5.43) (2 89) EM zsLS -1.21 (1.15) .48 .452 2.17 15 (1.54) (3.21) EX ZSLS -1.82 ( .25) .54 .838 1.50 15 (4.49) (2.07) JM OLS - .91 (1 53) .09 .751 1.84 14 (3.45) ( .85) 0x 2SLS -2.47 ( .54) 2.50 .942 1.08 14 2.18) (5.29) TABLE 4-14.--Metal Manufactures. EQN EST Price t P DIFF Income R2 on DF USM 015 .50a (2.25) .58 .927 1.28 15 (1.00) (2.70) usx ZSLS -3.51 (7.07) 3.17 .894 2.11 15 (5.75) (8.78) EM OLS - .52 (1.85) 1.27 .973 1.31 15 (2 27) (7 64) EX OLS -1 21 ( .52) .99 .931 2.28 15 (3.91) (4.81) JM b JX 2SLS -1.55 ( .80) 2.70 .942 1.79 14 (1.98) (2.30) aSubstitute -l.35 based on assumption (5), categories 11, 12, 13 and the E.E.C. bTotal is less than 10% of exports. Substitute -.41 based on assumption (5), categories 11, 12, 13 and the E.E.C. 74 TABLE 4-15.--Petroleum Manufactures. EON EST Price t P DIFF Income 82 0w DF USM OLS - .01a (2 40) 1.37 .914 .84 15 ( .01) (4.87) usx OLS -1.13 (5.41) - .19 .525 1.52 15 (2.85) (1 81) EM OLS - .50 - .32 .781 1.77 15 (1.90) (3.82) EX OLS - .23b ( .95) 1.37 .957 1.33 15 (1 50) + (4.85) JM OLS - .50 (1 09) .57 .548 1.18 15 (1.08) (4.73) 0x C aSubstitute -1.14 based on assumption (3), category 16, and the E.E.C. bSubstitute -.57 based on assumption (3), category 16, and the U.S. CTotal is less than 10% of imports. Substitute -l.69 based on assumption (7). TABLE 4-16.--Organic Chemicals. EQN EST Price t P DIFF Income 82 DW DF USM OLS - .02a ( .25) 3.08 .958 1.55 14 ( .08) usx 25LS -2.89 (4.59) .72 .875 2.51 15 (3 15) (2.41) EM OLS -1.12 (5.91) 2.14 .985 2.09 15 (5.45) (14.15) EX zsLS -1.45 ( .92) 1.53 .993 1.15 15 (5.59) (31.36) JM OLS -1.75 (2.00) .95 .830 1.75 15 (8.14) (8.55) 0x OLS - .55b (1.84) 2.77 .954 2.73 15 (1.51) (5.81) aSubstitute -2.13 based on assumption (4) and categories 17-21. bSubstitute -3.49 based on assumption (5), categories 17-21, and the E.E.C. 75 TABLE 4-17.--Inorganic Chemicals. EON EST Price t P DIFF Income P? on DF USM 2SLS -3.40 2.14 .770 1.03 15 (2 88) (5.18) usx OLS -1.51 (1.53) .99 .953 1.55 15 (8.49) (8.32) EM OLS -1.296 (13.05) 1.94 .981 1.28 14 (9.51) (15.18) EX OLS - .58 ( .28) 1.05 .971 1.53 15 (2.57) (5.14) JM zsLS - .79 ( .72) 1.29 .972 1.70 15 (8.03) (13.84) JX OLS -1.15a (2.31) 3.15 .958 1.94 14 (3.75) (8.85) aTransformed variables (Cochrane-Orcutt). TABLE 4-18.--DTC Materials. EQN EST Price t P DIFF Income 82 0w DF USM zsLS -3.71 ( .51) .95 .929 2.31 14 (4 54) (1.64) usx OLS .21a (1 31) .52 .945 1.40 15 ( .59) (2 92) EM ZSLS - .35b .54 .933 .97 15 ( .82) (2 76) Ex zsLS - .70 (1.29) 1.41 .989 2.29 15 (1.15) (10.33) JM zsLs -2.10 .55 .915 1.41 15 (4 91) (8.86) 0x OLS -1 43 (1.41) 2.53 .991 1.70 15 (4.11) (5.97) aSubstitute -l.O7 based on assumption (7). b Substitute -2.91 based on assumption (6). 76 TABLE 4-19.--Plastic Manufactures. EON EST Price t P DIFF Income 82 0w DF USM zsLS -2.32 (1.81) 2.57 .931 2.50 14 (5.97) (5.83) usx OLS -1.00 ( .51) .87 .890 1.57 15 (2.12) (5.35) EM OLS - .97 (2 24) 2.25 .989 2.54 15 (4.00) (15.22) EX 2SLS - .74 ( .58) 1.98 .995 2.20 15 (3.31) (24.84) JM OLS -1.94 (2.50) .72 .899 1.15 15 (5.35) (9.63) 0x 25LS -3.94 (2.17) 1.10 .981 1.15 15 (3.34) ( .97) TABLE 4-20.--Oils, Perfumes. EQN EST Price t P DIFF Income 82 DW DF USM ZSLS - .57 ( .19) 1.35 .889 1.58 15 (1.82) (4.39) usx OLS -1.05 (1.08) .34 .329 1.58 15 (3.51) (2.20) EM OLS - .48 1.33 .978 2.28 15 (4.11) (28.44) EX zsLS -1 01 (1 51) .97 .952 2.22 15 (2.01) (12.37) JM OLS -1 25 ( .11) .94 .953 1.22 15 (5.49) (19.53) 0x 2SLS -1.75a (5.33) - .44 .752 1.43 15 (2.43) ( .32) aSubstitute -1.11 based on assumption (7). 77 TABLE 4-21.--Other Chemical Products. EON EST Price t P DIFF Income R2 DW DF USM OLS - .55 (3.28) 1.38 .888 1.90 15 (l 55) (4.37) usx 2SLS - .55 ( .42) .41 .354 1.45 14 (1.34) (3.01) EM OLS - .85 (2.54) 1.88 .887 1.81 15 (4.54) (9.09) EX ZSLS - .59 ( .35) 1.28 .992 2.41 15 (6 64) (32.27) JM OLS -1.42 (2.55) .93 .953 2.50 15 (6 81) (17.85) 0x 2SLS -1.24 (1.13) 3.12 .977 2.73 15 (4.75) (18.52) TABLE 4-22.--Power Machinery. EON EST Price t P DIFF Income R2 DM DF USM OLS - .346 (2.85) 5.43 .949 1.03 15 ( .51) (7.84) usx OLS -1.94 (1.49) 2.33 .921 1.58 15 (5.50) (4.45) EM OLS .50b (2.90) .75 .875 1.83 15 (1.41) (1.63) EX OLS - .48c 1.48 .890 1.39 15 (l 54) (10.87) JM OLS -1 47 (1 01) 1.55 .819 1.55 15 (3.54) (4.85) 0x OLS -1.15d 4.05 .873 2.25 15 (1 55 (10.87) aSubstitute -2.17 based on assumption (5), categories 23-27, and Japan. b Substitute -l.83 based on assumption (6). cSubstitute -1 55 based on assumption (5), categories 23-27, and the U.S. dSubstitute -2.66 based on assumption and the U.S. (5), categories 23-27, TABLE 4-23.--Agricultural Machinery. 78 EQN EST Price t P DIFF Income R2 DM DF USM OLS - .94 (1 55) 1.08 .801 1.58 15 (4.61) (2 29) usx zsLs - .44a ( .54) .48 .298 2.18 15 ( .26) ( .42) EM OLS - .87 - .88 .539 1.11 15 ( .82) (2.07) EX zsLS - .88 (1 13) .22 .987 1.59 15 (8.99) ( .65) JM ZSLS - .55 (1 17) 1.59 .935 2.21 15 (3.50) (8.33) 0x OLS - .44 ( .82) 5.33 .961 1.22 15 (3.94) (5 97) attempt to improve the fit. (7). aAn industry-Specific activity variable was added in an Substitute -.66 based upon assumption TABLE 4-24.--Office Machinery. EQN EST Price t P DIFF Income R2 DM DF USM 2SLS -2.89 .30 .948 1.12 15 (5.58) ( .38) usx ZSLS -1.37 (1 70) .25 .839 1.57 15 (3 11) ( .42) EM OLS - .02a ( .49) 1.98 .927 1.88 15 ( .06) ( .88) Ex OLS - .15b ( .55) 1.35 .900 1.02 15 ( .76) (5.72) JM OLS -1.45 (1.56) - .10 .951 1.82 14 (4.95) ( .55) 0x OLS - .95 ( .03) 3.97 .725 1.11 14 (4.14) (3.73) aSubstitute -2.18 based on assumption (6). b Substitute -l.17 based on assumption (6). 79 TABLE 4-25.--Metalworking Machinery. EQN EST Price t P DIFF Income 82 DW DF USM 2SLS -1.29 (2.39) 1.13 .555 1.10 14 (2.18) ( .69) usx 2SLS - .70 (1.44) .55 .525 1.72 14 (2.58) (2 57) EM OLS -2.01a (2.80) .95 .815 1.49 14 (4.45) (3.65) EX OLS -1.01 ( .30) .83 .938 1.28 15 - (4.54) (3.24) JM zsLs -1.75 (2.83) 1.09 .854 2.33 14 (5.21) (4.70) 0x zsLs -l.81 (2.41) 3.88 .955 1.98 14 (7.32) (10.23) aTransformed variables (Cochrane-Orcutt)- TABLE 4-26.--Textile Machinery. EQN EST Price t P DIFF Income R2 DM DF USM OLS .43a (2.39) 1.84 .913 1.91 15 ( .90) (4.31) usx OLS -1 37 (1 79) - .51 .827 1.53 14 (4.42) (1 29) EM OLS - .03b (2.55) 1.35 .558 1.42 14 ( .06) (4 49) EX OLS - .57 ( .78) 1.49 .983 1.48 15 (3.93) (13.64) JM OLS - .54 (1.31) .75 .745 2.30 14 (1.93) (5.13) 0x OLS - .40c ( .34) 1.05 .825 1.59 14 (l 37) (2.67) aSubstitute -.90 based on assumption (5), categories 22-25 and 27, and Japan. bTransformed variables (Cochrane-Orcutt). Substitute -.77 based on assumption (6). CSubstitute -2.07 based on assumption (7). 80 TABLE 4-27.--Other Machinery. EQN EST Price t P DIFF Income R2 DW DF USM OLS - .72 (4.58) 1.98 .982 1.58 15 (3.53) (7 94) usx 2SLS -1 77 (2.15) 1.53 .815 2.57 15 (1 33) (5.83) EM zsLS -3.55a (1.33) 1.51 .751 2.18 15 (2.02) (3.39) EX OLS - .29b ( .70) 1.31 .981 1.11 15 (1.54) (6.25) JM 2SLS - .49 (1.55) .54 .927 1.20 14 (1.42) (4.57) 0x zsLS -2.48 ( .27) 2.05 .985 1.91 15 (3.40) (3.98) aSubstitute -.61 based on assumption (6). bSubstitute -2.13 based on assumption (6). TABLE 4-28.--Electrica1 Machinery. EQN EST Price t P DIFF Income R2 DW DF USM OLS -1 24 (4.49) - .04 .945 1.38 14 (2.41) ( .05) usx OLS -1.00 ( .50) .87 .970 2.15 15 (7.09) (11.60) EM OLS - .98 (1 51) 2.35 .985 2.20 15 (9 98) (19.87) EX OLS - .48 (1.52) 1.54 .972 2.07 15 (3.53) (9.43) JM zsLS -1.11 (1 90) 1.31 .987 2.72 14 (7.29) (11.20) .Jx ZSLS -1.09 ( .45) 2.55 .928 1.57 14 (5.34) (5.13) 81 TABLE 4-29.--Te1ecommunications Apparatus. EQN EST Price t P DIFF Income R2 DW DF USM 2SLS - .14a (4.08) 1.42 .955 2.18 15 ( .45) (2.14) usx OLS - .59 (5.41) 2.64 .975 2.35 15 (4 16) (15.35) EM zsLS -1.95 2.48 .923 1.25 15 (6 97) (7.82) Ex 2SLS -1.37 (3.33) 3.12 .958 2.35 15 (3.05) (5.38) JM b 0x zsLS -1.85 (1 12) 6.53 .955 1.84 14 (2.99) (5 66) aSubstitute -2.47 based on assumption (3), category 28, and the E.E.C. bTotal is less than 10% of exports. Substitute -2.22 based on assumption (3), category 28, and the E.E.C. TABLE 4-30.--Motor Vehicles. EQN EST Price t P DIFF Income 82 0w DF USM zsLS -2.66 (9.88) 2.54 .975 2.27 14 (4.90) (5.81) usx zsLs -2.93 (3.54) .00 .887 2.28 15 (4.43) ( .02) EM 2SLS -2.49 ( .77) 2.30 .893 2.30 15 (4.07) (2.75) EX 25LS - .74 (5.15) 1.14 .975 1.51 15 (2.03) (9.15) JM 3 0x OLS -4.21 (1.98) 5.25 .858 2.45 14 (2 04) (5.45) aTotal is less than 10% of exports. Substitute -2.32 based on assumption (6). 82 TABLE 4-3l.--Miscellaneous Transportation Equipment. EQN EST Price t P DIFF Income R2 DW DF USM USX EM EX JM JX 094910-109! aThe data for this category are generally inadequate for direct empirical estimation. As an approximation, substitute the relevant values for category 30 based on assumption (2). TABLE 4-32.--Precision Instruments. EQN EST Price t P DIFF Income 82 0w DF USM OLS - .75 (3.73) 1.22 .965 1.47 15 (3.45) (2 77) usx OLS -1 36 (5.35) 2.40 .940 1.49 14 (4.19) (15.95) EM 2SLS -1.20 ( .23) .39 .991 2.33 15 (5.12) (1 17) EX OLS - .33a 1.41 .980 2.17 15 (1 72) (8.53) JM OLS -1 17 (5.82) 1.57 .982 1.15 15 (11.71) (18.67) JX 2$LS - .96 ( .45) 1.34 .957 1.79 14 (1 15) (3.53) aSubstitute -l.l6 based on assumption (6). 83 TABLE 4-33.--Footwear, Travel Goods and Handbags. EQN EST Price t P DIFF Income R2 DW DF USM OLS - .15a (1.06) .55 .951 1.03 15 ( .53) ( .89) usx ZSLS -l.84 (1.48) .73 .459 1.50 15 (1 75) (3.61) EM ZSLS .40b (3.95) 2.01 .963 1.22 15 (1.25) (15.54) Ex ZSLS -l.81 (1.59) .81 .955 1.41 15 (2.67) (1.28) JM C ox 2SLS - .85 1.95 .754 1.58 15 (5.53) (4.90) aSubstitute -l.33 based on assumption (2) and category 1. bSubstitute -1.14 based on assumption (2) and category 1. CTotal is less than 10% of exports. Substitute -.94 based on assumption (2) and category 1. TABLE 4-34.--Photographic Manufactures. EON EST Price t P DIFF Income R2 DW DF USM OLS -2.40a 2.52 .933 1.33 15 (5.34) (7.50) usx OLS - .93 (2.25) 1.74 .981 1.33 14 (2 50) (14.59) EM OLS -1.95 ( .03) 1.84 .988 1.75 14 (4.97) (5 42) Ex 2SLS -1 25 1.31 .945 .99 15 (2 25) (15.99) JM OLS -1.15 (3.02) .98 .934 1.95 15 (4.99) (7 74) JX OLS -2.52a (1.54) .03 .959 1.17 15 (4.47) ( .04) aTransformed variables (Cochrane-Orcutt). 84 TABLE 4-35.--Furniture. EQN EST Price t P DIFF Income R2 DM DF USM OLS .30a (6.02) .30 .951 1.57 15 ( .76) (1 12) usx OLS -1 25 (2 72) .99 .885 2.19 14 (6.14) (4.05) EM 2SLS - .70 (3.37) 1.59 .952 1.51 15 (1 72) (7.16) Ex zsLs - .15b 2.15 .983 1.25 15 ( .40) (16.32) JM 0x aSubstitute -.81 based on assumption (1) and Kreinin, 1970 (Household Durables). bSubstitute -l.25 on the assumption that the U.S. and E.E.C. elasticities are the same. cTotal is very small. Substitute -.59 based on assumption (6). dTota1 is very small. Substitute -1.25 on the assumption that the U.S., the E.E.C., and Japan share the same elasticity. 85 TABLE 4-36.--Sound Manufactures. EQN EST Price t P DIFF Income R2 DM DF USM OLS -l.89 7.35 .905 1.38 15 (1 45) (5.71) usx OLS - .91 ( .30) 1.57 .919 2.04 14 (3.89) (5.28) EM zsLs -1.48 ( .56) 2.82 .951 1.42 15 ( .80) (4 86) Ex zsLS -1 94 ( .01) 1.79 .977 1.52 15 (4.81) (7.57) JM b 0x OLS - .10 ( .08) 8.29 .981 1.32 14 ( .24) (7.30) aTransformed variables (Cochrane-Orcutt). bTota1 is less than 10% of exports. Substitute -1.07 based on assumption (6). CSubstitute -l.94 on the assumption that the E.E.C. and Japanese elasticities are the same. TABLE 4-37.--Toys. “—1.. EQN EST Price t P DIFF Income 82 0w DF USM OLS - .12a (9.14) .39 .958 2.64 15 ( .52) (1 48) usx OLS .27b (1.55) .74 .800 1.95 15 ( .93) (3.98) EM 2SLS 1.12a (1.14) 2.00 .945 1.18 15 (1.92) (8.46) Ex 2SLS -1.49 (2.53) .99 .975 2.35 15 (3.69) (5 13) JM OLS -1.04 (1 47) .53 .943 2.46 14 (5.00) (4.21) 0x OLS -1.22 ( .57) 1.50 .954 1.50 15 (6 91) (10.27) aSubstitute -1.04 on the assumption that the U.S., E.E.C., and Japanese elasticities are equal. bSubstitute -l.36 based on assumption (7). 86 be the largest and the Japanese elasticity the smallest of the three elasticities. This is expected because imports compete with a larger volume of substitutes within the United States. This is one rationale behind assumption (6). On the export side one expects the opposite to be true for the demand for exports from the U.S. In this case, the price elasticity of demand for U.S. exports is expected to be the smallest due to its relatively large share of the market. Clearly, this is also a rationale for assumption (6). In some cases, however, the price elasticity of demand for U.S. exports is greater than the Japanese (or the E.E.C.) elasticity. In some groups of industries the price elasticity of demand for E.E.C. exports is consistently less than both the U.S. and Japanese elasticities. This is particularly true in textiles, petroleum manufactures, chemicals, and motor vehicles. These results can only be explained by differences in the composition of commodities in the category and differences in the composition of trading partners. The near-monopoly trading position of some E.E.C. countries with former colonies may be part of the explanation in some categories. In any case, these results are the rationale for assumption (7). 4.3 Autocorrelation If error terms are serially correlated, the standard errors of the estimated coefficients may be understated. Consequently, statistical tests based on these standard errors may not be valid. This is more vc 0f indL variab‘ the pr Proced lS nee ten p corre POsit Of SE \ OY‘n nifi a fe beCa 87 The Durbin-Watson statistic offers one method of testing for such autocorrelation. Many of the statistics computed for each esti- mating equation fall in the inconclusive region of the test because of the small sample size. There is evidence that when variables follow a trend without much fluctuation, the inconclusive region contracts toward the value of the upper boundary of the test.3 This is not as serious for trade variables because they are generally more volatile than comparable domestic variables. Moreover, indices of industrial production also tend to fluctuate more than income variables such as gross national product or gross domestic product. In cases where the Durbin-Watson statistic clearly indicated the presence of autocorrelation,4 a two-stage Cochrane-Orcutt procedure was used.5 This procedure was also used if the statistic is near the rejection limit for the null hypothesis (i.e., within ten per cent of the rejection limit when tested against negative correlation and within twenty-five per cent when tested against positive correlation).6 This should further reduce the probability of serious autocorrelation. 3Kmenta (54, p. 297). 4A two-tailed test against the alternative of either positive or negative correlation is used at the five per cent level of sig- nificance. 5See Kmenta (549 pp. 287-288) for a brief explanation. In a few cases the second-stage estimation has not been carried out because the original estimates did not warrant the effort. 6The percentages differ because the scales of the limits differ greatly for 14-16 degrees of freedom. 88 When autocorrelation appeared to be a major problem, OLS was chosen over ZSLS. The Cochrane-Orcutt procedure was then applied to the OLS estimates. This procedure is simpler and consistent with Monte Carlo results by Hurd (39, p. 573). One of his major conclusions is that when there is modest autocorrela- tion, OLS is generally superior to ZSLS. 4.4 Evaluating the Results Two hundred twenty-two estimates of price elasticities of import demand are required to estimate the effects of trade liberalization for the U.S., the E.E.C., and Japan in the thirty- seven categories. Of these two hundred twenty-two, empirical 7 One estimates are presented in Section 4.1 for two hundred four. hundred sixty-two of these (seventy-nine per cent) are actually selected for use on the basis of reliability. Thus, a total of sixty arbitrary elasticities are presented along with the empirical estimates. Of the price elasticities estimated directly, twenty- eight per cent are either positive or insignificantly negative at the five per cent level, and nineteen per cent at the ten per cent level. Many of the estimation difficulties are concentrated in "problem" industries. These categories appear to present an industry-wide pattern of poor results. These include Aluminum (12); Petroleum Manufactures (15); Dyeing, Tanning, and Coloring 7The remaining eighteen are not estimated because of their relative unimportance. 89 Materials (18); Power-Generating Machinery (22); Textile Machinery (26); Furniture (35); Sound Manufactures (36); and Toys (37). No characteristic common to all these categories which might offer an explanation is apparent. However, there are several concentrated industries among those in the list (e.g., Aluminum, Petroleum Manufactures, and Power-Generating Machinery). In these cases, the expected price responses may be distorted by market structures. It is interesting to note that a dummy variable entered in the U.S. equations for Motor Vehicles (30) to account for the Canadian-U.S. Automotive Products Agreement in 1965 was very insignificant. This does not mean, however, that the elimination of tariffs between the U.S. and Canada had no effect on the bilateral trade flow,8 since Canadian trade is excluded from the category. It does mean that the Agreement may have had little or no effect on trade with others. This is not really surprising given that much of the trade between the U.S. and Canada is intra- firm trade in intermediate products. The relative performances of OLS and ZSLS warrant special attention. Of the empirical estimates of price elasticities actually used, OLS was chosen as the "best" estimator in fifty- two per cent of the cases, and ZSLS was selected in the remaining forty-eight per cent. The two estimators, therefore, performed about equally well. Given the greater efficiency one expects of 81n fact, the Agreement has probably been a major factor in this bilateral trade. See, for example, Officer and Hurtubise (79, p. 325). 9O OLS in such small samples, the roughly equivalent performance of ZSLS indicates significant potential bias in using OLS estimates. The importance of relative price in obtaining reliable results is also a primary concern. The two price coefficients are significantly different from each other in about a third of the cases at the five per cent level and in about one-half the cases at the ten per cent level. Estimates using relative price were made for many equations for which the original estimates were poor. Of these, relative price improved the mean square error of the estimate in twenty-four cases. However, the estimates in ten of these instances were still unreliable and were replaced by arbi- trary estimates. Therefore, there are only fourteen equations in which the use of relative price appeared to be a critical factor in obtaining reliable results. This rather weak power overall, however, should not diminish the importance of relative price in the individual fourteen equations. It is difficult to compare the elasticity estimates presented here with previous estimates because previous estimates either do not exist or they refer to categories that are not comparable to those used here. A limited comparison is possible, however, with Kreinin's study (61) of disaggregate import demand functions for the U.S. A careful comparison of the categories in that study and the present study indicates ten categories which are almost exactly comparable. The respective estimates of price elasticity of demand for these categories are presented in Table 4-38. The estimates from the Kreinin study are all OLS estimates taken from the period 91 TABLE 4-38.--Comparison of Estimated Price Elasticities. U.S. Price Elasticities of Import Demand Category Stonea Kreininb 1 Leather mfgs -1.33c - .74 2 Rubber mfgs -2.22 - .39 18 DTC mats -3.71 -1.55 20 Oils, perfumes - .67 - .46 23 Ag mach - .94C - .67 25 Metal mach -1.29 - .98 27 Other mach - .72c - .92 28 Elect mach -1.24C - .92 33 Shoes, bags - .15C - .79 34 Photo mfgs -2.40d -1.08 aTaken from Tables 4-1 through 4-37, semi-annual data, 1963-1972. Unless otherwise indicated, estimates are for ZSLS and do not use relative price. bkreinin (61), quarterly data, 1953-1970. All estimates are for OLS and use relative price. c0LS. dRelative price variable and OLS. 92 1964 through the first quarter of 1970. Relative price and real gross national product were used as the price and income variables, respectively. In all but two instances, the estimates from the present study are higher, most by a substantial margin. These higher elasticities are most likely the result of the frequent use of ZSLS and the relatively rare use of relative price. The implications of this study regarding the performance of ZSLS relative to OLS and the performance of relative price tend to support this conclusion. Such a conclusion must remain tentative, however, due to the possibility of intervening factors. The elasticity for Photographic Manufactures, for example, is higher than the Kreinin estimate, but both are OLS estimates based on relative price. The higher elastici- ties, therefore, might be a "quirk" in these ten categories, although the fact that eight of ten are higher make this an unlikely explana- tion. The additional two and three-quarter years (19701-1972) in this study's sample period and the drastic exchange rate realignments during these years may also be an important factor. Still, these years represent only about twenty per cent of the sample. The shift in the price elasticity would have to be quite large to have a substantial impact on the estimated elasticity. Yet another possibility is that the use of semi-annual rather than quarterly data could explain the higher elasticities. This would be true, for example, if the Hnoisef from the unit value measurement problem has a greater tendency to "cancel" over 93 a six month period than over a three month period. It would also be true if the simultaneity bias were less in semi-annual data. This could be the case if export supply were substantially more elastic over a six month rather than a three month period. Some doubt is cast on this explanation, however, by Kreinin's rejection of preliminary estimates using a price variable lagged one quarter. Finally, the use of different activity variables may be Significant. An index of industrial production may be a better activity variable than real gross national product. On the other hand, it is difficult to believe that this could make such a con- sistent and substantial difference in the estimates. CHAPTER 5 STATIC EFFECTS OF TRADE LIBERALIZATION 5.1 Introduction This chapter presents estimates of the initial price effects of trade liberalization on tariffs, trade volumes, and welfare. A section estimating the employment effects in the U.S. labor market is also included. These results are presented primarily in tabular form. An attempt has been made to report the effects in each category in as much detail and from as many different perspectives as possible. The narrative, on the other hand, attempts to explain the origin of each table, the extraordinary elements of the detailed effects, and some of the implications of the results. 5.2 Tariffs and Tariff Changes Computation of Tariff Changes Table 5-1 presents the post-Kennedy Round MFN gg_valorem tariff rates and the tariff changes implied by Schemes A and B, respectively. Scheme A calls for a reduction of all tariffs by sixty per cent and elimination of tariffs of five per cent or less; Scheme 8 calls for a reduction in tariffs equal to the initial height of each tariff. The U.S. import tariffs are based on f.o.b. value, but the computed tariff changes were converted to c.i.f. 94 95 am. am. em. om.m a~._ mo.F m.m x5 55. em.a mm.e Fe. No.3 4F.F .o.F M.“ :5 mm. mm. om. mo.m 4N.P m~.. P.e xm mm. em.m mm.m o~.F am.m e~.F oo.P m.m 2“ oo.- oo.F so. om.a 3N.P oo.F m.~ xm: F_.F oo._ oo.~ 3F. Fo.P mo.P _m.e 0N.P 5., 2m: mace codes a oo. _o._ mm. me.m mo._ oo._ 8.5 xe mm. mm. m_.P o_._ NF.c mo.. oo._ N.op :6 oo._ _o._ mm. as.m mo.. .o._ m.m xm mm. oo._ mN.P mN._ ma.m mo._ oo.F w.oP 2m oo.F oo., mm. oe.m wo.P oo.~ m.m xm: _e. oo.P oo.P mN.F ep.m o-.P ma._ _o._ 5.3 gm: mace ace: m No._ F_.P mo. N~.m RN.P m~.P _.m xw a_._ so._ m_._ 38. wc.e mo., co._ m.“ 25 am._ om.F mN.F em.e ma._ _a.P m.m xm ao._ co.~ o~.P we. mm.e Fo._ oo.P m.m gm oo._ 00., mo. mo.o _o.F oo.P P.o_ xm: om._ oo.~ 00., mm. 32.3 mo.- ma._ Ne.F m.m 2m: meme coaasm N 00.2 mm. mm. _m.m ma._ _N._ o.m xw me. me._ om._ om._ No.5 a_.F oo._ N.F_ :6 mo._ mm. mm. _m.m mm.F a_._ m.m xu mm._ me. am. No. No.m ma.~ om._ 5.8 2m co.” oo.P mm. mm.m Na._ _N._ Fo.m xm: mm. oo.F 00., em. oa.m ao.P ._a.- NP.P _o.m 2m: mace coepeou F ua.m m < m < mou\uHu m < cewcae E.ES acomopeu x\z 4mm m: 3mm ea ce_cae a xoaeH cameo .uomeaeo mmecap bee uttwcae--.F-m mum

55555 55 55. 55. 55.5 55.5 55.5 55.5 5.5 55 55.5 55. 55. 55. 55.5 55.5 55.5 5.5 55 55. 55. 55.5 55.5 55.5 55.5 5.5 55 55. 55.5 55.5 55.5 55.5 55.5 55.5 5.55 55 55.5 55.5 55.5 55.5 55.. 55.5 5.55 555 55.5 55.5 55.5 55. 55.5 55.5 55.5 55.5 5.5 555 5555555 55 5555 5 5 5 5 5555555 5 5 555555 5555 55555555 555 555 55 555 55 555555 5 55555 55555 .Uw325ucounu._um m4m< ><><><><><3 X3><333>< 3 3X33 333 3 X3 ><><><333 3 33 X33 3 X 3X><><>< ><><33>< ><><3>< ><>< ><3 >< ><3 among twelve most import-biased categories. among twelve most export-biased categories. Note: M 110 11.10”] 5N. 5m. 5N. mm. mm.N m—.m xa m5.o_ o5.op mm. mp._ mN.e mN.m 25 mo. 55. 55. 5N. mm._ mN.m xm om.o mN.m 55. um. mN.N mm.5 :5 oo.p oo.5 Nm. 55.N o~.m om.o~ xm: oo.~ oo.5 mo. 55. 55. Nw. zm: 5555 55555 5 55. N5. N5. mm.5 PN.5 mo.m 55 m5.5 Nm.5 55.5 mm.N mm.w 55.05 25 mm. 55. mm. 55.5 no.5 05.5 55 mm. mm. mm. 55.5 Nm.5 55.5 25 oo._ oo.5 50.5 m5._ Nm.m 55.55 xm: oo._ oo.5 55.5 om.N om.m 55.o5 2m: 5555 coo: m 55.5 mN. N5. 05.5 Nm.5 om.w 55 55._ mm. 55. m5.5 m5.5 5m.m 25 mm. mm.5 5m. no.5 5o.m mm.m 55 No. 55. 5N. Nm. 55.5 5N.m :5 oo._ oo.5 5o. mN. 55.m 5N.5 xm: oo.5 00.5 N5. Nm. 55.5 55.w 2m: 5555 555555 N o5. mo. 55. em. 5o.m mm.m 55 mm. mm. mm. 55. N5.m om.m :5 mm. 55. mm. N_.— N5.m 55.5 55 55. Fm. mm. N5. 55.N 05.m 25 oo.p oo.5 55. mm.5 55.5 mm.m 5m: oo.p oo.5 55. 50.5 ©5.m mo.o 2m: 5555 cmgumm5 _ m < N 5 N F. 5555 xgommpmo m < - 55555 5 N m: 4mm < .mmE:po> munch :5 mmmcmgu mmmucmucwauu.mum 555 55555 55 55.5 55.5 55. 55.5 55.5 55.55 55 55.5 55. 55. 55. 55.5 55.5 55 55.5 55.5 55. 55.5 55.5 55.5 55 55.5 55.5 55.5 55.5 55.5 55.55 55 55.5 55.5 55. 55. 55.5 55.5 555 55.5 55.5 55. 55.5 55.5 55.55 555 5555555 55 5 5 5 5 5 5 5555 55555555 0UP»... < R I’. m: 4mm < 5555555555--.5-5 55555 118 55. 55.5 55. 55.5 55.5 55.5 55 55. 55. 55. 55. 55.5 55.5 55 55.5 55.5 55. 55.5 55.5 55.5 55 55.5 5555 55. 55.5 55.5 55.5 55 55.5 55.5 55. 55. 55.5 55.5 555 55.5 55.5 55. 55. 55.5 55.5 555 5555 55555 55 55.5 55. 55. 55.5 55.5 55.5 55 55. 55. 55. 55. 55.5 55.5 55 55.5 55. 55. 55.5 55.5 55.5 55 55. 55. 55. 55. 55.5 55.5 55 55.5 55.5 55. 55.5 55.5 55.5 555 55.5 55.5 55. 55. 55.5 55.5 555 555555555 55 55.5 55.5 55. 55.5 55.5 55.55 55 55.5 55.5 55.5 55.5 55.5 55.55 55 55.5 55.5 55. 55. 55.5 55.5 55 55.5 55. 55. 55.5 55.5 55.5 55 55.5 55.5 55. 55. 55.5 55.5 555 55.5 55.5 55. 55. 55.5 55.5 555 5555 55555 55 55. 55. 55. 55.5 55.5 55.5 55 55.5 55.5 55.5 55.5 55.5 55.5 55 55. 55. 55.5 55.5 55.5 55.55 55 55. 55. 55.5 55.5 55.5 55.5 55 55.5 55.5 55.5 55.5 55.5 55.55 555 55.5 55.5 55.5 55.5 55.5 55.5 555 5555 .55555 55 5 5 5 5 5 5 5555 55555555 55555 5 5 m: Aux < .555555555--.5-5 55555 119 55. N5. NN.5 5N.N no.5 55.x 55 ow. mu. o5. .55. 55.N 5o.m :5 5_._ N5.5 55.5 55.N mo.m em.o5 55 om. mo.5 mm. m5.5 mn.m mo.5 :5 oo._ oo._ m~.5 5m.~ mo.m mm.m 5m: oo.5 oo.5 no.5 N5.5 55.m 55.5 2m: 5555 mm m < N 5 N 5 5555 xgommumu m: 555 < 55555 < 5 .1555_5.5555--.5-5 555,: 120 Because of quantitative restrictions on textiles (5-7) and iron and steel (lO-ll), a change in the explicit tariff in these categories will have little impact on trade unless the quota is liberalized. Since explicit tariff revenues will be converted into implicit revenues to foreigners, explicit tariff reductions make little sense without accompanying quota liberalization. Section 5.5 addresses this issue. The most remarkable increase in U.S. imports occurs in Dyeing, Tanning and Coloring Materials (l8). The most remarkable increases in U.S. exports occur in Paper Manufactures (4), Clothing (7), Metal Manufactures (l4), and Miscellaneous Transport Equipment (31). The most responsive E.E.C. import categories appear to be Dyeing, Tanning and Coloring Materials (18) and Road Motor Vehicles (30). The only truly outstanding E.E.C. export categories are Shoes and Bags (33) and Toys (37), while the only outstanding Japanese import category is Hood Manufactures (3). The most responsive Japanese export categories include Clothing (7), Plastics (19), Road Motor Vehicles (30), and Miscellaneous Transport Equip- ment (31). There are also several surprises in these results. U.S. exports of Clothing (7), for example, expand proportionately more than imports under both Scheme A and Scheme B. This is primarily the result of the small U.S. share in the export market and the resulting high price elasticity of demand for U.S. exports. Exports also tend to expand proportionately more than imports in Organic. Chemicals (16) and Motor Vehicles (30). The controversy in recent 121 years regarding imports in these two categories makes this result particularly interesting. Absolute Changes By multiplying the percentage changes in trade by a base trade volume, the absolute change in trade is obtained. Table 5-4 presents these absolute changes for each category. As indicated earlier, 1974 is the base year for all value figures, and the changes reported in Table 5-4 are expressed in millions of 1974 U.S. dollars (f.o.b.). These changes represent the initial price effects of Scheme A and Scheme 8. In addition to the totals, the LDC constant market share of the changes is reported. This is useful for two reasons: First, it indicates the role of LDC's in the expansion of trade; and second, it provides a reference point for considering a non-constant market share of the change. The latter may be significant due to the unique role of LDC's in the multi-lateral negotiations. Again, the results for the textile (5-7) and iron and steel categories (10-11) are contingent upon liberalization of the relevant quotas. It is interesting to note, however, that only in textile category (7), Clothing, does the increase in imports exceed the increase in exports for the U.S. The greatest increases in U.S. imports under both Scheme A and Scheme B occur in Road Motor Vehicles (30); Iron and Stee1, Semi-Manufactures (11); Clothing (7); Petroleum Manufactures (15); Telecommunications Apparatus (29); Textile Semi-Manufactures (5); I. II-II'IIII' Ill! Ill" 11.! TABLE 5-4.-Absolute Changes in Trade Volumes (millions of dollars). LDC Total LDC Total LDC TotaTi LDC Total Prin .Categonx m m N!— Q' '3’ ‘0 (D USM USX EM EX JM JX 1 Leather mfgs m N P N com NN com LOQ' N m Q’LO VN N I’\ F P 0’) N on O O N F m 05 m LOLO NM moo Q'to Q"— I—LO USM USX 2 Rubber mfgs EM EX JM JX 122 [\O 23.7 1.3 26.0 45.6 4.1 3.7 2.1 7.2 13.5 MN 05¢ NKO Q’F- l\ 00¢ 05 Mr— Q’ ON VF- m.— mm 0 O O O O O Q!— USM USX EM EX JM JX 3 Wood mfgs 4 Paper mfgs 3.9 2.0 36.7 EM EX 1.0 1.1 16.3 15.4 8.3 30.2 Nd) RN 0 0 PP <7? MN cou- I—l\ TABLE 5-4.-Continued. Tota1 LDC Tota1 LDC Tota1 LDC Tota1 LDC Prin Category 5 Tex semi- mfgs 91. 178.8 EM EX JM JX 35. 102.7 17.0 134.6 33.1 38.0 171.7 260.4 74.1 332.2 [\ F [\ (V) F M O N 26.1 11.7 13.8 6.2 USM 6 Tex articles on O F 00 L0 <‘ N 05 O 8.6 20.7 4.6 24.3 13.0 38.6 USX 123 6.9 5.2 12.9 18.2 EM EX JM JX 22.2 8.9 16.6 N O P P F O m (I) m m m 03 CD 7 Clothing 109.8 m N 204.4 142.6 EM EX JM JX 22.5 37.6 67.9 [\ L0 15. 12.3 8.1 29.9 6.2 14.7 26.3 8.1 32.2 35.0 31.1 3.9 16.5 2.1 5.1 N 03 76.9 USM 8 Mineral mfgs 2.0 1.0 3.9 2.2 N Q 16.4 USX EM EX JM JX CD P <- m 01 oo O I\ MF- 0000 17. 0000 NO Q” TABLE 5-4.-Continued. Tota1 LDC Tota1 LDC Total LDC Total LDC Prin Category 4.5 8.5 1.4 2.5 13.5 25.5 2.6 USM 9 Glass mfgs 1.1 2.8 5.2 22.2 4.7 11.9 USX EM EX JM JX 1.8 12.5 N 0‘ 6.6 5.2 9.7 l!) m v 0 L0 l\ m r—d" m 03 LG (I) USM 10 1&5, O N F F USX EM EX JM JX unworked ON 0‘ L!) N <- 124 <- LO N RD P 1.5 1.3 3.5 8.7 270.2 63.6 2.5 18.7 1.3 17.1 536.4 r-CW mm OLD LON 23. 34. NM LON q 6 USM USX 123.2 semi-mfgs 11 1&5, (I) t\ O N m 60.3 182.8 EM EX JM JX N (\I d) N m 9 17.6 324.1 11.2 370.0 34.2 626.9 Nd’ moo com [\d’ 0—!— 0—!— mm Nw MN Q'O mm 0‘!— I—l\ mm .—m LON NO MN mm 33. 24. 3 4 USM USX EM EX 12 Aluminum LON 005 com #0 L000 JM JX -_._ TABLE 5-4.-C0ntinued. Total LDC Total LDC Total LDC Total LDC rm Category 3.1 2.0 17.7 15.6 7.3 47.9 30.0 95. 14. USM USX 13 Other metals 3.9 3.8 2.0 7.5 30.9 12.4 59.6 31.5 EM EX JM JX 6.3 1.5 15.9 O 0 MF- (13 O N (V) [\05 P0 63.2 163.3 122.0 14.3 109.2 312.3 USM USX 14 Metal mfgs 125 NO I—LO mm 16.1 189.0 EM EX l\ JM JX m 13.7 16.7 1.4 1.9 3.2 26.1 2.6 31.9 223.2 139.9 114.4 10.3 15.3 5.3 11.9 272.9 29. USM USX 15 Petrol mfgs 1.1 2.2 3.8 6.1 EM 39.7 23.1 20.4 EX JM JX 1.4 12.0 27.7 6.1 54.3 0 O O N ‘0 on L0 q. 05 Nm u—oo L0 (‘00 £00 NR r—r— 73.6 155.2 4 1 0N 11.9 105.9 7 141.4 298.5 84.2 180.5 USM USX EM EX l6 Org chem 52.7 193.3 JM JX LDC -.-—-_ .__.- .—-.- m 3.3 Total .7 1.9 1.8 LDC 3.5 4.7 6.5 1.6 Total 5.8 9.4 5.7 4.0 1.1 LDC 30.5 22.7 20.2 9.1 Total 11.5 18.3 11.0 7.9 LDC 60.5 39.1 18.0 Total Prin USM USX EM EX JM JX USM USX Category TABLE 5-4.-C0ntinued. l7 Inorg chem 18 DTC mat 126 mooo Cd" LOO‘ LDLD u—LD NV NN N l—f— mm 03F [\f— Mk0 Q'LO Q'N m l_l_ ON COO MN I'— N N r— r— N LOP-Q' Chm NLD F-m NN mm mm N f—‘mwF-Nm; '— (\I (‘0 PO MN LOT-D Ofl’ LOO [\N LOQ' LO NN 00¢ LDC” [\f— on L0 V t—N r-r- L0 00 NN (no NR r-m [\m Low LOO l\ V r—r— NO 0‘05 (*3 0—K l— F-N N 1" LOO NF- LOLD LON mm O‘LO #0 mm 000 [\L0 Or— Om CO r-w WV 0‘!— I'— Q'LD mm mm I'— r- ON Q'r- 01 OLD LON d’d’ ON N05 Nd’ w NOW QC Nil- r—LD r-LO r—v— r- N” Q' r-LO '— F m cm ON 0‘? Q'N (no ON Chd' OQ' Vt— com LOG mm ”LO Q” ”N NM l-‘F— No— LOO 00¢ '— N r- r— r- N XX XX XX XX (DU) 2X z>< (DU) 2X 2x LLJLIJ '7') D: LLILIJ '3'? =2 LIJLLJ '7'? q. S. m G) U Q. 0r— 4.) a U) m (U r— r— or- D. O 03 O I— N TABLE 5—4.-Continued. Total LDC Total LDC Total LDC Total LDC Pr1n Category ”V mm NV 6’ MN (Or- mm 0000 MV m f_l'_'-' LOG) LON (DR Pd) NO NN USN mm LOLO F- Nm Nd’ CDLO LON QR u—P' Q'LO Q'm Ln [\LO om ”N ON LO <' 9 45.5 USM USX EM EX JM JX 21 Other chem 17.0 v—m LOCI) mm mm: l—‘N 96.6 170.4 USM 22 Power mach USX EM EX 127 LOLO LOLO 81.5 126.9 2.7 3.7 8.6 16.6 20.1 .1 32.5 27.7 JM JX 1.6 3.1 5.2 54.2 USM USX EM EX JM JX 23 Ag mach 00¢ r-N mo— LOLO Q'LD Nfi’ 92.3 19. 26. 176.1 USM USX 24 Office mach LOF- d’l‘ 015 mm 7.1 72.6 18.3 36.3 140.4 67.6 EM EX TABLE 5-4.-Continued. Tota1 LDC Tota1 LDC Total LDC Total LDC r1n Category LO LO Q. LO 00 Q- P USM 9 Glass mfgs SO (I) N N m LO N USX EM EX JM JX 12.5 6.6 5.2 9.7 O LO l\ m r—Q‘ m 0‘ L0 00 USM 10 I&S, O N F F USX EM EX JM JX unworked . 2.2 2.6 4.2 11.4 124 1.6 6.9 5.0 18.7 1.5 1.3 3.5 9.6 8.7 270.2 1.3 2.5 45.7 17.1 536.4 F m O LO USM 11 I&S. 0‘ m LO l\ 123.2 67.3 63.6 USX EM semi-mfgs 60.3 182.8 EX JM JX 11.2 17.6 324.1 34.2 626.9 191.3 370.0 NQ' moo (”LON 1.4 COOLO 1.5 Q'ON COMM 2.8 33. 24. USM USX EM EX 12 Aluminum LON mm N'— 00‘ ”N #0 LOW JM JX TABLE 5-4.-Continued. Tota1 LDC Tota1 LDC Tota1 LDC Tota1 LDC Pr1n Category 3.1 USM USX 13 Other meta1s 30.9 30.0 15.6 7.5 3.9 3.8 2.0 59.6 EM EX 31.5 12.4 15.9 6.3 1.5 JM JX O 0’) 00 O O P l\ m 125 N03 l—Ch NR 14. 6 N00 l—KD w ,— NO [\N t—LO N N 63.2 5 5 163.3 8 8 14.3 109.2 NLO N LO 10 122.0 312.3 16.1 189.0 3 3 USM USX EM EX JM JX 14 Meta1 mfgs 272.9 223.2 139.9 114.4 31.9 26.1 16.7 13.7 USM 15 Petro1 mfgs 1.4 1.9 3.2 USX 29.6 10.3 15.3 5.3 2.6 EM 11.9 EX JM JX 1.1 2.2 3.8 6.1 23.1 20.4 39.7 1.4 12.0 27.7 6.1 54.3 C 00 NC” Fm LO LOLO 73.6 5 155.2 4 ‘DN 11.9 105.9 7 141.4 298.5 84.2 180.5 USM USX EM EX 16 Org chem 5.0 15.5 1 12.3 LOO 00 52.7 193.3 JM JX TABLE 5-4.-Continued. Tota1 LDC Tota1 LDC Tota1 LDC Tota1 LDC Prin Category 1.7 2.4 3.3 3.5 4.7 5.8 9.4 60.5 11.5 30.5 USM USX EM EX JM JX 17 Inorg chem 1.0 1.9 1.8 18.3 22.7 44.3 20.2 5.7 6.5 4.0 11.0 39.1 1.6 9.1 7.9 18.0 O 0 CD <3' (‘0 LO O (V) LO 01 N LO N O LOVN Q'MCS OF“!— mLO Q' “'05 l\ Cir- m r—MF- FNI— 126 N (V) or) LO (*3 I\ NC LOO QR ow l—'l-" r—V ON NV NV ,._ I—LO mm LOCO OQ' Q'r— NM 2x LOU) EX 2: UJLIJ 4.: (U E U *— CD CD I— 78.6 115.7 . 2.0 12.8 3.9 40.6 11.8 13.3 21.0 61.5 22.9 USM USX EM 19 P1astics 4.4 8.3 24.2 39.5 NLD F-N QC 6' 56.9 104.4 EX JM JX LOLO r—LO NF- (ON 0103 I\ 20.5 130.3 17.6 151.2 39.4 248.7 COLD coco (OI—N #0 00¢ 050 Pm Q'Q’ P0 USM USX EM EX 20 0115, perf NN mm 0100 OF- NOS 0 0 JM JX l1 7. :97 I. ill-Ir: TABLE 5-4.-Continued. Tota1 LDC Tota1 LDC Tota1 LDC Tota1 LDC Prin Category md' mm Nd’ V MN (‘01—- mm oooo NV 0 O O O C O m P FF LOG) LON ”N F-(n NO NN N CON ”LO LOU) '— Nm Nd‘ N . 00m LON ON F-F- Q'LO Gm LO [\LO OM LOO cox on mix LO '— 2x WU) EX 2X :3 LLJLIJ "JO E Q) .C U S. Q) .C 4.: O P N r—oo LOCO mm COLD r-N Nd’ <' 96.6 170.4 USM USX 22 Power mach 127 LOLO LOLO 81.5 126.9 EM EX 2.7 3.7 8.6 27.7 20.1 JM JX 1.6 3.1 5.2 16.6 32.5 54.2 OM ON OLD Na) um (DO USM USX EM EX JM JX 23 Ag mach NV r—N Or— LOLO Q'LD Nfl' 92.3 176.1 USM USX 24 Office mach 7.7 EM 140.4 7.1 72.6 3.7 17.6 18.3 EX JM JX 2.1 3.9 14.4 9.8 36.3 67.6 Q'O ON LOO mm OLD mm LDC Tota1 3.8 2.3 LDC 2.3 2.4 1.5 8.5 1.5 4.0 Tota1 LDC 2.7 12.0 8.3 7.6 8.8 7.0 27.2 5.7 14.1 Tota1 LDC 5.2 1.1 23.3 15.9 14.6 17.0 13.5 52.6 11.0 27.1 24.1 15.5 Tota1 Prin USM USX EM EX JM JX USM USX Category TABLE 5-4.-C0ntinued. 25 Meta1 mach 26 Tex mach 128 OQ‘ OO Nr— v—-|'\ F-f\ r—r- LOLO LON N m N m O LOO m LC) '— 00 V00 NLO OLO COO NP— Nl— LOF- ON d‘ LO (I'N ON r—O I-I-Q' LOr- M!— V LO r— r—N l—' '— Ol\ OM OLO F-Q' NLO LOLO OO NO <3" N N Ln «a mu r—O .—.— O LG I- v- r— r- I\L.D VB PO Q'N Q'LO r-U', NLO OO O O OO LON 00¢ F-LO OI— LOI— l—‘ N V l—N N F NO r—O om LOI— LON NO LOO OO .LD O r-O !— N NLO d’v— OLD r— N O N LO OLO O O !— N Q'LO OO LOO F'N OF- LOI— NN OO 06:5 .-o .—<2- mm 00 .-oo ox.— on m d’ ON NF- t—O OQ' Q'LO Nf‘ m d" t— '— Q'LO r—N OLD NO om em OLO LOLO N <' v—N F-LD I—O F-N OO [\f- m LO LO (\J N [\O LO 5 N 6’ r—- F" LOF- LOF- OLD Q'LO F-LO LOO LOQ' OLO LOO MO OO ON u—r- LOQ’ LON Or— LO l\ LON LOO NLO LOO OF- me’ xx xx Z>< 2X LOU) EX 2x mm 2x 2x LLILIJ O") O: LLJLLJ O") OO LLJLIJ OO .C .C U U M (U E E ‘- H G) U .C Q) +-’ v- 0 LL! N O N N 2.5 6.7 LDC 11.3 Tota1 LDC 4.7 Tota1 LDC 7 . 28.2 Tota1 122.4 400.0 361.5 LDC 106.0 21.6 5.2 121.9 15 Tota1 234.9 53.7 108.2 121.4 11.9 316.8 786.2 702.5 Prin USM USX EM EX JM JX USM USX EM EX Category TABLE 5-4.-Continued. 29 Te1ecom 30 Motor veh 129 4.1 19.8 10.6 1.2 7.9 82.2 2 37.5 158.7 62.0 20.2 156.0 312.0 2.2 2.4 45.2 10.5 9.9 1.6 4.4 “"31 22.8 4.7 87.5 19.7 19.3 .2 160.2 10.0 2.8 14.7 14.4 613.5 35.1 65.1 .5 310.8 5.4 28.5 41.7 27.8 1190.4 46. 87. 66.2 125.8 68.1 119.8 24.4 90.7 JM JX USM USX EM EX JM JX USM USX EM EX JM JX instr 31 Misc trans 32 Prec LDC r-:-- N l—‘l— Tota1 r-N OLO Q'LOr— LDC 1.4 27. 4.1 Tota1 8.1 3.7 LDC 14.8 47.7 5.9 2.5 Tota1 .9 LDC 15.7 15.1 6.8 10.9 4.7 16. 24. Tota1 Prin USM USX EM EX JM JX USM USX Category TABLE 5-4.-Continued. 33 Shoes, bags 34 Photo mfgs 130 Vt— MN 1 1 3.1 Q'N Nd" LON LOLO (OLD LOO OWLO NI— mm Q’F NN N [\m LOU) Q’ P COLD FLO Q'm moo LON Od’ 03F. LON LO N r-r- LO LON r-LO '— ,_. OQ' r—LO NN OSLO NLO LOI— P05 LON MLO LON OLD F-N l'-'f— Nl— r—O‘ NO f—l'_ '— m:—- N!— LO LDm PO 0000 LON NV LON LDC" r—N O .CI: NW) N r—r- O‘Q’ mm P!— l—' r—- N r—w N ON 000 MF- 06' OLD (DI- LOO r—.Q' LOF- mm NM M!— P” Q“) NN r—r— u—r— m LDN V“) 0‘ EX 2x Z>< Ex LOLO 22x 22x mm 2x Ex LIJLIJ O") :32) LLILLJ O") OO LLJLIJ O"? m OJ 0‘! 5- H- : E 4.: .,_ 'U C C $- 3 3 0 LL. (I) LO LD (*3 m TABLE 5-4.-Continued. Tota1 LDC Tota1 LDC Tota1 LDC Tota1 LDC rm Category LOO Mr— md' LOQ' LOO‘ LOF- [\m NCX) USM USX 37 Toys 25.7 11.9 13.6 6.3 6.3 2.9 3.3 1.5 8.3 4.4 EM EX 1.5 3.0 17.0 5.6 32.0 131 MLO v-LO LOG r—m JM JX Fart- - ... ..h 132 Organic Chemicals (16); Metal Manufactures (14); and Electrical Machinery (28). The greatest increases in exports, on the other hand, are in Road Motor Vehicles (30); Other Machinery (27); Metal Manufactures (14); Organic Chemicals (16); Electrical Machinery (28); Paper Manufactures (4); Office Machinery (24); Textile Semi— Manufactures (5); and Power Machinery (22). There is clearly intra- industry specialization in some categories, since some industries excel in both imports and exports. E.E.C. imports increase most substantially in Clothing (7); Textile Semi-Manufactures (5); Road Motor Vehicles (30); Office Machinery (24); Telecommunications Apparatus (29); Electrical Machinery (28); Power Machinery (22); Precision Instruments (32); and Plastics (19). E.E.C. categories exhibiting extraordinary export expansion include Other Machinery (27); Metal Manufactures (14); Iron and Steel, Semi-Manufactures (11); Organic Chemicals (16); Road Motor Vehicles (3); Precision Instruments (32); Telecommuni- cations Apparatus (29); and Textile Semi-Manufactures (5). The greatest increases in Japanese imports tend to be in Wood Manufactures (3); Textile Semi-Manufactures (5); Petroleum Adanufactures (15); Office Machinery (24); Clothing (7); Other Chemicals (21); Organic Chemicals (16); and Plastics (19). In- creased Japanese exports of Road Motor Vehicles (30) outstrip increases in all other categories by far under Scheme A and by a smaller margin under Scheme B. Other categories experiencing extraordinary export growth are Iron and Steel, Semi-Manufactures (11); Textile Semi—Manufactures (5); Telecommunications Apparatus 133 (29); Other Machinery (27); Plastics (19); Metal Manufactures (14); Organic Chemicals (16); and Miscellaneous Transport Equipment (31). The general pattern of these results for the three princi- pals is consistent with what one might have expected beforehand. The ranking of Road Motor Vehicles, Iron and Steel, and Clothing in U.S. imports, for example, will not surprise those who have lobbied intensively in recent years for greater protection in these industries. What is something of a surprise is the roughly equi- valent expansion of exports in the Road Motor Vehicle category. m‘--‘ 0' D. This is due in part to the relatively high price elasticity of demand for U.S. exports (Table 4-30). Aggregate Changes The aggregate effects of Schemes A and B have been computed by summing the effects of the individual categories. Table 5-5 contains these effects. The total effects are subdivided, however, into textiles (5-7), iron and steel (lO-ll), and the other categories as a whole to separate the impact of the quota-affected categories. The total effects for Scheme A suggest that imports in the thirty-seven categories will increase between 3.5 and 6.9 per cent for the U.S., between 2.8 and 5.4 per cent for the E.E.C., and between 3.0 and 5.8 per cent for Japan. The respective mid-range estimates are 5.2, 4.1, and 4.4. Alternatively, the total effects suggest that exports will increase between 3.8 and 7.3 per cent for the U.S., between 2.4 and 4.7 per cent for the E.E.C., and between 3.0 and 5.8 per cent for Japan. The respective mid-range estimates for exports are 5.6, 3.6, and 4.4. 134 ii. ..iii NmF mmm omN new mom «emu vowP mmme xa __ Rm Ne mo_ Pup owe mmm mom :6 Ne? NNm NNN o_m “mm mm~_ “mm, muqm xm em mo, _op «mm emm mmo mom mmfifi 2m mo_ mmm «om one mom ompm mNm_ “Npe xm: ¢e_ mam mom NNN 4mm meow mmop mmmm 2m: PQOOF om «mm exp mmq mmo _mop mow_ momm xv NP mo NN mm? mm Rem LN. “we :6 _N_ New «mm o_m Rm“ mmm_ owe. Mkom xm om ¢__ mm ONN MNP _mo 0mm o_mp 2m mm 0mm oo_ 44m moo oFa_ mNFF mofim xm: mm o_N LOP Noe mom mmmp mo“ “mam 2m: Lacuo mm we we a“ mm, mmm mxm eew xa _ m N m m LN mp mm 2a N N, _P mm me “a mm mm, xw _ m P m e Rm m NR 2m 4 N m «F mm mm mm mm_ xm: m Rm 0 PR om mNN Fm mmm 2m: _aapm a eocH Km mm ON ___ Pm_ CNN mmm owe xq op m_ m_ mm we 0N mm me_ sq m_ me RN RN _¢ m__ 0N N_N xm mm 0m «0 mm ~N_ mom mmm “Fe 2m a, we mm m“ mm mm_ om_ mam xm: om om_ mm, eew mop com mom qu 2m: mw__3xap gas _ap0h gas _wpop was .mHOH 894 _aHOH =_La scammgmu .AmLmPPOU $0 meow—.PwEv me:_.o> mbmgh cw mmmcwzu mu3POmn< 0:“ ....o hLmEEzmil.mlm “4m”: 135 The Scheme B effects represent only one "round" of reduc- tions by the height of the individual tariffs. To achieve the same aggregate import increase as Scheme A, the Scheme 8 reduc- tions would have to be multiplied by approximately 5.5 for the U.S., 5.6 for the E.E.C., and 5.2 for Japan. If textiles and steel are excluded, however, the factor increases to 7.3 for the U.S. and 6.0 for the E.E.C., but decreases to 5.1 for Japan. This reflects the height of the textile and iron and steel tariffs relative to other tariffs on manufactures. To achieve the same aggregate increase in exports as Scheme A, the Scheme 8 reductions would have to be multiplied by 6.5 for the U.S., 5.6 for the E.E.C, and 6.7 for Japan. Excluding the quota-affected categories, the factors are 6.8 for the U.S., 6.0 for the E.E.C., and 7.2 for Japan. Clearly, reductions of these magnitudes would require modification in many categories of the U.S. statutory limit to reductions of sixty per cent. Although the factors above would put the aggregate Scheme A and Scheme B effects on the same scale, the composition of the effects would be far different. At first inspection, the total figures for both Scheme A and Scheme 8 might suggest that the E.E.C. and Japan are likely to experience a significant trend toward an improved trade balance (upward pressure on the value of their currencies). This is not likely, however, because the relative expansion of total imports and exports in each case is in rough accordance with their original proportions in the thirty-seven categories. In other words, the implied trade balance effects will largely be mitigated by an 136 expansion in imports in categories other than these thirty-seven (e.g., petroleum and other raw materials). Suppliers in these other categories, therefore, should expect an expansion in the demand for their commodities roughly in proportion to the general expansion in trade for the principals. To the extent that this is not true, however, feedback price and income effects will tend to ameliorate the imbalance, and the final figures must be adjusted accordingly. The United i- States, for example, may experience a very slight positive trade balance effect under Scheme A, and the E.E.C. may experience a slight negative trade balance effect under both Scheme A and Scheme B. There is little indication that the Japanese trade balance will tend to move in either direction. An index of the import- or export-bias of Scheme 8 relative to Scheme A can be constructed at the aggregate level. This is the same index used at the category level in Table 5-4. For the United States this index is .85, indicating as expected the relative import-bias of Scheme 8 for the U.S. If the quota- affected categories are excluded, however, the index is 1.07, indicating that Scheme B has a relative export-bias for the remaining categories. Interestingly enough, this export-bias is due in large part to an expansion in exports of Road Motor Vehicles (30). The total index for the E.E.C. is .98, and the excluding textiles and steel is 1.00. These figures mean that Scheme A and Scheme B have little import- or export-bias relative to each other. The Japanese total index, however, is .78, and the 137 index excluding the quota-affected categories is .70. Scheme 8, therefore, is strongly import-biased relative to Scheme A for Japan. This result means that relatively high tariffs occur more frequently among the individual tariffs levied by Japan than among the tariffs levied by its major trading partners. Looking at this issue from the LDC perspective, one finds that the total index is 1.45, and the other index is 1.05. Hence, the export-bias of Scheme B for LDC's is primarily the result of the relatively greater tariff reductions on textiles implicit in ..h—f _ .1; ‘4 Scheme 8. This potential export4bias is constrained, however, by existing restrictions on textile imports. 5.4 Welfare Changes The generally low level of tariffs among the principals means that deadweight loss (gain) effects will be small and dominated by even the slightest terms of trade effect. The computation of these figures, however, does serve an important purpose in identi- fying those import-competing and export industries associated with the greatest welfare changes. Such information can be an important element in considering alternative trade policies affecting a particular industry. The ranking of the welfare effect in a parti- cular category is probably more significant than the actual numerical magnitude. The computation of these welfare effects was discussed in Chapters 2 and 3. As an indication of the general magnitude of welfare changes in each industry, estimates for Scheme A are 138 presented in Table 5-6. These are mid-range estimates, based on the assumption that export supply is twice as price-elastic as import demand. Obviously, this assumption may be more appropriate in some instances than in others, but it does offer approximate points of comparison. The calculation of the deadweight loss effect (DWL) requires that each individual price change be squared. As an example, this means that the Scheme A computation formula for the import OWL where k is the import share of the tariff change.4 The first term in parentheses is easily computed using the DI(B) adjustment factor from Table 5-1. The second term, which refers to tariffs five per cent or below, is approximated by the square of .025. This results in little error, however, because the tariffs are so small. The NR column represents the net loss or gain of tariff revenues to the country as a whole. For imports this equals the revenues from the new tariff times the change in trade less the loss in revenues to foreign suppliers through the terms of trade effort. For exports this equals the transfer of previously collected tariff revenues to the exporting countries. The sum of the DWL and NR columns equals the total welfare effect. 4Theoretically, the k is also a function of the individual tariff rates, but this is ignored since k is imposed by arbitrary assumption. 139 TABLE 5-6.--Mid-Range Estimates of Scheme A Welfare Effects (millions of dollars). Category Prin DWL NR Total 1 Leather mfgs USM .5 - 4.7 - 4.2 USX .l 2.6 2.7 EM .7 - 8.7 - 8.0 EX .2 5.2 5.4 JM .1 - 1.4 - l.3 JX .0 1.4 1.4 2 Rubber mfgs USM .6 - 5.6 - 5.0 USX .2 7.6 7.8 EM .1 - 3.0 - 2.9 EX .7 17.2 17.9 JM .1 -- .6 - .5 JX .3 7.9 8.2 3 Wood mfgs USM 1.5 - 5.2 - 3.7 USX .3 5.5 5.8 EM .8 - 5.2 - 4.4 EX .1 2.5 2.6 JM 1.3 - 4.7 - 3.4 JX .1 1.6 1.7 4 Paper mfgs USM .3 - 4.5 - 4.2 USX 1.3 15.9 17.2 EM .7 -10.4 - 9.7 EX .1 6.9 7.0 JM .3 - 2.5 - 2.2 JX .l 3.8 3.9 5 Tex semi-mfgs USM 7.3 -l8.6 -11.3 USX 1.2 27.7 28.9 EM 2 7 -25.6 -22.9 EX 1 1 39.4 40.5 JM 1.2 -1l.4 -lO.2 JX 3.1 36.2 39.2 . Total NR -10.1 17.2 DWL 140 Prin usm USX EM EX JM JX Category 6 Tex articles TABLE 5-6.--Continued. 0.. mflwmflmwumx 36 0 70 4 2] 2 - u 76 0 99 0 3 3 . _ 40 0 21! 6 .II MX 55 M UU E g n .1- h t 0 ul. C 7 24.8 23.9 EX JM JX USM USX EM EX JM JX USM USX EM EX JM JX USM USX EM EX JM JX 1&5, unworked 8 Mineral mfgs 9 Glass mfgs 10 Total -31.6 26.2 NR -38.7 25.7 DWL 6.7 141 Prin USM USX EM EX JM JX Category 1&5, semi-mfgs TABLE 5-6.--Continued. 11 61' 25 USM USX 12 Aluminum 13 44 42 44 EM EX JM JX 1.2 62 52 62 USM USX 13 Other metals EM EX 72 32 92 32 JM JX -17.6 24.1 -15.6 23.9 -20.2 22.4 23.7 -16.1 USM USX EM EX JM JX USM USX EM EX JM JX 14 Metal mfgs 15 Petrol mfgs Total -lO.8 27.4 NR -13.4 25.9 DWL 142 Prin USM USX Category TABLE 5-6.--Continued. 16 Org chem -12 9 32.0 -14.7 31 0 EM EX JM JX 7] 37 19 46 USM USX Inorg chem 17 .. 07 67 77 67 EM EX 92 12 02 22 JM JX USM USX l8 DTC mat 12. 15 12 42 EM EX 22 12 4] 112 JM JX -10.6 35.7 -12.1 35.1 .5 USM USX EM EX JM JX 19 Plastics 20 1|4 39 1'3 USM USX 20 Oils, perf 39 26 47 26 EM EX JM JX Total -l0.0 35.7 NR -10.7 35.2 DNL 143 Prin USM USX EM EX Category Other chem TABLE 5-6.--Continued. 21 35 63 04 73 JM JX \V -10.2 22.5 -10.7 21.9 USM USX EM EX 22 Power mach 63 25 9] 25 32 JM JX USM USX 23 Ag mach 45 13 45 «'3 EM EX JM JX 83 53 03 72 20 1] USM USX 24 Office mach -11.4 15.1 -13.2 14.6 .8 EM EX 70 58 58 77 JM OX 42 26 6] 26 USM USX 25 Metal mach EM EX 29 13 38 1|3 JM JX Total NR DNL 144 Prin Category TABLE 5-6.--Continued. 89 55 28 65 USM USX 26 Tex mach EM EX 30 1'0 36 «'9 JM JX -19.1 92.2 -19.1 97.9 -20.0 88.9 -19.7 93.8 3.3 4.1 USM USX EM EX 27 Other mach 96 96 -10 2 25.4 .3 2 1. JM JX -26.5 72.8 -21.6 62.2 -28.5 71.3 -22.6 61.5 .0 USM USX EM EX 28 Elect mach 32 73 95 72 67 JM JX -16.2 19.9 -19.5 19.6 3.3 USM USX 29 Telecom 31l- 83 -10 3 22.3 2.0 8 EM EX -67.9 61.8 - 5. 106. -72.8 59.3 -10.3 105.0 JM JX USM USX EM EX JM JX 30 Motor veh 145 TABLE 5-6.--Continued. Category Prin DNL NR Total 31 Misc trans USM .8 - 3.5 - 2.7 USX .7 7.6 8.3 EM l - .5 - .4 EX 3 11.0 11.3 JM .0 - .2 - .2 JX 1.0 7.5 8.5 32 Prec instr USM 1.8 -19.8 ~18.0 USX .6 23.2 23.8 EM 1.0 -12.4 -1l.4 EX .8 26.0 26.8 JM .5 - 5.4 - 4.9 JX .6 23.7 33.3 33 Shoes, bags USM 2.5 -16.5 -14.0 USX .1 1.2 1.3 EM .7 - 5.1 - 4.3 EX .9 12.5 13.4 JM .4 - 4 - 2.0 ‘ JX .0 1 4 1.4 34 Photo mfgs USM .1 - 1.7 - 1.6 USX .l 6.5 6.6 EM .3 - 2.7 - 2.4 EX .1 5.9 6.0 JM .4 - 1.9 - 1.5 JX .l 1.5 1.6 35 Furniture USM .2 - 4.6 - 4.4 USX .1 2.4 2.5 EM - l.3 - 1.2 EX 3 6.7 7.0 JM .0 - .9 - .9 JX .0 .6 .6 Mr . A l ' ... TABLE 5-6.--Continued. 146 Category Prin DNL NR Total 36 Sound mfgs USM .7 - 7.2 - 6.5 USX .1 5.9 6.0 EM .6 - 5.9 - 5.3 EX .2 5.0 5.2 JM .1 - 1.0 - .9 OX 1.4 12.7 14.1 37 Toy mfgs USM 1.0 - 9.6 - 8.6 USX .2 3.8 4.0 EM .6 - 5.2 - 4.6 EX .3 5.4 5.7 JM .1 - 1.9 - 1.8 OX .2 4.5 4.7 147 On the import side, the greatest DWL effects for the U.S. occur in Textile Semi-Manufactures (5); Clothing (7); Iron and Steel, Semi-Manufactures (11); Petroleum Manufactures (15); Tele- communications Apparatus (29); and Road Motor Vehicles (30). On the export side they occur in Metal Manufactures (14); Organic Chemicals (16); Other Machinery (27); and Road Motor Vehicles (30). The net effect in each industry can be calculated by summing the total effect for imports and exports. Those U.S. industries associated with the greatest net increase in welfare are Textile Semi-Manufactures (5); Organic Chemicals (l6); Plastics (19); Other Chemicals (21); Office Machinery (24); Other Machinery (27); and Electrical Machinery (28). Those associated with the greatest decrease in welfare are Clothing (7); Iron and Steel, Semi-Manufactures (ll); Petroleum Manufactures (15); Road Motor Vehicles (30); Shoes and Bags (33); and Toys (37). For imports, the greatest DNL effects for the E.E.C. occur in Textile Semi-Manufactures (5); Clothing (7); Organic Chemicals (l6); Plastics (19); Office Machinery (24); Telecommunications Apparatus (29); and Road Motor Vehicles (30). For exports, the greatest DNL effects are in Textile Semi-Manufactures (5); Metal Manufactures (14); Organic Chemicals (16); Other Machinery (27); and Road Motor Vehicles (30). The most interesting feature of these two lists is their similarity. E.E.C. categories associated with the greatest net decreases in welfare are Leather Manufactures (1); Paper Manufactures (4); Iron and Steel, Unworked (10); Other Metals (13); and Petroleum 148 Manufactures (15). The greatest net increases are in Iron and Steel, Semi-Manufactures (11); Metal Manufactures (l4); Plastics (19); Other Chemicals (21); Other Machinery (27); Electrical Machinery (28); Road Motor Vehicles (30); Precision Instruments (32); and Sound Manufactures (36). For Japan the greatest DWL effects on the import side are in Wood Manufactures (3); Textile Semi-Manufactures (5); Clothing (7); Petroleum Manufactures (15); and Office Machinery (24). On the export side the most substantial changes are in Textile Semi- Manufactures (5); Iron and Steel, Semi—Manufactures (ll); Plastics (19); Telecommunications (29); Road Motor Vehicles (30); and Sound Manufactures (36). Japanese categories associated with the greatest net decreases in welfare are Wood Manufactures (3); Aluminum (12); Petroleum Manufactures (15); Oils, Perfumes (20); and Other Chemicals (21). Alternatively, the most substantial increases are in Textile Semi-Manufactures (5); Iron and Steel, Semi- Manufactures (11); Other Machinery (27); Electrical Machinery (28); Telecommunications Apparatus (29); Road Motor Vehicles (30); Precision Instruments (32); and Sound Manufactures (36). Given tariff reductions of the same average scale for Schemes A and B, the values in Table 5—6 will generally be lower bounds for the Scheme B welfare effects. This is because the Scheme B OWL formulas involve the percentage change in price raised to the fourth power. This formula for imports is: 149 1:k k2 (Zwit 4 1 2 ) i where k1 is the import share of the tariff change and k2 is the scale factor required to equate the average tariff reductions of 5 The Scheme B OWL effects are greater in magnitude Schemes A and B. because high tariffs are reduced more than low tariffs. Because the OWL cost of protection rises with the square of a tariff, reductions based on the height of the tariff will reduce this cost more than across-the-board reductions. Table 5-7 presents a summary of the detailed effects in Table 5-6. Textiles (5-7) and iron and steel (10-11) are again listed separately. The total OWL effect is 88.9 million dollars for the U.S., 54.9 million for the E.E.C., and 38.6 million for Japan. The OWL for textiles accounts for about a third of the total OWL effects for the U.S. and the E.E.C. This proportion would be even greater using the Scheme B OWL formula. The large positive NR effect for the E.E.C. and Japan can be misleading. The general equilibrium aspect of these changes cannot be ignored. As indicated in Section 5.3, E.E.C. and Japanese imports in categories other than the thirty-seven con- sidered here will increase substantially. This means that the large positive NR effects for the E.E.C. and Japan will largely be dissipated in the form of higher payments for increased imports 5The average tariff change is leitg. Each tariff change is multiplied by the scale factor kg, and k2t2 is then squared as part of the basic OWL computation formula. 150 TABLE 5-7.--Summary of Mid-Range Estimates of Scheme A Welfare Effects (millions of dollars) Category Prin} OWL NR Total Textiles USM 20.3 - 66.3 - 46.0 USX 2.5 49.7 52.2 EM 9.5 - 65.7 - 56.2 EX 2.4 80.5 82.9 JM 2.8 - 23.4 - 20.6 JX 4.1 50.1 54.2 Iron & Steel USM 6.8 - 40.1 - 33.3 USX .6 26.9 27.5 EM 1.0 - 11.8 - 10.8 EX .8 73.1 81.9 JM .7 - 5.0 - 4.3 JX 2.7 92.3 95.0 Other USM 40.4 -378.6 -338.2 USX 18.4 532.4 550.8 EM 23.5 -225.6 -202.1 EX 17.7 673.9 691.6 JM 11.6 -122.6 -111.0 JX 16.7 336.6 353.3 Total USM 67.5 -485.0 -4l7.5 . USX 21.4 608.2 629.6 EM 34.0 -303.1 -269.1 EX 20.9 827.5 848.4 JM 15.1 -151.0 —135.9 JX 23.5 479.0 502.5 i “‘3‘“ 'h 151 of raw materials and other commodities. It is true, however, that these effects hold for the thirty-seven categories in this study. A final point is that the total effects for the U.S. in textiles and iron and steel are minimal. Moreover, the E.E.C. makes substantial gains in the textile categories. This, of course, ignores the effects of quantitative restrictions. 5.5 Employment Effects in the U.S. "1.1 Changes in employment are a primary concern in evaluating H the effects of trade liberalization. Based on the procedures . developed in Chapters 2 and 3, this section estimates the direct employment changes in the U.S. for each category. Table 5-8 presents these estimates for Scheme A and Scheme B. This table also includes the labor-output coefficient used in deriving the results. This coefficient expresses the number of "forty-hour- equivalent" workers per million dollars of 1974 output. The esti- mates themselves are mid-range estimates in two senses: First, the change in trade was computed on the assumption that export supply is twice as elastic as import demand; and second, changes in employment were derived on the assumption that production changes will account for three-fourths of the change in trade and consump- tion changes for one-fourth. The greatest net losses in employment occur in Wood Manu- factures (3); Clothing (7); Iron and Steel, Semi-Manufactures (ll); Petroleum Manufactures (15); Telecommunications Apparatus (29); Road Motor Vehicles (30); Shoes and Bags (33); and Sound Manufactures 152 TABLE 5-8.--Mid-Range Estimates of the Employment Effects in the United States. Jobs A Jobs Per Category Mill $ M/X A B 1 Leather mfgs 38.5 M - 602 - 97 X 353 55 2 Rubber mfgs 23.1 M - 684 - 64 x 523 89 3 Wood mfgs 25.2 M -1057 - 256 x 673 109 4 Paper mfgs 16.3 M - 138 - 19 X 2356 359 5 Tex semi-mfgs 28.8 M -2505 -l4ll X 2647 561 6 Tex articles 19.7 M - 295 - 79 X 438 123 7 Clothing 37.8 M -6560 -3371 X 2011 713 8 Mineral mfgs 24.3 M -107l - 434 X 228 54 9 Glass mfgs 28.5 M - 417 - 139 -"* X 365 86 10 I&S, unworked 29.2 M - 276 - 14 X 35 3 11 1&3, semi-mfgs 12.8 M -3872 - 513 X 897 101 12 Aluminum 12.4 M - 238 - 24 X 174 21 13 Other metals 10.2 M - 549 - 19 X 83 4 14 Metal mfgs 23.8 M -1653 - 389 X 4245 737 15 Petrol mfgs 5.3 M - 820 - 97 X 90 8 TABLE 5-8.--Continued. Jobs A Jobs Per Category Mill $ M/X A B 16 Org chem 7.3 M - 589 - 113 X 1242 199 17 Inorg chem 12.5 M - 427 - 25 X 314 33 18 DTC mat 13.5 M - 351 '- 151 X 105 17 19 Plastics 19.1 M - 854 - 144 X 1269 265 20 Oils, perf 9.5 M - 20 - 3 X 90 17 21 Other chem 14.8 M - 73 - 13 X 486 73 22 Power mach 18.6 M -1018 - 106 X 1799 177 23 Ag mach 19.3 M - 5 - 2 X 253 10 24 Office mach 21.4 M -1119 - 137 X 2148 379 25 Metal mach 30.6 M - 255 — 40 X 296 41 26 Tex mach 34.3 M - 471 - 75 X 302 39 27 Other mach 25.2 M - 873 - 132 X 9005 1166 28 Elect mach 25.1 M -2240 - 304 X 4070 665 29 Telecom 28.5 M —3819 - 557 X 875 164 30 Motor veh 11.7 M -5205 - 218 X 4669 507 154 TABLE 5-8.--Continued. Jobs A Jobs Per Category Mill $ M/X A B 31 Misc trans 17.7 M - 466 - 74 X 892 223 32 Prec instr 31.2 M -1185 - 353 X 2234 342 33 Shoes, bags 43.3 M -2613 - 693 X 221 65 34 Photo mfgs 14.6 M - 139 - 8 X 199 29 35 Furniture 33.9 M - 308 - 47 X 228 43 36 Sound mfgs 34.5 M ~1236 - 149 X 420 60 37 Toy mfgs 30.9 M - 799 - 226 X 592 147 155 (36). The largest net gains, on the other hand, are posted by Paper Manufactures (4); Metal Manufactures (14); Organic Chemicals (16); Power Machinery (22); Other Machinery (27); Electrical Machinery (28); and Precision Instruments (32). I Net gains or losses, however, can be misleading. An industry can experience significant labor turnover without a large net effect due to composition changes in the work force. Industries in which this appears to be the case include Leather Manufactures (1); Rubber Manufactures (2); Textile Semi-Manufactures (5); Plastics (19); J .-'o _ l"_‘—* 4 Miscellaneous Transport Equipment (31); and Toy Manufactures (37). To compare the relative employment effects of Schemes A and B independent of the scale of reductions, one can return to the index used for this purpose in Table 5-1. This index is a good approximation of whether Scheme B is import- (unemployment) or export- (employment) biased relative to Scheme A. Table 5-9 contains the aggregate employment effects implied by the changes in each industry. These results indicate that under Scheme A the mid-range employment losses will be 44,795 and the employment gains will be 47,173. Hence, a net gain of about 2,378 jobs can be expected under Scheme A if the limitations on textiles and steel are ignored. If these categories are excluded, however, the employment loss is 31,289; the employment gain is 41,144; and the net gain is 9,855. These changes are miniscule in relation to the total U.S. work force, but they are not insubstantial in many of the component industries. 156 TABLE 5-9.--Summary of Mid-Range Estimates of the Employment Effects in the United States. M X Category Net A B Textiles M - 9,358 - 4,861 X 5,097 1,397 Net - 4,261 - 3,464 Iron 3 Steel M - 4,148 - 527 X 932 79 Net - 3,216 - 448 Other M -31,389 - 5,107 x 41,144 5,134 Net 9,855 1,077 Total M -44,795 -10,495 X 47,173 7,660 Net 2,378 - 2,835 The Scheme B employment effects are less promising. These indicate a net loss of 2,833 jobs for just one "round“ of Scheme B. Most of this, however, is due to the substantial employment losses in textiles. Under Scheme A textile employment losses account for about 21 per cent of the total number of displaced workers. Under Scheme B, on the other hand, the employment losses in textiles account for almost half of all the displaced workers. The status of the restrictions on textiles, however, is critical to this analysis. An alternative method of looking at the aggregate employment effects of Scheme A versus Scheme 8 is to compute an employment-bias index similar to the import-bias index used in Table 5-1. This 157 index for the total employment effects is .69, substantially less than one. Scheme B, therefore, is significantly unemployment- biased relative to Scheme A. This bias is reduced to .92 if the quota-affected categories are excluded. A major category for which Scheme B is not relatively unemployment-biased is Road Motor Venicles (30). An index value of 2.59 indicates the strong employ- ment-bias of Scheme B for this category. 5.6 Liberalization of Textile and Steel Quotas The most significant import quotas affecting the principals of this study are the U.S. agreements restricting imports of iron and steel and the U.S. and E.E.C. quotas on textiles. In both instances, the administration of the quotas lies primarily with the exporting countries. The general studies of quotas by Mintz (73), Magee (69), and Bergsten (11) and the comprehensive study of iron and steel by MacPhee (67) provide much of the required infor- mation for this section. Mintz's estimate of the cost of the U.S. textile quota as interpreted by Magee (69) indicates that the tariff implicit in the quota is at least about 35 per cent. The E.E.C. restrictions imply a lower tariff because of the high proportion of relatively unrestricted imports of Textile Semi-Manufactures (5). If one assumes that the E.E.C. restrictions are roughly equivalent except for this category, the tariff implicit in the E.E.C. textile quotas is about half that of the implicit U.S. tariff. Table 5-10 presents the increases in U.S. and E.E.C. imports of textiles (5-7) if the textile quotas were eliminated. The increases implied by both an 158 TABLE 5-10.--Changes in Trade Due to the Elimination of Textile and Steel Quotas (millions of dollars). Category Prin 1 2 Textiles USM 1,024 591 EM 1,238 710 Iron & Steel USM 2,190 1,178 EX 887 477 JX 814 438 infinitely elastic supply and a less than infinitely elastic supply are presented. Bergsten (11) suggests that Japan would not share in the expansion of textile exports to the U.S. and E.E.C. caused by an elimination of textile quotas. For this reason, no Japanese export figures are presented. Magee's (69, p. 673) estimate of 17 per cent as the tariff implicit in the U.S. import quotas on iron and steel (10-11) is used to derive the increase in U.S. imports reported in Table 5-10.6 . Under the assumption of constant market shares, the E.E.C. and Japan shares in the expansion of exports to the U.S. are also reported. The magnitude of these effects for textiles and steel indicates that liberalization of these quotas will have a substantial effect on trade in these categories. These results are also signifi- cant even when compared to the total effects of general trade liberalization. 6There are also restrictions on some imports in Metal Manu- factures (14), but these are not considered. The most significant of these is the restriction on stainless tableware. 159 Quota effects are particularly important in the U.S. labor market where the mid-range estimates of the employment losses in textiles and steel are 21,076 and 17,682, respectively. These losses combined with the losses caused by reductions in the rela- tively high explicit tariffs make it clear that trade liberalization will, indeed, result in substantial displacements of workers in the textile and steel industries. As a final note, the consideration of the textile and steel quotas in this section is not meant to deny the significance of other non-tariff barriers in these and other categories. The difficulty in dealing with non-tariff barriers, of course, is that no two are exactly alike. In this respect the textile and steel quotas differ from other non-tariff barriers because their magnitude and scope make them more tractable than more subtle and diverse forms of non-tariff barriers. CHAPTER 6 SUMMARY AND CONCLUSIONS The two major objectives of this study were to obtain improved estimates of the relevant price elasticities of import demand and to use these elasticities to examine the static price effects of trade liberalization. In regard to the former, the estimates of price elasticities obtained in this study are "improved" estimates in several senses: First, the categories for which the estimates were made are comparable for the three principals; second, the price elasticities were estimated directly for the principal to which they are meant to apply; third, the consistency of the estimates was improved by considering the potential bias in using OLS when unit value and quantity variables are measured with error; and finally, to avoid potential bias the specification of the estimated equations was not generally restricted to the use of a relative price variable. The analytical framework used to examine the effects of trade liberalization is distinguished by three characteristics: First, the "industry" level effects of across-the-board versus harmonization tariff reductions and the effects of eliminating textile and steel quotas were computed; second, the possibility of rising supply prices was considered; and third, a critique of the traditional use of elasticity identities which relate trade elas- ticities to domestic demand and supply elasticities was presented. 160 161 The overall estimates of the effects of alternative tariff reductions indicate that the U.S. and Japanese interests (in terms of trade balance, welfare, and employment) are best served by an across-the-board tariff reduction rather than reductions proportional to the original height of the tariffs. The interests of the E.E.C., on the other hand, do not appear particularly sensitive to the difference between the two approaches. The results also indicate an expansion ranging from about three to seven per cent in imports and exports for each principal. The estimated results for the elimination of the textile quotas indicate that U.S. imports of textiles will rise by about twenty to thirty-three per cent and E.E.C. imports will rise by about thirteen to twenty-five per cent. The elimination of the U.S. restrictions on steel imports would result in an increase in im- ports by at least twenty-two to forty per cent, and E.E.C. and Japanese exports to the U.S. should rise by about the same proportion. The U.S. employment effects of the elimination of the textile and steel quotas are substantial and are not compensated in the short- run by any accompanying increases in employment in the export sector. Four points should be emphasized in evaluating the basic results of this study. First, the elasticities estimated here are generally higher than previous estimates. There is evidence that this is the result of the frequent use of 2SLS in this study and the relatively rare use of a relative price variable. This conclu- sion, however, is tentative due to the possibility of intervening factors. Second, the estimated effects for the thirty-seven categories 162 represent static price effects and do not include effects such as changes in income, tastes, technology, or market structure. Third, the actual effects of the negotiated tariff reductions will likely differ from any of those presented here. However, these estimates represent the effects of the two basic alternative approaches to tariff reduction. Finally, long-term comparative advantage in a particular category cannot really be inferred from any of the results presented here. The contributing factors to these results are much too complex to single out comparative advantage as a single factor. BIBLIOGRAPHY 163 10. BIBLIOGRAPHY Adams, F. G., H. Eguchi, and F. Meyer-zu-Schlochtern. ‘An Econometric Analysis of International Trade. Paris: Organization for Economic Cooperation and Development, 1969. Adler, F. M. "The Relationship Between the Income and Price Elasticities of Demand for U. S. Exports," Review of Economics and Statistics, August, 1970. Adler, J. H., E. R. Schlesinger, and E. Van Westerborg. Ih§_ Pattern of United States Import Trade Since 1923: Some New Index Series and Their Application. New York: Federal Reserve Bank of New York, 1952. Anjaria, S. J. "The MTN and Tariff Reduction," Finance and Development, December, 1975. Balassa, B. and M. E. Kreinin. "Trade Liberalization Under the 'Kennedy Round': The Static Effects," Review of Economics and Statistics, May, 1967. Baldwin, R. E. and J. H. Mutti. "Policy Problems in the Adjustment Process (U. 5.)" (Paper presented to the Seminar on Industrialization and Trade Policies in the 1970's, International Bank for Reconstruction and Development, October, 1972). Ball, R. J. and K. Marwah. "The U. 5. Demand for Imports, 1948-1958," Review of Economics and Statistics, November, 1962. Basevi, G. "The Restrictive Effect of the U.S. Tariff and its Welfare Value," American Economic Review, September, 1968. Basic Documentation for the Tariff Study. Working Group on the Tariff Study. Geneva: General Agreement on Tariffs and Trade, 1971. Baumol, W. J. Economic Theory and Operations Analysis. 3rd ed. Englewood Cliffs, N. J.: Prentice Hall, 1972. 164 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 165 Bergsten, C. F. "On the Equivalence of Import Quotas and 'Voluntary' Export Restraints,” Toward a New World Trade Policy: The Maidenhead Papers, ed. C. F. Bergsten. Washington: The Brookings Institution, 1974. Bergstrom, A. R. "An Econometric Study of the Supply and Demand for New Zealand's Exports," Econometrica, July, 1955. Bhagwati, J. "On the Equivalence of Tariffs and Quotas," Trade, Growth and the Balance of Payments, ed. Robert E. Baldwin and others. Chicago: Rand McNally, 1965. "More on the Equivalence of Tariffs and Quotas," American Economic Review, March, 1968. Branson, W. H. A Disaggregated Model of the U. 5. Balance of Trade. Staff Economic Studies, No. 44. Washington, D. C.: Board of Governors of the Federal Reserve System, February, 1968. Bulletin Mensuel du Commerce Extereur de L'Union Economique Belgo-Lexembourgeoise, L'Institut National de Statistique, Brussels, 1971. Burrows, James C. and Thomas A. Domencich. An Analysis of the United States Oil Import Quota. Lexington, Mass.: Heath Lexington Books, 1970. Cheng, H. 5. "Statistical Estimates of Elasticities and Propensities in International Trade: A Survey of Published Studies," International Monetary Fund Staff Papers, 1959-1960. Commodity Trade Statistics. Statistical Office of the United Nations, Series D, United Nations, New York, 1963-1976. Competitiveness of U. S. Industries. United States Tariff Commission, TC Publication 473, Washington, D. C., April, 1972. Deppler, M. C. "Elasticities of Demand for U. 5. Exports: A Comment," Review of Economics and Statistics, May, 1971. de Vries, B. A. "Price Elasticities of Demand for Individual Commodities," International Monetary Fund Staff Papers, April, 1951. Feldstein, M. "Errors in Variables: A Consistent Estimator with Smaller MSE in Finite Samples," Journal of the American Statistical Association, December, 1974. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 166 Ferguson, C. E. and M. Polasek. "The Elasticity of Import Demand for Raw Apparel Wool in the United States," Econometrica, October, 1962. Ferguson, C. E. The Neoclassical Theory of Production and Distribution. Cambridge: Cambridge University Press, 1969. Fisher, Franklin M. and Karl Shell. The Economic Theory of Price Indices: Two Essays on the Effects of Taste, Quality, and Technological Change. New York: Academic Press, 1972. Floyd, J. E. "The Overvaluation of the Dollar," American Economic Review, March, 1965. Foreign Trade. Federal Statistical Office, Wiesbaden, 1971. ForeignyTrade: Monthly Statistics. Statistical Office of the 1 European Communities, Brussels, various issues. Fortune, J. Neill. "Measurement of Tariff Elasticities," Applied Economics, March, 1971. Foster, C. D. and H. L. I. Neuburger. "The Ambiguity of the Consumer's Surplus Measure of Welfare Change," Oxford Economic Papers, March, 1974. Goldstein, M. and M. S. Khan. "Large Versus Small Price Changes 'and the Demand for Imports," International Monetary Fund Staff Papers, March, 1976. The Supply and Demand for Exports: A Simultaneous Approach, unpublished paper, International Monetary Fund, March, 1976. Gregory, R. G. "United States Imports and Internal Pressure of Demand: 1948-68," American Economic Review, March, 1971. Highlights of the U. S. Export and Import Trade. United States Bureau of the Census, Report FT990, December, 1974, U. S. Government Printing Office, Washington, D. C., 1975. Holzman, F. 0. "Comparison of Different Forms of Trade Barriers," Review of Economics and Statistics, May, 1969. Horner, F. B. "Elasticity of Demand for the Exports of a Single Country," Review of Economics and Statistics, May, 1969. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 167 Houthakker, H. S. and S. P. Magee. "Income and Price Elastici- ties in World Trade," Review of Economics and Statistics, May, 1969. Hurd, M. D. "Small-Sample Estimation of a Structural Equation with Autocorrelated Errors," Journal of the American Statistical Association, September, 1972. International Financial Statistics. Statistics Bureau of the International Monetary Fund, Washington, D. C., various issues. Iqbal, Z. "The Generalized System of Preferences," Finance and Development, September, 1975. Jadlow, J. W. The Effects of Reducing Trade Barriers in the Chemical Industry. Ph. D. dissertation, Oklahoma State University, 1976. Japan Statistical Yearbook. Bureau of Statistics, Office of the Prime Minister, Tokyo, 1963-73. Johnson, H. G. "The Gains from Freer Trade with Europe: An Estimate," The Manchester School of Economic and Social Studies, September, 1958. . "The Cost of Protection and the Scientific Tariff," Journal of Political Economy, August, 1960. . ”The Costs of Protection and Self-sufficiency," Quarterly Journal of Economics, August, 1965. Johnston, J. Econometric Methods. 2nd ed. New York: McGraw Hill, 1972. and M. Henderson. "Assessing the Effects of the Import Surcharge," The Manchester School of Economics and Social Studies, May, 1967. Junz, H. B. and R. R. Rhomberg. "Prices and Export Performance of Industrial Countries," International Monetary Fund Staff Papers, July, 1965. Khan, M. S. "Import and Export Demand in Developing Countries," International Monetary Fund Staff Papers, November 1974. . "The Structure and Behavior of Imports of Venezuela," Review of Economics and Statistics, May, 1975. and K. Z. Ross. "Cyclical and Secular Income Elastici- ties and the Demand for Imports," Review of Economics and Statistics, August, 1975. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 168 Klein, L. R., R. J. Ball, A. Hazelwood, and Vandome. .An Econometric Model of the United Kingdom. Oxford: Basil Blackwell, 1961. Kmenta, J. Elements of Econometrics. New York: The MacMillan Company, 1971. Krause, L. B. ”United States Import and the Tariff," American Economic Review, Papers and Proceedings, May, 1959. . “United States Imports, 1947-1958," Econometrica, April, 1962. Kravis, I. B. and R. E. Lipsey. Price Competitiveness in World Trade. New York: Columbia University Press, 1971. Kreinin, M. E. "Effect of Tariff Changes on the Prices and Volume of Imports," American Economic Review, June, 1961. "Price Elasticities in International Trade," Review of Economics and Statistics, November, 1967. . "More on the Equivalence of Tariffs and Quotas,” Kyklos, Vol. XXIV-l97l-Fasc. l. "Disaggregated Import Functions--Further Results," Southern Economic Journal, July, 1973. Kwack, S. Y. "The Determination of U. S. Imports and Exports: A Disaggregated Quarterly Model, 1960 III-1967 IV," Southern Economic Journal, January, 1972. Leamer, E. E. and R. M. Stern. Quantitative International Economics. Boston: Allyn and Bacon, 1970. Leser, C. E. V. A Study of Imports. Dublin: The Economic and Social Research Institute, 1967. Lovasy, G. and H. K. Zassenhaus. "Short Run Fluctuations in U. S. Imports of Raw Materials, 1928-39 and 1947-52," International Monetary Fund Staff Papers, October, 1953. Maandstatistiek Van de Buitenlandse Handel, per Goederensoort. Centraal Bureau voor de Statistiek, Gravenhage, 1971. MacPhee, C. R. Nontariff Barriers to International Trade in Steel. Ph. D. dissertation, Michigan State University, 1970. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 169 Magee, S. P. "Prices, Incomes, and Foreign Trade," Inter- national Trade and Finance: Frontiers for Research, ed. Peter B. Kenen. Cambridge: Cambridge University Press, 1975. "The Welfare Effects of Restrictions on U. 5. Trade," Brookings Papers on Economic Activity, 3:1972. Malach, V. W. "Elasticity of Demand for Canadian Exports,“ Review of Economics and Statistics, February, 1957. McGeehan, J. M. "Competitiveness: A Survey of Recent Literature," The Economic Journal, June, 1968. Mikhail, W. M. "A Comparative Monte Carlo Study of the Properties of Econometric Estimators," Journal of the American Statistical Association, March, 1975. Mintz, I. U. 5. Import Quotas: Costs and Consequences. Washington, D. C.: American Enterprise Institute for Public Policy Research, 1973. Monthly Bulletin of Statistics, Statistical Office of the United Nations, United Nations, New York, 1963-74. Morgan, D. J. and W. J. Corlett. "The Influence of Price in International Trade: A Study in Method," Journal of the Royal Statistical Society, Series A (General), CXIV’(Part III, 1951). Murray, T. and P. J. Ginman. "An Empirical Examination of the Traditional Aggregate Import Demand Model," Review of Economics and Statistics, February, 1976. Neisser, H. P. "The United States Demand for Imports," American Economic Review, Papers and Proceedings, May, 1953. and F. Modigliani. National Incomes and International Trade: A Quantitative Analysis. Urbana: University of Illinois Press, 1953. Officer, L. H. and J. R. Hurtubise. “Price Effects of the Kennedy Round on Canadian Trade," Review of Economics and Statistics, August, 1969. Ophir, T. "Tariffs and Quotas," American Economic Review, December, 1969. Orcutt, G. H. "Measurement of Price Elasticities in Inter- national Trade," Review of Economics'and Statistics, May, 1950. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 170 Overall Trade by Countries. Organization for Economic Cooperation and Development, Series A, Washington, D. C., 1974. Overseas Trade Statistics of the United Kingdom. Department of Trade and Industry, London, 1963-72. Piquet, H. S. Aid,yTrade, and the Tariff. New York: Thomas Y. Crowell, 1953. Polak, J. J. An International Economic System. Chicago: University of Chicago Press, 1953. Prachowny, M. F. J. A Structural Model of the U. S. Balance of Payments. Amsterdam: North-Holland Publishing Company, 1969. Prais, S. J. "Econometric Research in International Trade: A Review," Kyklos, Volume XV, Fasc. 3, 1962. Price, J. E. and J. B. Thornblade. "U. 5. Import Demand Functions Disaggregated by Country and Commodity," Southern Economic Journal, July, 1972. Quirk, J. and R. Saposnik. Introduction to General Equilibrium Theoryyand Welfare Economics. New York: MacGraw Hill, 1968. Ramsey, J. B. "Limiting Functional Forms for Market Demand Functions," Econometrica, March, 1972. Reimer, R. "The United States Demand for Imports of Materials, 1923-1960," Review of Economics and Statistics, February, 1964. Rhomberg, R. R. and L. Boissonneault. "Effects of Income and Price and Changes on the U. S. Balance of Payments," International Monetarnyund Staff Papers, March, 1964. "The Foreign Sector,“ The BrookingsyQuarterly Econometric Model of the United States, ed. J. S. Duesenberry and others. Chicago: Rand McNally, 1965. Sargan, J. 0. "Production Functions," Qualified Manpower and Economic Performance, Part V, ed. P. R. G. Layard and others. London, 1971. Sawa, T. "Finite-Sample Properties of K-Class Estimators," Econometrica, July, 1972. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 171 Shibata, A. "On the Equivalence of Tariffs and Quotas," American Economic Review, March, 1968. Shinkai, Y. “Price Elasticities of the Japanese Exports: A Cross-Section Study," Review of Economics and Statistics, May, 1968. Statistica Mensile del Commercio con L'Estero, Instituto Centrale di Statistica, Roma, 1971. Stern, R. M. "The U. S. Tariff and the Efficiency of the U. S. Economy," American Economic Review, Papers and Proceedings, May, 1964. Steuer, M. D., R. J. Ball, and J. R. Eaton. "The Effect of Waiting Times on Foreign Orders for Machine Tools," Economica, November, 1966. Theil, H. Principles of Econometrics. New York: John Wiley and Sons, 1971. Trade Barriers: Report to the Committee on Finance of the United States Senate and its Subcommittee on Inter- national Trade. United States Tariff Commission, Washington, D. C.: U. S. Government Printing Office, April, 1974. Trade by Commodities, Series C. Organization for Economic Cooperation and Development, Washington, D. C., 1963-72. Trade Statistics of Ireland, Central Statistics Office, Dublin. various issues. United States Commodity Exports and Imports as Related to Output. U. S. Department of Commerce, Bureau of the Census, Washington, D.C.: United States Government Printing Office. United States Expprts: Schedule B Commodity by_Country, Report FT 410. United States Bureau of the Census, Washington, D. C.: United States Government Printing Office, 1963-74. United States General Imports: Schedule A Commodity by Country, Report FT 135. United States Bureau of the Census, Washington, D. C.: United States Government Printing Office, 1963-74. Wallace, T. C. and Toro-Vizcarrondo, "Tables for the Mean Square Error Test for Exact Linear Restrictions in Regression," Journal of the American Statistical Association, December, 1969. 109. 110. 111. 112. 172 Weaver, J. H. and I. Winakur. "Impact of United States Cotton Textile Quotas on Underdeveloped Countries," Southern Economic Journal, July, 1968. Wemelsfelder, J. "The Short-Term Effect of the Lowering of Import Duties in Germany," The Economic Journal, March, 1960. Wholesale Prices and Price Indexes, Bureau of Labor Statistics, Washington, D. C.: United States Government Printing Office, 1963-72. Wilson, J. F. Yet Another Econometric Model of U. 5. Imports, 1958-1971, Disaggregated by End-Use Commodity Groups and Region of Origin. Ph. D. dissertation, University of Pennsylvania, 1974. 146 0748 1293 03 3