University of Southern California Dissertations and Theses

Export Instability And Economic Development: A Statistical Verification

(USC Thesis Other)

Export Instability And Economic Development: A Statistical Verification

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content 71 - 12,337 GLEZAKOS, C o n s ta n tin e , 1936- EXPORT INSTABILITY A N D ECO NO M IC DEVELO PM EJJT: A STATISTICAL VERIFICATION. U n iv e r s ity o f Southern C a lif o r n ia , Ph.D., 1970 E con om ics, g e n e r a l University Microfilms, A X E R O X Company, Ann Arbor, Michigan Copyright by CONTANTINE GLEZAKOS 1971 THIS DISSERTATION HAS BEEN MICROFILMED EXACTLY AS RECEIVED EXPORT INSTABILITY AND ECONOMIC DEVELOPMENT: A STATISTICAL VERIFICATION by Constantine Glezakos A D issertation P resented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In P a rtia l Fulfillm ent of the R equirem ents fo r the D egree DOCTOR OF PHILOSOPHY (Econom ics) August 1970 UNIVERSITY O F SO UTHERN CALIFORNIA THE GRADUATE SCHOOL UNIVERSITY PARK LOS ANGELES, CALIFORNIA 9 0 0 0 7 This dissertation, written by CONSTACTINE GLEZAKOS.............. under the direction of h..i$... Dissertation Com mittee, and approved by all its members, has been presented to and accepted by The Gradu ate School, in partial fulfillment of require ments of the degree of D O C T O R OF P H I L O S O P H Y 7 - Dean Date AUgUSt 1 9 7 0 ACKNOWLEDGMENTS I w ish to ex p ress my sin cere appreciation to all m em bers of my d issertatio n com m ittee, P ro fe sso r G erhard T intner, P ro fe sso r Jeffrey Nugent, and P ro fe sso r C edric Pounders. E specially I w ish to ex press my gratitude to P ro fe sso r Jeffrey Nugent for inspiration, guidance, invaluable help and fo r his precious tim e which he so profusely afforded me at all stages of my work. Finally, I w ish to thank the Com puter Science L aboratory of the U niversity of Southern C alifornia for th e ir cooperation and a ssistan ce with the com putations. TABLE OF CONTENTS Page ACKNOWLEDGMENTS............................................................................................ ii LIST OF TABLES .................................................................................................... iv C hapter I. INTRODUCTION................................................................................... 1 The Problem The Im portance of the Problem The A pproach and Hypothesis of this D issertation II. PROBLEMS OF MEASURING INSTABILITY: INSTABILITY IN D E X E S .............................................................. 16 Review and Evaluation of Export Instability Indexes D esirable P ro p erties of an Export Instability Index Choice of an Instability Index P ossessin g D esirable P ro p erties The Role of Expectations in Export Instability III. PROBLEMS OF MEASURING INSTABILITY: SELECTION OF D A T A ................................................................. 40 M easurem ent of Export Proceeds Selection of Countries IV. STATISTICAL TESTING OF THE EFFECTS OF EXPORT IN STA BILITY ................................................................................... 68 The Growth Rate of Exports The Size of the Export Sector The Role of Foreign R eserves V. PRICE STABILIZATION AND ECONOMIC DEVELOPMENT . 94 The R elative Im portance of the P ric e and Quantity Instability The A rgum ents About P ric e Stabilization iii C hapter Page The E m pirical Evidence Conclusion BIBLIO G RA PH Y ........................................................................................................ 121 APPENDIXES A. Independence of the Instability Index Involving Expectations From T r e n d ...................................................................................... 131 B. A B rief Exposition of the Rationale Behind the C ochrane- O rcutt Iterativ e Method for C orrecting for A utocorrelation in the E rro r T erm ....................................................................... 136 C. Statistical Appendix ......................................................................... 140 iv Page 49 59 72 74 77 79 82 87 91 96 99 115 141 143 LIST OF TABLES OLS R egressions of Im ports on E x p o r t s ...................................... Sum m ary R esults of R egressions of Im ports on Exports C orrected for A utocorrelated R esiduals ................................. Sum m ary of P e r Capita GDP Growth Rate and E xport Instability R egressions ................................................................. R egressio n s of the P e r Capita GDP Growth R ate on the Growth Rate of E x p o rts.................................................................. R egressions of the Export Growth Rate on the Export Instability Index ............................................................................. Sum m ary R esults of Income G row th-Export Instability and Export Growth M ultiple R egressions .................................... Sum m ary R esults of G row th-Instability and Size of the Export Sector M ultiple R e g r e s s io n s ....................................... Sum m ary R esults of Income Growth, Export Instability, and Foreign R eserves R egressions ................................................ R eg ressions of Income Growth, Export Instability and Fluctuations in Foreign R e s e rv e s ............................................. Value, Quantity and P rice Instability In d e x e s ........................... Sum m ary R esults of the Export P rice-Q uantity E lasticitie s and T h e ir C orrelations (Less Developed C ountries) . . . . Sum m ary R esults of the Effects of Export Quantity and P ric e Instability on the Income and Export Growth R a te s ............ Instability Indexes ............................................................................. Values of V ariables Used for S tatistical A nalysis of Export Instability and Economic Growth (1 9 5 3 -1 9 6 6 )....................... v CHAPTER I INTRODUCTION International trade has been considered to be an instrum ent of national income expansion since the days of the M ercantilists and many argum ents have been developed in support of the gains from trad e. Recently, however, som e controversial questions have been raised as to the relevancy and validity of the classical theory of international trad e. In the context of dynam ic and stru c tu ra l problem s of today's le ss developed co u n tries, such controversy has a rise n because of the static character* an d /o r 2 the underlying assum ptions of the theory. B ecause in the post World W ar II period an ev er increasing num ber of le ss developed countries is taking d elib erate steps tow ard accelerated economic grow th, questions directly re lated to the export se c to r and its role in the econom ic grow th of these countries becom es of g reat im portance. Justifiably, th e se questions have attracted the in te re st of both the academ ic econom ists and the econom ic p ractitio n ers. * G erald M. M eier, International T rade and Development (New York: H arper & Row, 1963), p. 153. Also, Shu-Chin Yang, "F oreign T rade P rob lem s in Economic Developm ent," Scottish Journal of P olitical Economy, XI, No. 2 (June, 1964), 116-135. 2 Gunnar M yrdal, An International Economy (New York: H arper & B ro th ers, P u b lish ers, 1956), pp. 337-340. The Problem The validity of th e theory of export led grow th, especially with reg ard to less developed countries, has been challenged on the grounds that exports may be subject to fluctuations which may have detrim ental effects on econom ic growth. B ecause le ss developed countries export m ainly prim ary products whose p ric e s suffer larg e and som etim es violent fluctuations and also because of the high concentration of th e ir exports in a very sm all num ber of com m odi tie s , it has been claim ed that the le ss developed countries experience high instability in th e ir export proceeds. This claim was firs t substantiated by the U. N. study, Instability in Export M arkets of U nder-Developed C ountries, * the resu lts of which dem onstrate the high d egree of instability in p rice s and export proceeds of a la rg e num ber of p rim ary com m odities and exporting countries. But, even before these re su lts w ere known, it was not difficult to su s tain such a claim on an a p rio ri b a sis, since it is generally accepted that sup ply and demand e lasticitie s of p rim a ry products a re very low and that th e re is a significant num ber of uncontrollable facto rs lying behind th e ir supply and *U. N. D epartm ent of Economic A ffairs, Instability in Export M arkets of Under-Developed C ountries, Sales No. 1952, n , A (New Y ork: U. N. D epartm ent of Economic A ffairs, 1952), 1. demand. P rim a ry commodity p rices a re also "m ore influenced by specula tive buying and selling, and changes in inventories often tend to aggravate o instead of dam pening fluctuations in dem ands. " The aforem entioned U. N. study, n ev erth eless, "is confined to the m easurem ent of fluctuations" only and does not deal "with the problem s 3 raised by . . . instability. " However, sev eral authors have developed a This can be easily shown with the u se of an exam ple. If, fo r sim plicity, we assum e constant (price) elasticity demand and supply functions fo r a given p rim ary com modity, i. e . , if (1) D A P ^ and (2) S = C P , then a t equilibrium it would be: (3) P = A C 1 P+y From Equation (3) we obtain: (4) dP P dA A 1 p+y and (5) dP P dC C 1 p+y Now, from Equations (4) and (5) it is obvious that the relativ e change in p ric e w ith reg ard to the rela tiv e change of the autonomous facto rs is la rg e r the sm a lle r the p ric e e lasticitie s of th e demand and supply functions. 'G unnar M yrdal, p. 238. 3 U. N. D epartm ent of Economic A ffairs, Instability in E xport M arkets of U nder-Developed C ountries, p. 1. num ber of argum ents suggesting that excessive fluctuations in the export sec to r of the le ss developed countries a re detrim ental to the stability and eco nomic growth of these c o u n trie s.1 A fter careful exam ination, the argum ents of the effects of export in stability can be classified into two g eneral categ o ries: (1) those related to the p ric e instability of the p rim ary products p e r se and (2) those related to the fluctuations of the export proceeds of the p rim ary products exporting less developed countries. Synoptically, the argum ents of the fir s t type advocate that due to ex cessiv e p rice fluctuations of the p rim ary com m odities exported by the less developed countries, firs t, the economic risk s of devoting reso u rces to p ro duction for exports a re high, and, th erefo re, production fo r exports becom es relatively le ss profitable and consequently le ss attractiv e. So in essence, it is argued that p rice instability re ta rd s the grow th of the export se c to r by d is couraging investm ent in p rim ary production fo r exports. B rainard and Cooper p re se n t the counterpart argum ent that excessive p ric e fluctuations frequently cause hedging behavior on the p a rt of the fa rm e rs which induces them to de vote th e ir reso u rces into subsistence farm production. They claim that "the hazards of being faced with high food p rice s in y e a rs of low income a re suffi ciently disastrous in many p a rts of th e world to induce fa rm e rs and nonfarm ers * G erald M eier, p. 181. Also R agner N urkse, "T rade Fluctuations and Buffer Policies of Low-Incom e C ountries." Kyklos, XI (1958), 141-154 (espe cially p. 143); A. K. C airn cro ss, F actors in Economic Developm ent (New York: P raeg er, 1962), p. 213; and Gunnar M yrdal, op. c i t . , p. 229. alike to retain land for home use. This risk -av ertin g behavior helps perpetu ate extrem ely low productivity in agriculture. "* Second, th e re is a danger that highly unstable p rice s may induce the developm ent of synthetic products fo r the 2 replacem ent of natural raw m a teria ls o r that they may force im p o rters to shift to other v a rie tie s by modifying th e ir production p ro c e sse s. F o r exam ple, P o rte r states that "in the p o st-w ar period when p ric e s of Egyptian cotton w ere both high and unstable, spinners in L ancashire and W estern Europe bought other v a rie tie s and adapted th e ir m achines and the quality of th e ir products accordingly. Once such a change has been m ade it req u ires a considerable fall in p rices to persuade spinners to re s e t th e ir m achines and they a re not, in any case, likely to do so until . . . they can be assu red of a reasonable degree of O stability in p ric e s in the future. " On the other hand, the basis of the argum ents of the second category is the dependence of the economic grow th of the less developed countries on im p o rts of investm ent goods which have to be paid by export earnings, unless a country has other so urces of foreign exchange. F irs t, it has been argued that *W illiam C. B rainard and R ichard N. Cooper, "U ncertainty and D iver sification in International T rade, " Studies in A gricultural Econom ics, T~ade, and Developm ent, VIII, No. 3 (1968), 257-285. 2 H. W. Singer, "Introductory Statem ent, " Kyklos, Second Symposium, XII, F asc. 3, (1959), 271-282 (especially p. 277). 3 R. S. P o rte r, "C om m ent," Kyklos, F irs t Symposium , XI, F asc. 2 (1958), 231-243. for each less-developed country th e re may be a m inim um of im ports required for its economy to operate at full capacity. * If, th e re fo re, due to frequent fluctuations in export proceeds this m inim um cannot b e obtained, the economy would not be able to operate and grow at its optimum ra te of growth. A. H irschm an, on this point, adds that "for the tim e being the problem is be coming ra th e r m ore serio u s as the exporters of p rim ary products a re indus trializing. In the course of this p rocess raw and sem ifinished m a teria ls tend to replace finished consum er m anufactures as a p rin cip al im port category. Sudden sh arp declines in export proceeds becom e of g re a te r concern under such conditions since they m ean d ra stic curtailm ent of econom ic activity and em ployment ra th e r than, as before, postponem ent of final purchases fo r con sum ption and investm ent. Second, the m o st frequently heard com plaint is that strong fluctuations in export proceeds ren d er effective planning extrem ely difficult, sin ce it is intuitive that with stable export earnings " it is e a s ie r to take a long view and plan dom estic investm ent without the constant interruption 3 that destroys half its value. " Third, excessive fluctuations in export trad e *S. B. L inder, T rade and T rade Policy for D evelopm ent, (New York: F. A. P ra e g e r P u b lish ers, 1967), p. 9. 2 A lb ert O. H irschm an, "P rim ary P roducts and Substitutes: Should Technological P ro g re ss be P o liced ?" Kyklos, XII, F asc. 3 (1959), 354-361 (especially p. 355). 3 A. K. C airn cro ss, p. 214. W allich also put th is argum ent in its pro p er p ersp ectiv e by stating that "Instability turns program m ing, at b est a guessing gam e, into a rank speculation. " Henry C. W allich, "Stabilization of Proceeds from Raw M aterial E xports, " in Economic Developm ent for Latin A m erica, ed. Howard S. E llis (London: M acm illan & C o ., 1961), pp. 342-36L also may have a detrim ental influence on long-run p riv ate investm ent since they "may well be a m ajor cause of the speculative attitude and the 'g e t-ric h - quick' m entality so w idespread among businessm en in underdeveloped countries. Finally, it has been claim ed that the national incom e of the less developed countries is m o re sensitive to the ill effects of instability in the ex p o rt proceeds because these countries do not have at th e ir disposal effective m onetary and fiscal policy instrum ents to counterbalance such instability. The m ost im portant im plication of the above argum ent is that fluctuations in export proceeds a re expected to c re a te p arallel p ric e fluctuations dom estically. For, since th ese countries a re assum ed to be ineffective in counterbalancing the sharp changes in th eir export earnings, th e ir dom estic demand is expected to move accordingly. Now, given the low sh o rt-ru n supply e la stic itie s, especially of farm products in the less developed cou n tries, it is expected that changes in demand will re su lt in significant dom estic p ric e changes in these countries. M ore im portant, however, it may be argued that in the le ss de veloped cou n tries, fo r political and institutional reasons as well as because of g re a te r effectiveness in handling deflationary than in handling inflationary p re ss u re s, governm ents m ay be inclined to re a c t only to the fo rm e r and by so doing tran sfo rm what would have been a p ric e fluctuation re su lt of export in stability into a long-run inflationary process. ^ a g n a r N urkse, p. 143. All the above argum ents concerning the effects of export instability on economic grow th w ere m ade on an a p rio ri basis with none, o r at best, very little em p irical evidence. R esults of recent em p irical studies that have dealt with the problem of export instability, however, do not show th at commodity concentration is a m ajo r facto r in determ ining the strength of p rice fluctua tions of exports as one should expect on a p rio ri c o n sid e ra tio n s,1 o r th at the differences in export instability between developed and less developed countries 2 "a re much le ss than commonly supposed. " M ore im portant a re th e ir a s s e r tions that from the findings th ere is no sta tistic a l evidence th a t fluctuations in export proceeds "inflict any significant dam age on the stability and grow th of 3 the average underdeveloped country, " of that th e re is "any relation between 4 grow th in p e r capita re a l incom e and export instability. " We have se v e ra l reasons to believe that statem ents as the above, e s pecially those pertaining to the effects of export instability on econom ic growth, a re not justified o r a t le a st that they cannot be considered as proven by the lype of analysis used. M. M ichaely, Concentration in International T rade (A m sterdam : N orth-Holland Publishing C o ., 1962), pp. 73-74, and Benton F. M assell, "E xport C oncentration and Fluctuations in Export Earnings: A C ross-Section A n aly sis," A m erican Economic Review, LIV, No. 2 (M arch, 1964), 47-63. O A lasdair I. Mac Bean, Export Instability and Economic Development, (Cam bridge: H arvard U niversity P re s s , 1966), p. 34. 3M acBean, p. 32. 4 Joseph D. Coppock, International Economic Instability (New York: M cGraw-Hill Book Company, 1962), p. 107. 9 With regard to the claim that export instability does not seem to be p articu larly related to concentration, B rainard and Cooper have explained that this is due to the fact that countries tend to produce a range of products which have sim ila r ch arac te ris tic s , e . g . , th e ir exports all depend p rim a rily on industrial produc tion in the m ajo r countries o r they a re all subject to the vicissitudes of rainfall and sunshine. These countries, though ostensibly 'd iv e rsifie d ,' have not diversified properly; sim ply adding com m odities to the list of exports is not sufficient. Some of them m ay have diversified properly through im port substitution, of co u rse, but this would not be reflected in the figures on export earnings and hence would not influence the findings of M assel and o th e rs .1 The em pirical analyses which intended to verify o r refute the a p rio ri argum ents of the effects of export instability on econom ic growth utilized c e r tain instability indexes. As m ight be expected, these indexes vary rath e r significantly according to the form ula used fo r th e ir estim ation. So, for instance, despite M acBean's claim that "as long . . . as each index is calcu lated for the sam e variable over the sam e period of tim e, the resu lts a re 2 invariably highly co rrelated , " we found that his and Coppock's export in sta bility indexes for the sam e sam ple of 37 countries have a co rrelatio n coeffi cient of .440 and a Spearm an's rank c o rrelatio n coefficient of .436, which 3 cannot be considered to be very high. ^B rainard and Cooper, p. 271. 2 MacBean, p. 34. 3 MacBean, p. 40, Columns (1) and (2) of Table 2:3. 10 MacBean based his conclusion that th e re a re no significant effects of export instability on econom ic growth on the following pieces of evidence: firs t, the lack of association between the instability index of exports and national incom e fo r 35 less developed countries using Coppock's data, and second, the lack of significant reg ressio n coefficients in the m ultiple re g re s sion analysis between the rate of growth of GDP and the export instability index, the rate of growth of im port capacity, the ratio of foreign trad e to in com e, e t c ., for 23 less developed countries fo r the period of 1950 to 1958. * S im ilarly, Coppock considers as evidence the lack of correlatio n be tw een his export instability index and the ra te of grow th of GNP (1951-1957) o r the percentage in crease in p e r capita GNP (1951-1957) for all countries fo r which he had data. Each analysis, however, seem s to suffer from sev eral draw backs which we believe have rendered resu lts u nreliable and which have distorted the tru e relationships these authors intended to estim ate. • Coppock's instability index, which is m erely the antilogarithm of the log-variance of the yearly ra te s of change of a tim e se rie s, as 2 it would be shown below, is greatly influenced by the choice of the f ir s t y e a r and the la st y ear of the s e rie s . It can be shown that ^M acBean, p. 65 and pp. 123-124. . 2 See C hapter II, p. 21 of this study. 11 if the la s t y e a r of two otherw ise identical tim e s e rie s is different only in direction from the trend, the instability index of the se rie s with the low er la st y ear value would be la rg e r than the other, som etim es significantly, depending upon the v alu e of the la s t year. We feel, th erefo re, that the m easurem ent of instability with such an index, especially when it is used fo r a sh o rt-ran g e tim e s e rie s , is sim ply a random estim ate. • M acBean's instability index, using a fiv e-y ear moving average for co rrectin g fo r the trend, m ight be biased upw ards due to the use of moving averages that can introduce perio d icities o r fluctuations which a re not p resen t in the original data, according to Slutzky's th e o re m .1 • F o r the incom e grow th rate , Coppock used the "percentage in c re a se in GNP p e r y e a r" and the "in crea se in p e r capita GNP, " 2 both adjusted for p rice changes, while M acBean used the "com - 3 pound annual ra te of grow th of GDP" in c u rre n t p ric e s. Both authors, th erefo re, invited distorting e rro rs as discussed below. F irs t, since the le ss developed countries have higher population *See E. M alinvaud, Statistical Methods of E conom etrics (Chicago: Rand McNally & C o ., 1966), p. 378. 2 Coppock, pp. 106-107, and p 161. 3 MacBean, pp. 118*119, Table 4:1. 12 grow th rates than do the developed countries, the use of the total GNP o r GDP grow th ra te s, ra th e r than the respective p e r capita grow th figures, introduced an upwards bias for these countries. Evidence of this e r r o r is shown in Coppock’s resu lts. His c o r relation coefficient between export instability and GDP growth rate is -. 003. However, when he used the p e r capita GNP figures, the co rrelatio n coefficient was -. 15, despite the fact that his "growth in p e r capita real GNP" figures a re nothing e lse than the total percentage change of this variable over the period 1951 to 1957. * Second, since inflation in the le ss developed countries is usually stro n g er than in the developed countries, the use of GDP data in cu rren t p rice s for the estim ation of the incom e growth 2 rates actually o v erestim ate the grow th rates of the form er. • Additionally, Coppock m akes two other methodological e rro rs . On the one hand, while his export instability indexes have been estim ated for the period 1946 to 1958, his incom e grow th figures cover the period 1951 to 1957. On the other hand, the sam ple of countries he uses to estim ate the relationship of export instability and income grow th includes a la rg e num ber of developed countries 1Coppock, p. 106. j I 2 | Coppock states that his incom e data, obtained from an unpublished | so u rce, a re adjusted fo r p ric e changes without mentioning how this co rrectio n ! is m ade. Coppock, pp. 106 and 161. I 13 which, according to the a p rio ri argum ents, a re not significantly affected by export instability. F or these reaso n s, we feel a different approach, fre e from these draw backs, should be used fo r em prical verification o r refutation of argum ents concerning the effects of export instability on econom ic developm ent. The Im portance of the Problem The idea of stabilization of international com modity m ark ets is not new; efforts to im plem ent international com m odity agreem ents have dated from the end of the nineteenth century. E specially a fte r W orld W ar II, how ever, sev eral stabilizatio n schem es on a national o r international scale have been proposed, although few have actually been im plem ented in an attem pt to reduce the fluctuations in the volum e of trad e and the p rice s of the p rim ary com m odities. A lso, se v e ra l conferences have been organized fo r the purpose of finding solutions and many books and a rtic le s have been w ritten offering suggestions and rem ed ies on how the le ss developed countries can go about dim inishing instability in th e ir exports o r m itigating its effects. All of these, of co u rse, have taken place under the explicit o r im plicit assum ption that the a p rio ri argum ents about the d etrim en tal effects of export instability fo r the less developed countries a re valid. Should em pirical evi dence not support th ese argum ents, it is obvious th a t all endeavors for stab ili zation of export p ric e s o r export proceeds fo r p rim a ry products have been in vain and that both international organizations and le ss developed countries should refocus th e ir efforts fo r stim ulating econom ic growth. 14 The Approach and Hypothesis of this D issertation On the basis of the observed m ethodological e rro rs of the previously m entioned em pirical studes, our hypothesis is that a m ore satisfacto ry ap p ro ach should show that export instability has m ore serious consequences in te rm s of economic growth fo r the less developed countries than shown by these studies. This hypothesis will be tested herein by the following ste p s: In C hapter n , we m ake a survey and critica l evaluation of the instability indexes which have appeared in the lite ra tu re and form ulate certain c rite ria that a good instability index should satisfy. A new instability index which fulfills these c rite ria is also developed. In C hapter III, te sts a re perform ed on the relatio n ship between im ports and exports in o rd e r to se le c t and use in fu rth er s ta tis tic a l testing only those countries for which such a relationship is found to be significant. Such a step seem ed appropriate, inasm uch as the m o st frequently m entioned effect of export instability is said to be exercised through the capac ity to im port. C hapter IV p resen ts sta tistic a l testing of the effects of export instability on economic growth, both for all the countries fo r which appro p ria te data w ere available and for the developed and less developed countries sep arately . The ro le of other related v ariab les, such as the grow th ra te of exports, the relativ e size of the export se c to r, and the amount of foreign re se rv e s is also investigated. Finally, C hapter V shows the relative 15 im portance of p ric e and quantity fluctuations and privides som e policy recom m endations on the basis of our findings. CHAPTER n PROBLEMS OF MEASURING INSTABILITY: INSTABILITY INDEXES B ecause em pirical studies that deal with the problem of instability of exports depend significantly on the m easurem ent of instability, it is im perative to begin our study with a review and evaluation of the instability indexes used in such studies. We shall propose certa in c rite ria for the evaluation of in stab ility indexes and develop an index which is shown to p o ssess desirable p ro p e rtie s. In view of the fact that our study is concerned m ainly with sh o rt-ru n instability, we w ill lim it our attention to those instability indexes which are specifically designed to m easu re y e a r-to -y e a r variation. We will not be con cerned with m easu res of cyclical fluctuations and p eriodicities. Review and Evaluation of Export Instability Indexes Instability indexes can be classified in th ree general categ o ries. In the f ir s t category belong those indexes which m easu re instability as an arith m etic m ean of y e a r-to -y e a r fluctuations. The second category includes those indexes which m easu re instability as an average of y e a r-to -y e a r fluctuations a fte r co rrectin g fo r the trend factor. The th ird category includes all those indexes which m easu re instability as an average deviation around a fitted trend line. 16 17 It m ust be stre ss e d that th ere is no general c rite rio n according to which we can say that the one category of instability indexes is b e tte r than another. The choice of an index depends mainly upon the type of variable whose instability we want to m easu re and the use we intend to m ake of the in dex. F o r instance, th e re is no reason why an instability index of a city 's te m p eratu re should have a co rrectio n for trend. T here a re not many in sta bility indexes in the fir s t category. In the U. N. study, Instability in Export M arkets of Underdeveloped C ountries, instability was m easured as the average of the absolute values of the relativ e y e a r-to -y e a r changes, with the change estim ated as p ercen t of the higher value between the beginning and the term in al 1 y e a r, e.g., i = - L V * IXt ~ Xt - i * 1 n- 1 / i X T t = 2 L w here X is the la rg e r between X and X . L t t"l 2 M ichaely also used an instability index which belongs to this category. His instability index is the average of the absolute values of the y early relative changes. Its form ula i s : 1In this rep o rt, a ris e was m easured as a percentage of the term in al high point, ra th e r than of the low er startin g point, of an in crease. United N ations,D epartm ent of Economic A ffairs, Instability in Export M arkets of U nderdeveloped C ountries, p. 77. 2 M. M ichaely, p. 68. 18 I. 2 A lternative form ulations of indexes of the fir s t category are : I 1_ X 1 o r n-1 I 2 1 X n-1 We w ill not be g reatly concerned about th ese kinds of instability indexes for sev eral reasons. F irs t, th ese kinds of indexes show a higher degree of instability for the sam e variation the g re a te r the trend. Second, these indexes have very few of the features which a satisfactory instability index m ust have, as discussed below. Finally, the indexes m entioned above w ere used for m easuring mainly p ric e instability, in which case the trend factor was not considered to be of g re a t im portance. In the second g en eral category of instability indexes, the category in cluding indexes involving som e co rrectio n for the trend, it is obvious that th e re would be sev eral different types of indexes, depending not only upon the way the co rrectio n fo r the trend is m ade, but also upon the type of trend assum ed. At this point we m ust m ake c le a r that co rrectin g fo r the trend by sim ply "subtracting the average annual in c re ase . . . from the actual in crease 19 o r d e c re a se ”* leads to a serio u s m istake. It is easy to show that such a co rrectio n depends exclusively on the difference between the firs t and the la st observed value of the resp ectiv e variable. According to this type of suggested trend elim ination, the co rrected value of the yearly change is defined as: * x ; = v x t-i - a x w here AX is the average annual in crease. In other words n .) 11 t=2 f(X -X J + (X„ - X J + . .. + (X -X J ] n-1 lv 2 l ' ' 3 2 ' n n-1 7 = X n ~ Xl n-1 th e re fo re, the two extrem e values of the s e rie s will greatly influence the value of such an index using co rrected firs t differences for m easuring 2 instability. 3 M ass ell defines the following instability index as: * Mac Bean, p. 102, footnote a. O Such co rrectio n should be avoided, not only in the construction of in stability indexes, but also if one wants to u se co rrected fo r trend firs t d iffer ences fo r reg ressio n analysis. If, fo r instance, the firs t differences of two v ariab les Yt and Xt a re co rrected for the trend in such fashion and reg ressed , the reg ressio n coefficient of AYt on AX^ is b = [(n -l)^ 2 (Y t -Yt-i)(Xt-Xt _ i) - (Yn-YD(Xn-X ij/([n-l) t£2(Xt-X t_1)2-(X ir X1 )^ and obviously depends g reatly on the extrem e values of both v ariables. 3 Benton F. M assell, "Export Concentration and Fluctuations in Deport 20 E w t * ^ Vl - U « 3 n-1 n-1 W Max Xt ’ Xt-1 w here Max Xt> Xt+1 m eans the la rg e r of the two values, and U ^ . = X^-Xt = X - (a + bt). 1 It can be easily shown that his index belongs to the second category of instability indexes since it reduces to: n-1 I„ =J_ v l xt + i ’ xt + i ‘ <V xt > l 3 n-1 A --------------- --— ----------- t=l Max Xt, Xt+1 n-1 =* 5 n-1 ^ If Vi - W .- V l “ v Vl V i-x t-b M axX t,X t+1 Ig therefore, is m erely the average of the absolute values of the yearly p e r- 2 centage changes, co rrected fo r the trend. An alternative form ulation of this index would be: '3 E<Xt _ X , . , -_b>2_ n-1 Earnings: A C ross-S ection A nalysis, AER, XIV, No. 2 (M arch, 1964), 50. 1 A X = a + bt, is the lin ear trend fitted by the ordinary le a st squares (OLE) method. 2 The problem with this index is that th ere is no reason, and its author does not explain, why the yearly change should be estim ated as p ercen t of the highest of the two term in al values. 21 Coppock used an instability index which m akes co rrectio n for exponen tia l tr e n d .1 Despite the discrepancy between the definition of his instability index and the form ula he gives fo r it, Coppock's instability index can be w ritten as follows: I4 = 100 [antilog w here Vlog = ^ I > < Xt+l /X t> ' m)2 - d r Z h (X t+ i/xt>- „Ti2 > (X t+ i/xt> ]2 fo r t = 1,2, . . . n-1. It can be easily shown that Coppock’s index has a serio u s defect in that it greatly depends on the two extrem e values, since , X 1 > , t+1 1 , n m = ^ h l o g v = ^ 0 8 \ This is the sam e defect we mentioned above w ith regard to the co rrectio n of the yearly differences by sim ply subtracting the arithm etic average of all dif ferences. Coppock's index, by assum ing an exponential trend, will also show higher instability the higher the export trend of a country, even if the country has a perfectly stable lin e a r export trend. The third category of instability indexes also includes sev eral of those indexes that differ p rim arily by the type of trend the deviations a re m easured from as well as the norm alization used. ^Coppock, p. 24. 22 M assell, for instance, used as a second instability index the standard 1 e rro r of estim ate divided by the m ean of the observations i - y . ^ v V h w here = a + bt, that is the lin ear trend fitted by the OLS method, and X - XXj/n, fo rt =1,2, ... n. A lternative form ulations of the above index a re: n i A i 2 | v x * I5 = n ^ — X— and t=l t ft I, E i v *,i l t=l X Instead of m easuring the deviations from a lin e a r trend, it is possible to m easu re them from a moving average trend line. This type of index was 2 3 used in an IMF study and by MacBean. Both studies used as an instability index the arith m etic m ean of the annual percentage deviations from a five-y ear moving average. '''M assell, p. 49. 2 "Fund Policies and P rocedures in Relation to C om pensatory Financing of Commodity Fluctuations, " IMF Staff P a p e rs, VIII, No. 1 (November, 1960), 1-76. 3 M acBean, p. 34. 23 w here for t = 1,2, .. n. This type of index has the advantage of not assum ing any specific shape fo r the trend. However, it also has som e disadvantages. F irs t, it loses four observations, which if the tim e se rie s available is not very long, m ight be an im portant handicap. Second, there is som e a rb itra rin e ss as to the num ber of y ears to be used in the estim ation of the moving average, and th ird , th e re a re certain theoretical qualifications with reg ard to the use of moving averages for the estim ation of the random disturbances, since it is possible fo r periodicities which a re not in the original s e rie s , to be introduced in the co rrected s e r ie s .1 An alternative of the above type of index is We can also con stru ct an instability index which m easu res instability as the average deviation from an exponential trend. n ■ I, 6 w here X I, 7 100 antilog An alternative of this index would be: I, 7 *See C hapter I, Introduction, p. 11. 24 ^ A X in both cases is the lin ear logarithm ic trend log X = log a + tlog b, v V estim ated by the OLS method. D esirable P ro p erties of an Export Instability Index As mentioned before, evaluation of an instability index depends greatly on the nature of the variab le whose instability is being m easured and on the use intended to be m ade of the index. With this in mind, we tried to establish certa in c rite ria that an insta bility of export index m ust satisfy if it is to be useful in a study such as ours whose p rim ary purpose is to evaluate the effects of such instability on eco nomic development. We propose that a good instability index m ust have the five following p ro p erties. F irs t, it m ust be fre e of the trend. This m eans th at the index should have the sam e value fo r the sam e variation around a tren d , indepen dently of the trend slope. F o r instance, if equal to I (see p. 25). We believe this property is indispensable in m ea- 13 surin g the influence of the instability of exports net of the effect of the growth in exports. Second, the index should take into consideration the relativ e im portance of the variation. In other w ords, in the case of two countries with exactly sim ila r absolute variation, the index should show a sm a lle r value fo r w here x f = a + bt, x f = c + dt and b^d, then if 5 ^ = X6 , I m ust be 25 the country having the la rg e r average am ount of exports. X t s o l f x f - x f = x f - x f A A A D w here Xj = a = bt, X t = a ’ + bt and a > a' — A — B which m ean X *X , then 1^ m ust be less than Ig . X T hird, if two export tim e s e rie s a re sym m etric with regard to a com mon tim e trend, th e ir respective instability indexes m ust be equal. That is, A * * • B * if X^ - X^ = X^ - X^ w here X^ = a + bt, th e ir common trend line, then 1 ^ m ust be equal to IB . 26 X X = a + bt t Fourth, if two tim e se rie s a re rev ersed with regard to tim e—th a t is, if they have the following relationship: < - Xt “ Xn+1- , - V m f t - 1. 2 . . . . . » > w here X = a + bt, i. e . , th e ir common tren d , then I m ust be equal to I . X A ij x X f = a + b t t Finally, if one tim e s e rie s is a m ultiple of the other, that is if B A X = cX , then I. m ust be equal to I . This property obviously assum es X t A J3 th at the effect of instability is of relativ e ra th e r than absolute im portance. 27 Choice of an Instability Index P ossessin g D esirable P ro p erties In view of the above p ro p erties an export instability index m ust have, we have adopted as our instability index the following: x _ 100 Z I Xt " Xt-1 ~ b M ~x i A n - 1 ' 1 w here b is the slope of the lin ear trend, = a + bt, fo r t = 1, 2 , . . . , n. It is obvious that our instability index is the arith m etic m ean of the ab solute values of the yearly differences co rrected for the lin e a r trend, and ex p ressed as the percent of the average of all observations. Next, we will briefly show that ou r index satisfies all five of the above- m entioned d esired p roperties. Xf - ^ = Xf - Xf (1) 2 w here X ^ = a + bt, = c + dt and X ^ = X^, then . < * > *We have labeled this index as our "m echanical" index of instability. 2 — . X = E X /n , i. e . , the average of the respective tim e se rie s. 28 Subtracting (2) from (1) we have: Xf - *f-l> * Xf - A - l - Xt !l> < 3> o r < - < . - b - xf - x t!i - d <4) (4) would also be c o rre c t for the absolute values of the differences and fo r all t (t = 2, 3, . . . , n). th e re fo re , n - h | ■ £ K - x « t =2 t=2 - d (5) Now if X* = 5 ? then (n-1) 3^ = (n-1) X B , (6) Finally, dividing (5) by (6) and m ultiplying both sides by 100 we get: I . = I . A B A * A B AB 2. If Xt - X f = Xt - X t (6) » A * B w here X^ = a + bt, X^ = a ’ + bt and a > a* then, as was shown above (3), < - x f - i - (X* - x f-i> - - < 1 - < x f - ^ - i > ° r K - xt-i - b = xf - xt-i - b < 7» 29 (7) is tru e fo r the absolute values of the differences and fo r all t (t = 2, 3, . . . , n), th e re fo re ,- n Z l < • < i - b t =2 Z lxf - xt-i - b i < s> t=2 If a > a' it is easy to show that X^ > XB, from which follows that (n-1) ^ > (n -1) X8 (9) Dividing (8) by (9) and m ultiplying by 100 we have I* < A B 3. If A * x t - x t B x t - x t (10) w here Xt = a + bt then according to (4) it w ill be; A A A A xt - X,-1 - ‘ V ’W A A V x t-i - ( X f - x t - i* °r X A _ x A x t x t - l - b b " < Xf “ X ® 1> (11) from which follows that b - < x t - x ,-i» - b (12) Summing (12) over all t (t = 2, 3, . . . , n) we get n t=2 t=2 - b If = a + bt is the common trend of the two tim e s e rie s then zx = zxx z X t o r X ^^ = X®, and - 1) X^ (n -1) (n-1) Now by dividing (13) by (14) and m ultiplying by 100 we get IA = Ifi. 4. If x f - X = X“ , A — X , 4 t t n + l-t n + l-t A B w here X^ = a + bt, the common trend of X^ and X j., then A T> T3 X - X = X - X = X - X t-1 t-1 n + l-(t-l) n+1 - ( t - 1) n+2- t n+2- t Subtracting (16) from (15) we get: Xf - < 1 - <*t ' Xt-1» = Xn+l-t - Xn+2-l " < V l - t - W x A . - * • - B rB X / - b = X , - X „ + b t-1 n + l-t n+2- t Taking the absolute values of (17) we have: 31 (18) holds for all t (t = 2, 3........... n). T herefore, sum m ing up (18) for a ll t we have: E I xf - X A - b I = Z|XB - XB - b 1 t t -1 1 1 n+2- t n + l-t *|XB - - b (19) If we sum (15) for all values of t (t = 1, 2, . . . , n) z x f- = E X® - E X = E X® - E X t t n + l-t n + l-t t t (20) from which it is obvious that EX^ = E x f (21) since X^ = a + bt is the common trend fo r X ^ and X® . From (21) it is easily shown that: (n-1) X* = (n-1) X® (22) finally dividing (19) by (22) and m ultiplying both sizes by 100 we get I = I . J \ D 5. If X® = c XA (23) VB A then since e X = c EX , V I Xs = c XA (24) if XA = a + ' ‘ ’ "® Also if X * = a + bt and X = A + Bt, it can be easily shown that A = ca l » I 32 and B = cb by virtu e of (23) th erefo re, ° lx f - x t - i - B I - i : K - cx tA - i - cb t=2 t =2 n (25) = o ^ | x A - X * - b t =2 from (24) we have (n-1) X® = c (n -l) XA . Dividing now (25) by (26) and m ultiplying both sides by 100 we get: (26) 100 Z | X® - X® x - B | c (100) E | XA - X® ± - b | X® (n-1) cX* (n-1) xoo n-1 o r I = i 1 B A Not only does our index satisfy all five proposed p ro p erties of a good index, but it can also be given som e logical econom ic interpretation. One could consider, fo r exam ple, th at the whole in c re ase in export proceeds from one y e a r to the next m ight not be unexpected and th erefo re disturbing. P a rt of the in c re ase m ight be expected, if an in creasin g trend in export proceeds has been observed in the past. So, we m ay assum e that not all the difference This is a p articu larly useful p roperty fo r an instability index, because it can be used to estim ate the instability involved in a tim e se rie s of index num bers, since it would give the sam e value independently of the chosen base y e a r of the se rie s. 33 X - X is considered as disturbing, but, ra th e r, only its unexpected p a rt £t X d = |X - X - b I (see below). X u X x t A lso, if a decline in the export proceeds occu rs, we may assum e that the e x p o rte rs' fru stratio n will be la rg e r than the observed decline by the am ount of the expected annual in crease. In other w ords, the size of the d is turbance is not only the difference X - X , but this difference increased by O u the am ount of fru strate d expectations, e . g . , d = I x - X - b | . It is c le a r that our index belongs in the second g eneral category of in stability indexes and also that its lim its a re zero and infinity. Of co u rse, this index w ill be z e ro in the case of an export s e rie s having a p erfect lin ear trend, independently of its slope. The Role of Expectations in Export Instability A fourth category of instability indexes which has not been explored in the lite ra tu re of export instability, and which we feel is pertinent in the con text of export instability with reg ard to econom ic developm ent, is indexes 34 involving expectations. The obvious reason fo r the appropriateness of an instability index involving expectations in a study of export instability is the fact that one of the a p rio ri argum ents with reg ard to the effect of export in stability on econom ic developm ent is that such instability precludes any suc cessful planning on the p a rt of the LDC’s, since planners cannot clearly antic ipate the foreign exchange earnings of the country. We feel, therefo re, that an export instability index based on som e kind of expectations is an appropriate m easure to be used fo r the verification o r refutation of such a p rio ri hypothe se s with reg ard to the effects of instability on developm ent. Expectations h ere a re seen as attitudes and judgm ents with regard to future p ersp ectiv es, and as such, they have two dim ensions; uncertainty and tim e. In our study, since we a re concerned with the sh o rt-ru n type of in stability in exports, we w ill consider the tim e horizon as fixed at one y ear, and we w ill tu rn our attention m ainly to the uncertainty. In oth er w ords, we w ill try to determ ine the rationale o r logic behind an expectations model which m ight explain in a sim ple fashion how expectations a re shaped in the minds of governm ent planners o r exporters. Now, on an a p rio ri b asis, it m ight be expected that two types of forces would shape the expectations fo r next y e a r’s exports in the minds of planners o r ex p o rters. The firs t one would be the in e rtia around a "norm al" o r "average" level and the o th er the "dynam ic" force of a recen t change in the level of exports. This general fram ew ork in which expectations a re conceived can be ex p ressed in sev eral m athem atical form s. F irs t, th e re a re different ways in which the "norm al" o r "average" level of exports can be determ ined and, second, th e re are sev eral ways in which p ast "tren d s" o r "extrapolations of recen t changes" m ight be taken into account. Inasm uch as our study is in the exploratory stage we w ill not go deeply into the experim entation with regard to the altern ativ e form s in which the above g en era l m odel of expectations can be presen ted . Our intention is not to determ ine the best expectations m odel as such, but to exam ine how an in stability index based on expectations com pares w ith the previously discussed altern ativ es with regard to th eir ability to explain the effects of instability of exports on econom ic development. We have chosen as an "average" o r "norm al" level of expected exports fo r a c e rta in y e a r the average of the exports of th e th ree previous y ears. This "norm al" level is then adjusted according to the change which took place in the previous y e a r's exports and which m ight be considered as the ex tra polating dynam ic force. The m athem atical form used fo r the expected amount of exports fo r a p a rtic u la r y e a r t can be w ritten as follow s: xf = x t * + b (x t_1 - x t. 2) E w here: is the expected am ount of exports at y e a r t, * X is the average exports of the previous th ree y e a rs from 1 -1 36 b is a coefficient of adjustm ent. In other w ords, b is the degree o r intensity with which la st y e a r's change in exports will be considered by the exporters or planners in the shaping of th e ir expectations. Now, the g re a te r the b the stro n g er the dynamic o r extrapo lating force would be in the shaping of expectations. * Of co u rse, the adjustm ent coefficient m ust depend on p ast experience and consequently it m ight be defined in sev eral ways. We have chosen to define b as the coefficient which m inim izes the sum of squared deviations of observed and expected exports, that is I (Xt - X^5 )2 = Minimum Since then E (Xt - X f )2 = E <X t-X *)2 - 2b 2 < Xt"X*)(Xt_1 - X ^ ) + b2 1 ^ , - X ^ — - -2 £ (Xr xt><xt-rxt-2) + 2b E < x t- i - x t-2> 2 =■0 3b from which b = E (Xt-X*)(Xt _1-X t_2) £ < Xt - l - Xt - 2»2 *Here we clearly m ake use of the concept of "adaptive expectations. 37 The fact that b becom es indeterm inate when th e re is no trend o r change in the export se rie s under consideration does not ren d er our expecta tions m odel u seless. Actually, one m ight argue-that in an expectations model like the one assum ed here, it would be reasonable fo r som e indeterm inancy to exist with regard to the dynamic elem ent in the form ulation of expectations, if th ere is not any change at all. F or in such a case, it would not be c le a r w hether it is the lack of adaptation o r the lack of dynamic elem ents which determ ines expectations. It m ight also be added that inasm uch as no such case has been observed, an objection to our expectations m odel on grounds of indeterm inancy would not have p ractical validity. It is recognized, however, that the assum ption of a constant coefficient of adjustm ent throughout the tim e period under consideration, reg ard less of the way it is estim ated, constitutes a draw back fo r the estim ation of expectations. N evertheless, given the exploratory stage of the concept of expectations in the export instability index construction on the one hand, and the p ossible re duction of our form ula of expected exports into a weighted average of p ast exports on the o th er, we feel that our expectations form ula does not signifi cantly violate the basic theoretical considerations of expectations. It is our objective to com pare only the effectiveness of the m echanical instability index to that involving expectations, reg ard less of how crude such an index m ight be. We mentioned above that our form ula for the expected exports fo r y e a r t can be reduced to a weighted average of p ast exports. This can easily be 38 seen if one expands the form ula of expected exports. E T herefore, in final analysis Xt is nothing but the weighted average of exports of the p ast th ree y e a rs . 1 This is p articu larly im portant since re se a rc h in different a re a s of econom ics supports the hypothesis th at expectations a re r e lated in som e way to the p ast (m ost frequently in th e form of weighted average A nother advantage of our m odel of expected values is that it allows fo r predicted long-run growth. This is con trary to the approach that considers the expected value as revised over tim e at a rate proportional to the difference between expected and actual values. According to the la tte r approach, the ex pected value—being an average of e a rlie r observations—is n ecessarily always *It is also relevant to m ention h ere that the sum of weights is equal to one since; ( 1/3 + b) + (1 /3 - b) + 1/3 = 1. 2 L. M. Koyck, D istributed Lags and Investm ent A nalysis (A m sterdam , N orth-H olland Publishing Company, 1954); P. Cagan, "The M onetary Dy nam ics of hyperinflation, " in Milton Friedm an (e d .) Studies in the Quantify Theory (Chicago, U niversity of Chicago P re s s , 1956); M. F riedm an, A Theory of the Consumption Function (New York: N ational B ureau of Economic R esearch, 1957); and M. Nerlove, "D istributed Lags and Demand A nalysis for A gricultural and other Com m odities, " A gricultural Handbook, No. 141 (U.S. D epartm ent of A griculture, 1958). of p ast experience^ 2 between the low est and the highest observed values. * Our instability index based on expected exports is defined as follows: s ■ - t ( i i ) n-3 t=4 ' X® ' This index is the average of the annual deviations from expected exports, ex p ressed as p ercen t of the expected exports. One disadvantage of this index is that in its estim ation th e re is a loss of th re e observations. However, this loss is over com pensated for by the prop erty of the index to be independent of the underlying trend in the data. In oth er w ords, reg ard less of w hether an export s e rie s has a lin e a r o r expo nential trend, its instability would be zero so long as its tren d is constant 2 throughout the period under consideration. A m o re im portant lim itation of the instability index involving expecta tions, especially in cro ss-co u n try 'co m p ariso n s, is th at it assum es a sim ila r expectation p attern for all countries. This may be a very strong assum ption and if it is not fulfilled fo r a few countries, it m ight reduce o r d isto rt the explanatory pow er of the index. This is why we w ill not m ake extensive use of this index in subsequent testing. *M. Friedm an, pp. 143-144. 2See Appendix A. CHAPTER III PROBLEMS OF MEASURING INSTABILITY: SELECTION OF DATA Before we begin sta tistic a l verification of the effect of export insta bility on econom ic developm ent, two p ertinent problem s w ill be discussed. The fir s t is related to m easurem ent of the export proceeds and the second to the countries to be included in the sam ple. M easurem ent of Export Proceeds The f ir s t problem concerns the unit of m easurem ent of export proceeds. Since export proceeds a re expressed in money te rm s , it is im portant for the purpose of our study to specify the m ost appropriate currency in which we should m e asu re th ese proceeds. This problem n ece ssarily a ris e s because different cu rren cies a re not o r have not been exchangeable a t constant ex change rate s and because not all international tran sactio n s a re m ade in a common currency. The m easurem ent of export proceeds in dom estic currency will not have a significant effect on com parisons with reg ard to the ra te of change fo r any individual country. However, since our intention is to make meaningful international com parisons, we m ust use a m ore o r less common currency fo r export pro ceed s, and as such, we have chosen the U. S. dollar. 40 41 It is tru e that the conversion of dom estic cu rren cies into dollars with the u se of official exchange rate s m ight introduce a la rg e r e r r o r than the one it intends to elim inate. B ecause, exchange controls, p referen tial ra te s, and tran sactio n s with foreign exchange obtained in the free m arket, ren d er the official exchange rates u seless fo r com parison of purchasing power. It is our conviction, how ever, that this is not exactly the case with the value of export proceeds expressed in U. S. d o llars. B ecause, fo r a g reat num ber of countries, invoices of international transactions a re recorded in U.S. d o lla rs, o r som e other strong international currency which is in constant exchange ra te with the dollar. Then the value of export proceeds in dom estic cu rren cy is estim ated eith er by using the official exchange ra te o r by adding the dom estic curren cy equivalent fo r the different categories of exports using the p ro p e r differential ra te s. This is evidenced by the fact th at in the U. N. T rade S tatistics Yearbook, as well as in other sources of international trad e sta tistic s (for y ears during which exchange ra te s w ere varying o r m ultiple exchange rate s existed). The "conversion fa c to rs" a re estim ated from the relationship between total value in d o llars and the value in dom estic currency equivalent. An additional difficulty which is encountered with reg ard to m easu re m ent of export proceeds, especially in m aking international com parisons in purchasing pow er obtained through exports, is w hether o r not exports should be 42 adjusted fo r changes in the purchasing power of export proceeds. As mentioned b efo re, 1 one of the a p rio ri assum ptions with regard to the effects of export instability on economic developm ent claim s th a t the m ain effect of such in stability is exercised through the destabilizing results on the capacity to im p o rt of the exporting country. T herefore, an attem pt fo r verification or refu tation of such a hypothesis m ust be based on data m easuring im port capacity in te rm s of constant purchasing power. Such an adjustm ent, however, would not only req u ire the availability of a satisfacto ry p rice index of the com m odities involved in international trad e, but a sep arate index for each country as well. That is , it would req u ire an index fo r each country which would m easu re th e p rice changes of all the com m odities im ported by the country. Such an index would be, of co u rse, the unit value index of im ports of the country. U nfortunately, such indexes do not ex ist for many of the le ss developed countries, and a good num ber of those that do exist a re not reliable. This m eans th at the co rrectio n of the existing data with such indexes would probably 2 introduce m ore bias in the com parisons than it would elim inate. We believe, however, that the om ission of such a correction does not ren d er our findings m eaningless. This is because, as stated above, we have *See C hapter I, Introduction, pp. 5-6. 2 If reliable inform ation existed, an ideal co rrectio n would have been th e m ultiplication of each country's export proceeds by its "term s of trad e. " 43 chosen to m easu re export proceeds in U. S. dollars. Consequently, we feel th a t this in som e way may be a satisfactory alternative fo r making meaningful international com parisons of purchasing power. F irs t, because the U. S. d o llar has not been devalued during the period under consideration, and second, it is not fa r from tru e that two countries with equal amount of dollars from export proceeds have equal purchasing power of foreign goods. Some question may a ris e w ith reg ard to possible bias in the published international trad e sta tistic s due to (the exclusion of) smuggling. Unfortu nately, no inform ation is available regarding the rela tiv e size o r direction of contraband. Deductively, one m ay argue that sm uggling m ight be relatively m o re im portant in the trad e of the le ss developed countries because of the probable lack of b o rd er policing and system atic custom detection o r because of the am ount of th e ir trade. Although today's high ta riffs in alm ost every country m ay induce num erous individuals to re s o rt to "am ateu r" smuggling, organized contraband on an international scale seem s to deal alm ost entirely in prohibited goods such as a rm s, drugs, and diam onds. However, it is in the le ss developed countries w here, due to sev ere shortages of foreign exchange, there a re d ra stic im port re stric tio n s on luxuries such as w atches, jew elry, cosm etics, and nylons, combined with extrem ely high sales taxes on such im ported item s th a t m ake smuggling profitable and w orth a risk. 44 It is conceivable, th erefo re, that the published im port figures fo r a larg e num ber of less developed countries may underestim ate th e ir real im p o rts. N evertheless, as fa r as the bias that may be introduced in the export instability indexes from the exclusion of contraband is concerned, we feel that it is of no significance. F irs t, it does not seem plausible that th ere m ight be a significant am ount of sm uggling in exports of p rim ary products in the le ss developed countries, except possibly of som e staples between b o rd er te rrito rie s . Of cou rse, this is not tru e fo r those countries exporting drugs. Second, if th e re a re export restric tio n s because of an international agreem ent such as a buffer stock agreem ent o r because of a u n ilateral p rice stabilization schem e, such re stric tio n s a re norm ally put into effect when the demand for exports and international p ric e s are very low. In such c a se s, it does not seem reasonable to expect any significant export smuggling. On the con trary , when world demand and p rices a re high it does not seem logical that th e re would be any sev ere export restric tio n s to induce sm uggling and the only reason for which sm uggling m ight be attractiv e in such a case m ight be an appreciable difference between dom estic and world p ric e s. Third, excessive export duties that also may induce export smuggling do not seem to prevail when the demand for exports and p ric e s a re low. It is m ore plausible, however, to expect high export duties and possibly som e smuggling only when w orld demand and p rice s a re relatively high. 45 From the above one may suspect that, if any bias is involved in the estim ation of the export instability indexes for the less developed countries because of the exclusion of smuggling, it would be ra th e r toward underestim a tion of that instability. As far as the developed countries a re concerned, one could reasonably expect that the value of smuggling in th e ir trade on the one hand constitutes but a very sm all portion and on the other that it does not fluctuate enough to in tro duce an appreciable bias in the estim ated instability indexes. Selection of Countries The second problem we a re faced w ith is that of selecting the countries to be used fo r the sta tistic a l verification of the effects of export instability. The p ro p er selection of countries is p articu la rly im portant due to the fact that the stro n g er argum ent on the resu lts of export instability considers the lack of im port capacity as the factor m ost responsible fo r affecting economic develop m ent. Hence, if in the attem pt to te st the effects of export instability we in clude countries whose im port capacity does not depend on th e ir exports, we m ight expect the obtained resu lts to be distorted. F o r selecting the countries whose im ports depend in some significant way on th e ir exports, we used reg ressio n analysis. That is, we reg ressed im ports on exports for all countries fo r which data w ere available. N everthe le ss, since we a re dealing with tim e se rie s data, sim ple reg ressio n s e s ti m ated w ith the ordinary le ast squares m ethod m ost likely would have yielded 46 m isleading resu lts due to possible lack of independence of the random e rro rs . We here make reference to the known problem of se ria l co rrelatio n of the random disturbances inherent in tim e s e rie s an aly sis. 1 It is known that the le a st squares method yields b est lin ear unbiased estim ates only under the assum ption that the random e rro rs a re independent. If this assum ption is not tru e and the e r r o r te rm s a re indeed autocorrelated, then the least.sq u ares estim ates, although they a re not biased, a re not the best estim ates of the population p a ra m e te rs. That is, th e ir variances a re la rg e r than n ecessary. The m o st im portant consequence resulting from the u se of the le a st squares method fo r the estim ation of population p aram eters if random e rro rs are not independent, is that the sam pling variances of the obtained 2 estim ates a re m ost likely underestim ates of the actual variances. This im plies that we m ight obtain significant re g re ssio n coefficients in cases w here 3 this is not true. *We m ight note here th at th ere is not com pletely uniform term inology on this subject. In m ost c a se s, however, the term "autocorrelation" o r "auto correlated resid u als" re fe rs to the problem in relation to the population, while the term "se ria l co rrelatio n " is related to the sam ple. 2 H. Malinvaud, S tatistical Methods of E conom etrics (Chicago: Rand McNally & C o ., 1966), pp. 420-421. 3 It should be stre s se d h ere that these consequences a re not the re su lt of the existence of autocorrelation in the independent variab les included in the reg ressio n equation. 47 The m ost reasonable factors which m ake us believe, on an a p rio ri b asis, that th ere would be autocorrelated resid u als in the reg ressio n equation of im ports on exports a re: • F irs t, the obvious om ission of other econom ic o r noneconomic factors affecting o r determ ining the level of yearly im ports in addition to the y early exports. Off hand, such factors a re incom e, foreign exchange re se rv e s, im port duties and re stric tio n s, level of technology, im port p ric e s, and production m ethods. Most of the previously m entioned economic tim e s e rie s a re believed to have som e positive autocorrelation, high o r low. By om itting these facto rs from the attem pted estim ation of the relationship between im ports and exports it is reasonable to expect that the resulting com posite e r r o r te rm s would be significantly autocorrelated. However, even if the om itted factors a re not highly autocorrelated, as is probably the case of im port duties, it is still p ossible that th e ir om ission m ight introduce autocorrelation in the e r r o r te rm s over tim e. * • Second, the possible e rro rs of m easurem ent in the dependent v a ri able. This is because such e rro rs a re m ostly e rro rs of coverage *Wold, fo r instance, found that while som e factors may be com pletely random , th e ir general influence over tim e may be system atic. H erm an Wold, A Study in the A nalysis of Stationary Tim e S e rie s. 2d ed. (Stockholm: A lm qvist and W iksell, 1954), pp. 171-174. 48 and, as such, a re expected to be com m itted in successive y e a rs, and a re also expected to follow the sam e p ro p erties of the tim e s e rie s of the dependent v ariab le with regard to autocorrelation. Since, as indicated above, th e re w ere strong reasons to believe--and actual estim ates have shown that th e re is significant autocorrelation in the resid u als of the estim ated relationship between im ports and exports of the countries under study--w e trie d to avoid the estim ation e rro rs due to auto co rrelatio n in the residuals by using an iterativ e estim ation procedure sug- 2 gested by D. Cochrane and G. H. O rcutt which seeks to m ake the e r r o r te rm random by transform ing the variables according to the au to reg ressiv e stru c tu re assum ed fo r the e r r o r term . This iterativ e estim ation procedure co n sists briefly of the following step s, which a re given h ere in referen ce to the equation we attem pt to esti m ate, that i s : Yt = a + BXt + ut (1) w here Y and X a re the annual dollar values of im ports and exports, resp ec- I v tively, and ut is the e r r o r term . F irs t, equation (1) is fitted to the data with the method of ordinary le a st squares: 1See Table 1. 2 D. Cochrane and G. H. O rcutt, "Application of L east Squares Re g ressio n to R elationships Containing A utocorrelated E rro r T e rm s, " Journal of the A m erican Statistical A ssociation, XLIV, No. 245 (M arch, 1949), 32- 61. 49 TABLE 1 OLS REGRESSIONS OF IMPORTS ON EXPORTS S. E. of Country Constant (t-value) b (t-value) i? Estim ate V D-W Statistic A rgentina 1084. 95 (4. 83) .08 (.43) -. 047 201.30 1.68 A ustralia -278.31 ( .88) 1.21 (8.31) .791 318.19 2.26 A ustria 24.36 (.42) 1.25 (21.40) .962 110.32 0 0 + Belgium 78.52 (1. 06) 1.03 (52. 62) .994 126. 83 2.30 Bolivia 43.08 (2.57) .41 (2.56) .236 18.03 . 50+ B razil 528.07 (1.17) .62 (1.98) . 139 239.53 1.61 B urm a -3.78 (.04) .89 (2.25) .185 45.30 1.14* Canada 239.13 (.82) .99 (19.61) 9.55 388.96 . 96+ Ceylon 203.39 (1.61) .39 (1. 12) 0.14 49.87 1.56 Chile 22.42 (.41) .87 (7.94) .775 66.35 1. 03+ Colombia -11.41 (. 12) 1.05 (5. 60) .628 74.15 1.40 C osta Rica -40.62 (3.19) 1.66 (11.17) .873 14. 12 1.84 TABLE 1—Continued 50 Country Constant (t-value) b (t-value) S. E. of E stim ate S yx D-W Statistic Cyprus -20.90 (1.17) 2.27 (6.77) .713 19.72 . 79+ D enm ark -47.77 (1. 01) 1.23 (37.21) .987 79.25 2.04 Dominican Republic -22.71 ( .73) 1.08 (4. 65) .534 31.02 1.50 Ecuador 4.72 (.53) .77 (11. 02) .870 11.88 1.78 Egypt -. 90 (. 004) 1.39 (3.23) .343 132.25 . 64+ El Salvador -35.26 (4. 91) 1.23 (21.16) .961 10.31 1. 20* Ethiopia -21.18 (2.80) 1.46 (11.62) .882 13.34 1.17* Finland -117.24 (3.44) 1.25 (22.65) .966 73.53 2.77* France 402.98 (1. 15) .99 (17. 53) .945 602.71 . 99+ Germany 319.42 (1. 02) .90 (30.23) .981 729.40 1. 03+ Ghana -184.22 (2. 01) 1.81 (5.01) .572 61.75 1.47 G reece 24.46 (.45) 2. 80 (9.66) .837 107.44 . 63+ G uatem ala -3 .5 8 ( -23) 1. 12 (8.73) .807 21.93 GO 00 + 51 TABLE 1— -Continued Country C onstant (t-value) b (t-value) R2 S. E. of E stim ate Syx D-W Statistic Honduras -5 .8 1 ( -71) 1.08 (11.32) .876 9.77 1.30* Iceland 19.68 (4.26) .96 (15.70) .932 7 .90 1.32* India -1396.93 (2.59) 2.46 (6. 32) .688 304.70 1. 08+ Indonesia 277.63 (2 .12 .47 (2.93) .296 127. 63 1.69 Iraq 95.60 (3.80) .39 (8.77) .808 50. 11 1.17* Ireland 174.64 (3.77) 1.22 (10.97) .869 70.72 . 97+ Isra e l 235.43 (12.57) 1.35 (16.25) .939 47.58 1.61 Italy 700.22 (3.12) 1.05 (17.82) .946 533.18 1 . 13+ Jam aica 20.28 (2.47) 1.20 (20.44) .959 16.20 1.27* Japan 768.28 (3.80) .96 (20.27) .958 537. 78 1.53 Kenya 89.14 (4.08) 1.00 (4. 88) .573 29. 89 1.47 Libya 66.16 (4. 90) .35 (8. 80) .809 51.55 . 24+ 52 TABLE 1—Continued Country Constant (t-value) b (t-value) S. E. of E stim ate Syx D-W Statistic W. M alaysia 179.60 (1.52) .52 (3. 39) .411 104.34 . 54+ M auritius 19.97 (1.91) .64 (3.68) .410 10.18 1.19* Mexico -126.49 (1.35) 1.47 (12. 68) .899 106.04 1.29* M orocco 354.27 (7.22) .25 (1. 69) .093 49.78 .81* N etherlands 222.14 (1.18) 1.15 (22.81) .966 354.26 1.50 New Zealand -55.43 (.56) 1.05 (8. 37) .793 87.10 2.42 N icaragua -.9 7 (.25) 1.19 (22.67) .966 8.05 1.98 N igeria -48.19 (.78) 1.13 (8. 39) .794 84.28 . 57+ Norway 103.95 (2. 03) 1.50 (25.71) .973 82.66 1.16* P akistan 160.43 (.71) .91 (1. 86) .121 227.50 . 44+ Panam a 22.38 (2. 20) 2.62 (11. 56) .886 19.06 . 55+ Paraguay 7.13 (1.18) .66 (4. 04) .460 5.51 1.31* P eru 46.76 (1.73) .94 (14.42) .920 51.98 1.31* 53 TABLE 1—Continued S. E. of Country Constant b E stim ate D-W (t-value) (t-value) S Statistic yx Philippines 294.88 (4.50) .70 (5.56) .624 88.26 1.59 Portugal 27.03 (.67) 1.53 (13. 19) .906 63.91 . 99+ Spain -1045.17 (5.35) 3.61 (11.98) .888 307.55 1.53 Sudan 13.10 (.30) .97 (3. 77) .424 47.03 1.57 Sweden 78.17 (1.40) 1.06 (46.61) .992 92.74 1.54 Switzerland -186.32 (1. 83) 1.34 (24. 07) .970 172.46 + 0 0 0 0 Syria 60. 86 (1.51) .97 (3. 45) .421 33.08 . 91+ Tanganyika 46.43 (4. 38) .46 (5. 90) .653 14. 85 1.61 Thailand -278.15 (4. 46) 1. 81 (12. 08) .890 79.77 1.57 Trinidad and Tobago -17.61 (1. 87) 1.09 (30.67) .981 17.18 1. 31* T unisia 167.62 .20 - . 043 33.83 C O + Turkey 12.10 (. 17) 1.39 (6. 64) .705 64.74 1.51 United Kingdom 1441. 00 (2.54) 1. 05 (18.45) .950 601.43 2.17 54 TABLE 1—Continued Country Constant (t-value) b (t-value) i 2 S. E. of E stim ate S yx D-W Statistic United States -583.15 (.48) .81 (12. 37) .894 1545.00 1.36* Uruguay 159.41 (3. 85) .27 (1.23) .028 42.11 2.04 V enezuela 377.74 (2 . 00) .37 (4.19) .479 233. 05 < 1 0 0 + Yugoslavia 104.11 (3.06) 1.21 (20. 53) .959 79.81 2.15 Source: IM F, International Financial S tatistics, various issu es. Note: + indicates positive autocorrelation at 5 p e rc e n t. * indicates indeterm inate situation with reg ard to autocorrelation at 5 percent. and from (2) the residuals u = (Y -Y ) (t= l,2 , ..., n) a re calculated. v V I Second, assum ing the resid u als in (1) a re generated by a firs t o rd e r autore g ressiv e schem e: ut = eV i + E t < 3 ) w here c. is a random disturbance and p the autocorrelation coefficient, an estim ate r of p is obtained by the ordinary least sq uares method applied to the reg ressio n flt = ' “ t - i + ct • (4) w here e . is an estim ate of e . T hird, the estim ated autocorrelation V I coefficient r is used to tran sfo rm the original variables Y and X into t I x! = Y - r Y and x' = X - r X . The transform ed v ariables a re su b se- V I V "X v l “ l quently used to estim ate: » Y^ = a f + b x ' t + iit (5) Fourth, a second estim ate of r is calculated from the resid u als of (2) using the revised estim ates of a 1 and b from (5). The above iterativ e procedure is con tinued until the estim ates of P converge, that is , until two successive r 's a re Since our observations a re annual, a firs t o rd e r au to reg ressiv e p rocess is probably sufficient to re p re se n t the autocorrelation of the e rro rs . Sargan, fo r instance, found that fo r the data he used, his te st "gives no reason to p re fe r a higher o rd e r au to reg ressiv e equation to the firs t o rd e r equation. " J. D. Sargan, "Wages and P ric e s in the United Kingdom: a Study in Econo- 56 not significantly different. The reason an d /o r th e purpose of this iterativ e estim ation procedure will becom e c le a r from the following b rie f explanation. If we lag (1) by one period and m ultiply it by p we get: pYt - i ap + 6p Xt - l + pUt - l ^ Subtracting (6) from (1) we obtain: Yt ' pYt - l = « - ap + 6 ( ^ - PXM ) + ut - PU t-1 (7) = a' + 6 (Xt - P X w ) + ct by v irtu e of (3), where a' = a - ap = ( l - p)a. Now (7) can be also w ritten as follow s: Yt = + pY t - l + 6 Xt ‘ epXt - l + e f (8) To estim ate (1) under the assum ption (3) we m ust find the estim ates of 3 and p which m inim ize the sum of squared resid u als in (8), th at is the estim ates which make the z ( E ,)2 = z (Y - a' - pY - 6 X - 6pX )2 , minim um . v t V ™ 1 X T ~ J. To obtain these estim ates of 6 and p , we have to differentiate this sum of squared residuals, le t's call it Q, with resp ect to 8 and p and se t the firs t derivatives equal to zero. This p ro c e ss, n ev erth eless, leads to m e tric M ethodology," in Econom etric A nalysis fo r National Economic Plan ning, P. E. H art, G. Mills and J. K. W hitaker (e d s .) (London: B utterw orths, 1964), pp. 25-54. 57 nonlinear equations in g and p . * To g et around the difficulty presented by the nonlinearity in the coefficients 3 and p , the above described iterativ e estim ation p ro cess has been followed, which is equivalent to setting P con stant and solving only with resp ect to g . Then, using the estim ated value of 3 found from the fir s t solution, it is po ssib le to solve fo r the value of p which m inim izes the sum of squared resid u als Q. Each stage in this iterativ e p ro cess req u ire s only the solution of a system of lin e a r equations and it takes 2 usually a few iterations fo r the convergence of the estim ates of 3 and p . Sargan has shown that this type of iterativ e solution will always lead to a con verging value of Q, at which Q is m inim um with resp ect to 3 and p s e - 3 parately. The estim ated values of g and p m ight, however, correspond to a saddle point or local minim um of Q in the ca se Q has m o re than one m inim a. F o r th e estim ation of the reg ressio n coefficients of (8) following the iterativ e p ro c e ss, we used a com puter p rogram known as AUTOECON devel oped by the F ederal R eserve Bank. This program is arranged in such a way as to stop eith er when it find two successive r 's being different by less than .001, o r after twenty iterations. 1 2 See Appendix B. Malinvaud, pp. 442-443. g Sargan, pp. 29-30, and Appendix A, pp. 49-51. 58 In using the com puter program to estim ate the relationship between im ports and exports fo r the countries under study, it com pleted 20 iterations in only two out of 68 cases. * In these two cases the r 's reached at the tw entieth iteratio n w ere never ap art from the previous r 's by m ore than . 017. When the value of the estim ate of p is found to be close to one, then it is n ecessary to suppress the constant te rm , since in such a situation the au to reg ressiv e schem e (3) reduces sim ply to: u. = u + e (3a) t t-1 t ' and (7) becom es: V Y t-i - e(xt-xt_x) + et m which is nothing else but: AYt = 6 AXt + e t . (7b) In such cases we a re able to obtain b est unbiased estim ates of 6 by simply reg ressin g the fir s t differences of im ports on the f ir s t differences of exports. From Table 2 data one can see that th ere w ere only five ca se s out of 68 for which it was n ecessary to suppress the constant te rm . As can be seen from Table 1 data, which show re su lts of the sim ple reg ressio n s of im ports on exports fo r 40 out of 68 co u n tries, that is fo r about *The respective countries a re G reece and G uatem ala. 59 TABLE 2 SUMMARY RESULTS OF REGRESSIONS OF IMPORTS ON EXPORTS CORRECTED FOR AUTOCORRELATED RESIDUALS Country Constant (t-value) Explanatory V ariables x t Xt-1 (t-value) (t-value) -2 R S. E. of E stim ate syx D-W Statistic Argentina* 1293.70 (5. 62) 12“ (.60) -.0 3 3 181.60 1.90 Argentina 1023.73 (4. 47) . 12“ (.59) -. 134 181.69 1.81 A ustralia -250.27 ( -82) 1.21 (8.78) .772 322.42 2.11 A ustria -307.72 (2.61) 1.53 (15. 74) .989 59.81 2.12 Belgium 30.74 1.04 (72.69) .995 109.79 2.02 Bolivia 48. 97 (2.42) .46 (3.19) .655 12.25 2.35 B razil 268.51 ( -54) .83 (2. 34) .247 222.08 1.88 Burma* 52.34 ( -65) .6 9 “ (1.97) .361 38.15 2.11 Burm a 72.98 ( -80) .5 9 “ (1.51) .298 40.01 2.17 Canada 621. 42 (1.36) .93 (12.97) .967 322.88 2. 14 Ceylon 66.68 ( .62) .78 (2.60) .213 44.77 1.90 Chile 323.74 i O .42 / o .846 53.80 2.32 60 TABLE 2—Continued Country C onstant (t-value) Explanatory V ariables Xt x t - l (t-value) (t-value) S. E. of E stim ate syx D-W Statistic China 76.44 (4. 48) 1. 15 (15.56) 954 29.42 1.93 Colombia 73.60 (.71) .90 (4.30) 509 82.21 1.82 Costa Rica -41.59 (2. 73) 1.67 (9. 59) 859 14.42 2.00 Cyprus -25.58 (1.03) 2.47 (5.17) 760 18.13 1.99 Denm ark -63.55 (1.30) 1.23 (36.84) ,987 79.71 2.03 Dominican Republic -11.48 ( -42) 1.01 (4.87) 380 35.73 1.89 Ecuador Egypt 2.72 (.25) suppressed .79 (9. 32) 1.40 (11. 10) 859 12.03 1.99 ,660 98.21 2.07 El Salvador -45.48 (5.35) 1.31 (19.73) ,972 8.63 2.21 Ethiopia suppressed .24 (2 . 12) 979 5.52 2.00 Finland -204.92 (5.63) 1.28 (32.73) ,971 67.81 2.25 F rance -129.97 (.25) 1.07 (13.71) 963 491.63 1.56 61 TABLE 2—Continued Explanatory V ariables Country C onstant (t-value) x t (t-value) x . - i (t-value) S. E. of E stim ate S vx Germany 125.00 (. 22) .91 (18.97) .984 641.64 Ghana -195.25 (.83) 1.89 (6.93) .755 44. 51 G reece 5.78 ( -03) 3.19 (4. 54) .892 87.74 Guatemala* 144.51 (2. 09) 181.97 (2. 04) . 32 0 1 (1.27) .2 9 “ (.89) .895 .892 15.94 16.15 Honduras 7.28 (. 74) .93 (8.07) .901 8.48 Iceland 20.39 (3. 34) .96 (12.46) .941 7.50 India -180.63 ( -30) 1.66 (4. 20) .788 256.32 Indonesia 257.09 (1. 92) .50 (3.12) .370 118.63 Iraq 87.56 (3. 23) .43 (9.14) .919 32. 53 Ireland 132.66 (2. 08) 1.30 (9. 03) .919 57.09 Isra e l 240.01 (8.71) 1.32 (11.81) .931 49.65 Italy 922.19 (2.31) .99 (10. 74) .954 490.91 D-W S tatistic 1.90 1.72 1.89 1. 21* 1.44* 2.40 1.84 2.04 2.02 1.71 2.02 1.62 1.68 62 TABLE 2—Continued Country Constant (t-value) Explanatory x t v (t-value) V ariables Xt -1 (t-value) S. E. of E stim ate S yx D-W Statistic Jam aica 27.93 (2.93) 1.24 (17.28) .943 18.77 1.84 Japan 852.37 (3.84) 1.11 (18. 93) .961 506.74 1.94 Kenya 83.93 (2.48) 1.11 (4. 01) .755 20.08 1.23“ _ . . # Libya suppressed .0 7 “ (1.16) .971 20.37 2. 98* Libya suppressed - . 12 0 1 (1.69) .972 19.92 3.16* W. M alaysia 890. 92 (2.57) .25 (2. 99) .797 57.32 1.83 M auritious^ 76.06 (5.07) - .0 8 “ ( .66) .775 6.17 2.46 M auritious 52.72 (4.21) .2 3 “ (1.97) .814 5.60 2.58* Mexico -156.68 (2. 28) 1.60 (17.94) .936 79.11 1.76 M orocco -116.87 (. 56 1.34 (3.15) .503 37.04 1.31* N etherlands 102.58 (.43) 1.18 (19. 35) .969 335.47 2.01 New Zealand -75.71 (1.09) 1.12 (12.29) .860 68.37 2.20 N icaragua -1.89 (.60) 1.19 (28.61) .967 7.84 1.84 63 TABLE 2— Continued Country Constant (t-value) Explanatory V ariables X X , t t-1 (t-value) (t-value) R2 S. E. of E stim ate S yx D-W S tatistic N igeria* 509.34 (2.67) .3 4 “ (1.73) .936 45.15 2.27 N igeria 806. 07 (2. 95) .1 4 “ ( .66) .928 47.74 2.14 Norway 61.33 (. 73) 1.67 (16.72) .953 108.94 1.92 Pakistan 484. 74 (1. 44) .72 (2.45) .783 115.46 1.94 Panam a suppressed .5 0 “ (1.45) .978 8.25 2.00 Panam a suppressed - . 1 4 “ (-.38) .976 8.71 1.90 Paraguay 7.59 (. 99) .67 (3.27) .525 5.19 1.97 Peru 35.60 (1. 10) 1.08 (13. 55) .961 35.59 1.57 Philippines 185.90 (2.98) .94 (7.72) 8.31 60.87 2.54* Portugal -45.82 ( • 94) 1.85 (13. 35) .965 39.69 1.90 Spain -979. 39 (3.78) 3.49 (9. 22) . 893 305.02 1.97 Sudan* 221.14 (3. 91) - . 04“ (.23) .0 3 “ (2.04) .709 32.56 1.87 Sudan 231.38 (3. 23) .763 29.42 1.68 Sweden 44. 52 ( -67) 1.07 (40.75) .993 87.72 1.70 64 TABLE 2—Continued Country Constant (t-value) Explanatory V ariables x t x t - i (t-value) (t-value) E 2 S. E. of E stim ates S yx D-W Statistic Switzerland -252.46 (1.49) 1.35 (16. 36) .980 140.72 1.23* Syria* 153.55 (2. 25) .4 7 “ (1.16) .446 29.04 2.15 Syria 248.17 (3.33) - .0 6 “ (.15) .388 30.53 2.44 Tanganyika 43.06 (3. 25) .48 (5.14) .641 15.08 1.77 Thailand -351.44 (5.57) 2.10 (13.22) .865 85.64 1.84 T rinidad and Tobago -12.27 (1.54) 1.15 (36.44) .986 14. 54 2.15 Tunisia 438.41 (4. 41) - . 70 (3.04) .806 14. 97 1.89 Turkey - 22.66 ( -24) 1.48 (5. 59) .714 63.08 1.79 United Kingdom 1420. 36 (2. 48) 1.05 (18.52) .946 617.00 2.02 United States 471.93 ( -32) .77 (9. 94) .906 1406.83 1.82 Uruguay 107.31 (2.76) .55 (2.71) .280 37.25 1.84 Venezuela -24.23 (. 04) .54 (2. 25) .658 188.56 1.79 65 TABLE 2—Continued Country Constant (t-value) Explanatory V ariables X X t t-1 (t-value) (t-value) R2 S. E. of E stim ate S yx D-W Statistic Y ugoslavia 116.05 (4.46) 1.20 (26.62) .969 69.83 2.23 Source: IM F, International Financial S ta tistic s, various issu es. Note: # indicates country fo r which no significant reg ressio n coefficient of im ports on exports was found. + indicates positive autocorrelation a t 5 percent. * indicates indeterm inate situation w ith regard to autocorrelation at 5 percent. a indicates not significant a t 5 percent. 66 59 p ercen t of the countries included in our sam ple, the D urbin-W atson statistic j indicates positive autocorrelation o r an indeterm inate situation. j Comparing these resu lts with those found from the application of the ite ra tiv e solution shown in Table 2, it can be seen that only th ree countries a re le ft with indeterm inate situation with reg ard to autocorrelation at 5 percent level of significance. * It is c le a r, th erefo re, that the transform ation of the v ariables ac cording to the firs t o rd e r au to reg ressiv e schem e nearly elim inated the problem of autocorrelation in the resid u als in the relationship between im ports and exports. Consequently, in selecting countries to be used fo r fu rth er testing the effects of export instability, the inform ation obtained from Table 2 ra th e r than that from Table 1 should be used. The p rim ary c rite rio n fo r the selection of countries to be used in our subsequent testing was the significance of the reg ressio n coefficient of im ports on exports of the sam e y e a r o r on exports of the previous y e a r. This is because fo r som e countries, due probably to stru ctu ra l fa c to rs, it seem s that im ports depend m ore heavily on exports lagged by one y e a r than on the exports of the cu rren t y ear. Table 2 data, th erefo re, show the b etter of the two re g re s sions, except fo r the countries fo r which no significant dependence of im ports * These countries are Guatem ala, M orocco, and Switzerland. 67 on exports was found, in which case we have included both reg ressio n s—that is , the reg ressio n of im ports on cu rre n t y e a r’s exports and that of im ports on exports lagged by one year. As can also be seen from Table 2 data, no significant relationship be tween im ports and exports was found for nine countries included in our sam ple. The countries for which no significant reg ressio n coefficient of im ports on exports was found at 5 p erce n t level of significance a re A rgentina, B urm a, G uatem ala, Libya, M auritius, N igeria, Panam a, Sudan, and S yria. 1 This m ight be due eith er to sev ere re stric tio n s on im ports in these countries o r to the existence of other so u rces of foreign exchange such as tourism , foreign aid, and ro y alties, which allow for im ports to vary alm ost independently of exports. H ere, however, we w ill not attem pt to explain system atically the reasons fo r which such independence between im ports and exports exists in th ese countries. Such an attem pt is outside the scope of our study and could very well be the subject of in terestin g case studies. In addition to the above-m entioned nine countries, we have also ex cluded China (Taiwan) from our sam ple because of the la rg e proportion of foreign aid it received throughout m ost of the period under consideration (well over 30 percent of her im ports). 1In o rd er to te st the sensitivity of the resu lts of this p a rtic u la r decision to re s tric t the num ber of countries included in the sam ple, we subsequently com puted com parable re su lts when all these nine countries w ere included. CHAPTER IV STATISTICAL TESTING OF THE EFFECTS OF EXPORT INSTABILITY Having explained the m e asu re s of export instability in C hapter n and th e choice of data in C hapter HI, we can now proceed to the final testing of the effects of export instability on econom ic growth. As stated in Chapter HI, fo r the estim ation of instability indexes we have chosen the m easurem ent of export proceeds in U. S. dollars. The various issu es of the International M onetary Fund (IMF) publication, Inter national Financial S tatistics, w ere the so urces of our data. F o r consistency w ith the period fo r which data on incom e in constant p rice s ex ist, the in sta b ility indexes have been estim ated m ainly fo r the period 1953 to 1966. F rom a casual look a t Table 13 in Appendix C, it can be seen th at the export instability indexes of the le ss developed countries a re , in general, higher than are those of the developed countries. * What a casual look does *The distinction between developed and le ss developed countries is m ade solely on the basis of th e ir re a l p e r capita GDP. We have included in th e developed countries those countries from our sam ple whose real p e r capita GDP was $600 o r m ore in 1966. The conversion of local cu rren cies to U. S. d o llars was made with the u se of the official exchange rate s published in the U. N. publication, International T rade S tatistics. 68 69 not reveal however, is that the average instability index (1^ ) fo r the le ss developed countries is approxim ately 10, while that for the developed coun trie s is 5. 3. This im plies that, on the average, the exports of the le ss developed countries fluctuate tw ice as m uch as the exports of the developed ones. With reg ard to the export instability index involving expectations (I ), E the relativ e m agnitudes between less developed and developed countries a re in the sam e o rd e r. 1 Table 13 data also reveal that only five of the le ss developed countries have export instability slightly low er than the average export instability of the developed countries. The m o st serio u s data problem we w ere faced with in conducting a sta tistic a l testin g of th e effects of export instability was the lack of data on incom e, especially in constant p rice s. We have clearly explained in the intro duction why th e use of such data is im perative if findings a re to be valid. Even though we cannot claim that all the data used on incom e a re com pletely hom ogeneous, we a re confident that they a re by fa r b e tte r, fo r the purpose of o u r study, than a re c u rre n t incom e data used by o th er authors. We also feel th at definitional differences o r e rro rs of coverage with reg ard to incom e at constant p ric e s between the different countries included in o u r sam ple a re n eith er extrem ely lim iting nor invalidating to our findings. This is because such differences do not figure v ery im portantly when one is in terested in the 1F o r the LDC, ^ = 1 2 .43 while fo r the DC, I£ = 7. 96. 70 the com parison of the rate of change of the individual country’s incom e ra th e r than in the com parison of the absolute incom e levels of the various countries. F or estim ation of p e r capita incom e figures, we have used re a l GDP data, mainly because this is the m ost available m easure of national incom e in constant p ric e s, especially fo r m ost of the less developed countries. Sources of our income d ata, as explained in Appendix C, Table 14, a re the U. N. National Income Accounts S tatistical Y earbook, various issu e s, and the OECD N ational Accounts of L ess Developed C ountries 1950-1966. The so urces fo r the population figures a re the U. N. Dem ographic Y earbook, various issu e s, and the above OECD publication. To avoid repeating th e m istake m ade in previous stu d ie s1 w ith regard to the estim ation of the grow th rate, we have chosen to estim ate the "per capita real incom e grow th rate " by fitting a logarithm ic trend to re a l p e r capita 2 income figures fo r each country for the period 1953 to 1966. Using the infor m ation of Appendix C—T ables 13 and 14—we have estim ated the re g re ssio n equations of the p e r capita re a l income grow th rate s on the instability indexes fo r the total num ber of countries of our sam ple as well as fo r the developed and le ss developed countries separately. ^ e e C hapter I, pp. 11-12. 2 t The fitted trend line is Yf = ab , w here Y|. denotes the annual re a l p e r capita GDP and t the tim e in y ears. Now, it is obvious that if b = 1 + r, r would be the estim ated o r fitted annual grow th ra te of the re a l p e r capita GDP. As can be seen from re su lts of the re g re ssio n equations presented in Table 3, export instability seem s to have a significant negative effect on the real p e r capita incom e grow th ra te of the countries included in our sam ple. From the determ ination coefficient of Equation (1) we see th at the export in stability index explains about 22 p ercen t of the observed v ariatio n in the grow th rates of the countries under consideration. The im plication of Equation (1) is that a 1 p e r cent in crease of the export instability of a country w ill, on the average, cause a q u a rte r of 1 p e rc e n t d ecrease in the cou n try 's real p e r capita incom e grow th rate . More striking inform ation, however, is provided by Equations (3) and (5), which show the relationship between grow th rate and export instability for the less developed and the developed countries, respectively. The re su lts of these two equations convincingly confirm the a p rio ri hypothesis that export instability is m o re harm ful to the economic growth of the le ss developed countries than it is in the developed ones. Actually, our findings do not show any negative effects of export instability on the economic grow th of the de veloped countries. As the resu lts of Equation (5) show, if th e re is any effect, it seem s to be positive ra th e r than negative. * From the above findings it is also c le a r that the export instability index involving expectations is a less satisfactory explanatory variable than is *These findings com e in com plete co n trast with those of Coppock's, for instance, who states that "the evidence does not indicate any relation between grow th in p e r capita incom e and export instability. " J. Coppock, p. 107. TABLE 3 SUMMARY OF PER CAPITA GDP GROWTH RATE AND EXPORT INSTABILITY REGRESSIONS Equation Dependent V ariable Constant Coefficients of the Explanatory V a rie tie s I ? F -T e st D. F. 1. All Countries x i 4.634 (8.503) -. 258 (3.968) .218 15.75 (1,52) 2. All Countries x i 3.513 (5.064) -.087 (1.348) .015 1.82 (1,52) 3. L ess Developed Countries x i 4.483 (6.144) -.2 5 9 (3.352) .226 11.24 (1,34) 4. Less Developed Countries x i 2.558 (2.849) -. 038 (.492) - . 022 .24 (1,34) 5. Developed Countries x i 3.067 (2.494) .095 (. 433) -.0 5 0 .19 (1,16) 6 . Developed Countries x i 2.937 (2. 522) .080 (.577) -.0 4 1 .33 (1,16) N otes: W here: X j = re a l p e r capita GDP grow th rate , X2 = export instability index (1^), and X3 = export in sta bility involving expectations (Ie ). In all tables that follow, R is the coefficient of determ ination co rrected fo r the degrees of freedom . The figures in parentheses below the reg ressio n coefficients a re the t-v alu es of the coefficients, and below the F -te s t a re the respective degrees of freedom . Source: Tables 13 and 14 of Appendix C. 1 73 the m echanical index. This m ight be due eith e r to the m odel of expectations we have assum ed fo r its estim ation o r to the fact that export expectations do not have any appreciable effect on the grow th ra te of the countries under con sideration. * As was m entioned in C hapter II, this re su lt was not unexpected. The Growth R ate of Exports It has been claim ed that export instability, in addition to its inhibition on the econom ic grow th of le ss developed countries through the destabilizing effects on the capacity to im port investm ent goods vital fo r th e ir economic grow th, has also som e negative effects by being detrim ental to the grow th ra te 2 of exports. To show the validity of this argum ent we felt we m ust show the signifi cance of the grow th of exports on the econom ic grow th and also the effects of th e export instability on the growth ra te of exports. The re su lts of the reg ressio n s of re a l p e r capita incom e grow th ra te on the exports grow th ra te , presented in Table 4, cle a rly show that the la tte r is a very significant facto r fo r the economic grow th of all countries, since, as it can be seen from Equation (1), it explains about 50 p ercen t of the observed v ariatio n in the grow th rate s. It is also noticeable that the growth rate of ■ * T t is obviously very difficult for one to p a ss judgm ent on which of the two reasons m ight be m o re responsible fo r the lack of association between the incom e grow th ra te and the export instability index involving expectations. 2 F o r the theoretical argum ents in support of this claim se e C hapter I, pp. 4-5. 74 TABLE 4 REGRESSIONS OF THE PER CAPITA GDP GROWTH RATE ON THE GROWTH RATE OF EXPORTS Equation Dependent V ariable Constant Coef. of the Explanatory V ariable X4 R2 F -te s t D. F. 1. All Countries X1 .611 .356 .499 53.84 (1.907) (7.337) (1,52) 2. L ess Developed X1 .856 .282 .350 19.88 C ountries (2.380) (4. 459) (1,34) 3. Developed Countries X1 -.5 4 6 .530 .630 29.94 (.693) (5.472) (1,16) Note: W here: = real p e r capita GDP annual grow th rate. X j = annual grow th rate of exports. Source: Table 14 of Appendix C. 75 exports seem s to be a m o re im portant contributing facto r fo r the economic grow th of the developed countries than it is for that of the le ss developed ones, as the results of Equations (2) and (3) indicate. This re su lt m ight be m o re o r le ss expected on a p rio ri considerations. F o r even though it is highly probable that a fast grow th of the export secto r fo r m ost countries will re su lt in a noticeable growth of th eir econom ies as a whole, the degree with which any expansion in the export se c to r would be tran sm itted to the other secto rs depends, as M eier puts it, "on the c h a ra c te r of the country's export b a se , and on the degree of dom estic m a rk e t im perfec tio n s, interpreted in a wide sense. nl It is reasonable, th e re fo re, fo r one to expect that due to the p a rtic u la r type of production fo r exports, which fo r the m ajority of the le ss developed countries is m ainly p rim ary production ch aracterized by lim ited "backw ard" linkage effects and the higher deg ree of m a rk e t im perfections which prevent som e of the potential "forw ard" linkage effects to m a te ria liz e , that the sam e grow th rate of exports would not c re a te incom e grow th in c re ases of the sam e 2 m agnitude as in the developed countries. ^Gerald M. M eier, International T rade and Developm ent (New York: H arper and Row, P u b lish ers, 1963), p. 176. 2 However, it should be pointed out that te sts fo r both the equality of the en tire regression functions (2) and (3) as w ell as for the equality of the re g re s sion coefficients (b) of the sam e re g re ssio n equations do not show any signifi cant difference a t the 5 p ercen t level. The firs t te s t gave a value fo r F = 2.267* 3.18 = F. 05(2,50) and the second, F = 3.622 < 4.03 = F_ 05 (2,50) • F o r the details of these te sts see C. R. R ao, L inear S tatistical Inference and Its Applications (New York: John Wiley & Sons, I n c ., 1967), pp. 237-240. 76 Now, since it has been clearly shown that the growth ra te of exports plays such a significant role in determ ining the economic grow th of a country, the next reasonable step is to investigate the relationship between export instability and the grow th ra te of exports. The resu lts of the reg ressio n equations of the export grow th ra te on the export instability index p resented in Table 5 a re in accordance with the the o retical statem ents that export instability is m ore harm ful to the export grow th of the less developed countries than it is to th at of the developed countries. Indeed, as shown by Equations (2) and (3), the effects of export in stability a re only significantly negative on the export grow th fo r the less developed countries while fo r the developed ones they a re not only insignificant but, in addition, in the positive direction. * Explanation fo r such differentiated effects of export instability between developed and less developed countries can be given with the conventional argum ents which claim that exports in the le ss developed countries a re m ore easily discouraged by the fluctuations. This is because less developed As was mentioned before, we attem pted to te s t the sensitivity of the resu lts presented in Tables 3 and 5, by running sim ila r reg ressio n s with data including seven of the nine countries (for which data w ere available) which have been left out from previous testing. (These countries w ere excluded because of lack of association between th e ir im ports and ex p o rts.) F rom the resu lts p resented below, it is evident that the coefficients and in general the relation ships estim ated are in ferio r to those shown in Tables 3 and 5, p articu larly fo r the le ss developed countries. T herefore, the evidence seem s to justify our approach to exclude those countries. (See C hapter III, p. 6 7 .) However, the c rite rio n for the selection of the sam ple would not seem to have been respon sible fo r our resu lts. (Continued on page 78 . ) 77 TABLE 5 REGRESSIONS OF THE EXPORT GROWTH RATE ON THE EXPORT INSTABILITY INDEX Equation Dependent V ariable Constant Explanatory V ariable * 2 5 F -T e st D. F. 1. All C ountries X4 9.345 -. 475 .180 12.66 (8.371) (3.448) (1,52) 2. Less Developed X4 8.105 -.3 8 9 .096 4.70 Countries (4.779) (2.167) (1,34) 3. Developed C ountries X4 7.078 . 130 -. 053 .15 • > (3.775) (.388) (1,16) Note: W here: X4 = annual grow th ra te of exports. X = export stability index (I ). £ t M Source: T ables 13 and 14 of Appendix C. 78 countries, having few er ways to defend against fluctuations in export proceeds, sh ift th e ir productive reso u rces to other lines of production in which they are m ost probably less efficient. F o r the developed countries, however, insta bility of export proceeds m ight be considered as a stim ulating factor to th eir export grow th, if one reasons along the lines of A. H irschm an on unbalanced growth. * The resu lts presented in Table 6 indicate the com bined effects of export instability and export grow th on the p e r capita incom e grow th rate. The con clusions one can draw from th ese findings of the m ultiple re g re ssio n equations w ith reg ard to the v ariables involved and the significance of th e ir effects shown by th e re g re ssio n coefficients a re exactly the sam e as those draw n from the re su lts of the single reg ressio n s between the individual v ariab les. The only noticeable difference m ight be that the reg ressio n coefficient fo r the export Coef. of Expla- „ ,, Dependent _ . . natory V ariab les F -T e st Equation „ . , . Constant - 2 ■ ■ ^ V ariable X r d . F. 1. All C ountries X1 4.391 -.2 2 6 . 186 14. 73 1 (8.647) (3.839) (1,59) 2. All C ountries X 4 8.855 -.4 1 1 .141 10.86 * ± (8.227) (3.296) (1,59) 3. L ess Developed X1 4.046 -.2 0 6 . 159 8.93 Countries (6.118) (2. 989) (1,41) 4. L ess Developed X4 7.310 -.2 9 3 .049 3. 17 C ountries (4.629) (1.781) (1,41) Source: T ables 13 and 14 in Appendix C. "^Albert O. H irschm an, The Strategy of Economic Developm ent (New Haven: Y ale U niversity P re s s , 1958), pp. 173-175. TABLE 6 SUMMARY RESULTS OF INCOME GROWTH-EXPORT INSTABILITY AND EXPORT GROWTH MULTIPLE REGRESSIONS Coefficients of Explanatory V ariables Equation Dependent V ariable Constant X2 X 3 X 4 R2 F -T e st D. F. 1. All Countries X1 1.731 - . I l l .311 .595 30.32 1 (2.661) (1. 962) (5. 890) (2,51) 2 . All Countries X1 1.456 - . 083 .335 .521 29.83 X (2.616) (1.837) (7.479) (2,51) 3. Less Developed X1 2.622 -. 170 .230 .430 14.20 Countries 1 (3.238) (2. 398) (3.627) (2,33) 4. Less Developed X1 1.924 -. 105 .307 .385 11.94 C ountries X (2.679) (1.702) (4.846) (2,33) 5. Developed Countries X1 -.672 .027 .528 .606 14.09 X (. 649) ( .196) (5.261) (2,15) 6. Developed Countries X1 -. 463 -.0 1 4 .533 .606 14.07 X (. 477) (. 158) (5.221) (2,15) Note: W here: X , X , X , and X denote the sam e v ariables as above. 1 ^ 0 4 Source: Tables 13 and 14 of Appendix C. 80 instability index involving expectations (X^) becom es a little m ore significant than before, but it rem ains le ss significant than the m echanical export in stability index (X ). To this point, our findings indicate that export instability influences econom ic grow th through its detrim ental effects on the grow th of exports for the le ss developed countries. In trying to determ ine additional factors involved in the relationship between export instability and economic grow th, we r e so rted , because of lack of data on investm ent in constant p rice s fo r m ost of the le ss developed countries, to o th er explanatory v ariables on which inform a tion was relatively m ore available. The Size of the Export Sector It has been claim ed th at one of the facto rs that m akes export instability m ore harm ful for the less developed than the developed countries, is the size of th e ir export se c to r relative to th e ir national incom e. S everal authors have suggested th at the sam e degree of export instability would have different effects on econom ic grow th depending on the relative m agnitude of exports to the country’s income. C. W. Reynolds, fo r instance, states that "the key to the relationship between export instability and economic grow th lies in the relationship between the fluctuating value of export earnings and th e changing level and com position of its dom estic sh are. 1C lark W. Reynolds, "D om estic Consequences of Export In stab ility ," A m erican Economic Review. P apers and P roceedings, IIII (May, 1963), 93- 102. M eier also states that "the uncertainty and v ariability of export p rice s 81 To te st the role of the size of the export se c to r in the relationship of export instability and econom ic grow th we employed m ultiple reg ressio n analy sis. As a m easure of the relative size of the export se c to r we have used the average yearly proportions of exports to GDP for the period 1953 to 1966 inclusive. The reason fo r using th e en tire p erio d 's average instead of a single y e a r's proportion is obviously to avoid any potential bias involved in case the chosen y e a r constitutes an extrem ely high o r low proportion in relation to the norm al one. Because of lack of data in constant p ric e s we have estim ated the ratio of exports to GDP fo r c u rre n t p ric e s. The resu lts of m ultiple re g re ssio n equations of the grow th ra te on ex p o rt instability and the size of the export secto r to GDP presented in Table 7 do not indicate any conclusive evidence on the ro le of the la tte r in the relation ship between export instability and econom ic grow th. F or the total num ber of countries as well as fo r th e less developed and the developed countries sepa rately , no significant re g re ssio n coefficient was obtained fo r the v ariab le standing fo r the size of the export secto r. The reason for th e lack of conclusive evidence on the influence of the relativ e size of the export secto r, con trary to the a p rio ri assum ptions, m ight be due to the solely quantitative tre a tm e n t of the sh a re of the exports to the national income in our testing, which im plicitly suggests sim ilarity in the and earnings can certainly have ad v erse effects on the development program of a p a rtic u la r country when . . . the country's exports a re la rg e rela tiv e to its national product. " G erald M. M eier, The International Econom ics of Developm ent (New York: H arper and Row, 1968), p. 266. TABLE 7 SUMMARY RESULTS OF GROWTH-INSTABILITY AND SIZE OF THE EXPORT SECTOR MULTIPLE REGRESSIONS Equation Dependent V ariable Constant Coefficients X2 of Explanatory V ariables X X 4 5 R2 F -T e st D. F. 1. All Countries X1 4.617 258 -.0008 .202 7.72 1 (6.717) (3. 903) (.041) (2,51) 2. All Countries X1 1.783 -. 112 .311 -.0027 .516 19.84 X (2.466) (1.950) (5. 836) (. 170) (3,50) 3. Less Developed X1 4.191 -.2 5 4 .014 .214 5.77 Countries (4. 930) (3.247) (. 6 86) (2,33) 4. Less Developed X1 2.474 -. 168 .227 .008 .416 9.31 Countries X (2.812) (2.340) (3.524) (.461) (3,32) 5. Developed Countries X1 3.801 .051 -.0 2 8 -. 101 .22 (1.987) ( .211) (. 510) (2,15) 6. Developed Countries X1 .239 - . 031 .533 -.0 3 7 .612 9.94 ( .182) ( .218) (5. 347) (1. 108) (3,14) Note: W here: X^, X^, and X^ denote the sam e v ariables as before. X = average ratio of exports to GDP. S ) Source: Tables 13 and 14 of Appendix C. 0 0 to 83 the degree of economic developm ent of the countries under consideration. It is im portant to indicate h ere that it is highly probable that the sig nificance of the sh a re of the export secto r in the national incom e m ight be closely related to the country's stage of economic developm ent. F o r le ss developed countries which a re in a p rim itiv e stage and p re dominantly ag ricu ltu ral, a sm all o r la rg e export se c to r relativ e to th e ir na tional incom e m ight not m ake a significant difference as fa r as the effects of export instability a re concerned for th e ir econom ic growth. This is because in such countries econom ic grow th does not depend heavily on the im ports of capital goods. On the other hand, as a developing country m oves to higher stag es of econom ic developm ent, its further econom ic grow th becom es in creasingly dependent upon its im ports of capital goods, which in th e ir tu rn , rely on the country's capacity to im port—that is , on the country's export proceeds if the country does not have any other source of foreign purchasing pow er. It is w orth noting a t this point that, on the average, the sh are of exports to GDP fo r the less developed countries included in ou r sam ple is not different than that of the developed countries. The av erage sh a re of exports to GDP for 42 less developed countries was found to be 17.14 p erce n t and for the developed ones 17. 54 p ercen t, while th e ir m edians w ere 15. 82 p ercen t and 16. 59 p ercen t, respectively. These estim ates m ay contradict som e g u esses o r estim ates based on one y e a r's inform ation, that fo r the le ss developed 84 countries exports "tend as a rule to be la rg e r in proportion to the national in com e, than in the highly developed co u n tries. 1,1 An additional possible reason for which the sh a re of exports to GDP failed to show any effect on the relationship between export instability and econom ic grow th m ight be the fact that we have estim ated this ratio from data in c u rre n t p ric e s. This could have resulted in downward biased estim ates of the size of the export se c to r fo r the countries with high ra te s of inflation in 2 th e ir dom estic m ark ets, which m ost probably a re the le ss developed ones. Since the resu lts of the reg ressio n analysis did not indicate any signifi cant role fo r the relative size of the export se c to r in the relationship between export instability and economic growth, it m ay be argued that the effects of the s iz e of the export se c to r could be of ordinal ra th e r than cardinal sense. To te s t th is, we partitioned our sam ple of 36 le ss developed countries into th re e and four subgroups according to the relativ e siz e of th e ir export sector. Percentage of Exports to GDP 30+ 20-29.99 10-19. 99 Below 10 20+ Total C o rr. Coef. r 1a -.6 4 7 -. 568 -.4 6 3 -. 266 -.5 9 1 -.498 Rank C orr. Coef. -.7 0 0 -.476 -.3 0 1 -. 017 -.607 -.398 No. of C ountries 5 8 14 9 13 36 in the Group M. M ichaely, Concentration in International T rade (A m sterdam : N orth Holland Publishing C o ., 1962), p. 107. 2 As stated in the Introduction, one of the expected possible re su lts of export instability is dom estic inflation. 85 The re su lts obtained from the partitioning of the sam ple a re revealing. W hether the num ber of subgroups is three or four, the relationship between the export instability index and the real p e r capita incom e growth rate becom es stro n g er—that is, the negative effects of instability are m o re powerful for those countries with the relatively la rg e r export secto r. T hese findings, th e re fo re, support the argum ent stated at the beginning of this section that export instability is m ore harm ful the la rg e r the export se c to r of the country, if the size of the export secto r is understood in the ordinal sense. In passing, it is interesting to m ention h e re that Equation (6) vaguely confirm s H. Johnson’s theoretical m odel according to which, other things equal, a country’s "equilibrium rate of grow th w ill be -higher . . . the lower the ratio of exports to output. | The Role of Foreign R eserves i As has been stated before, one of the facto rs which is considered to be | ' aggravating to the detrim ental effects of export instability on the economic grow th of the le ss developed countries is th e ir lack of adequate foreign exchange re se rv e s. The justification of this argum ent lies in the fact that the higher the level of foreign re serv es available to a country, the m ore insensitive the economy of that country is to the destabilizing effects of export instability, i *H. G. Johnson, International T rade and Economic Growth (Cam bridge, j M ass. : H arvard U niversity P re s s, 1958), C hapter V, p. 127. 86 Since adequacy of foreign re se rv e s seem s to be a rela tiv e m agnitude, we have chosen to approxim ate this m easu re by the average ratio of foreign re se rv e s to annual im ports for the period 1953 to 1966 inclusive. By using the average relativ e size of foreign re se rv e s one can avoid any p ossible biased m easu res due to extrem e conditions which m ight p rev ail fo r one y e a r only, selected even a t random. The data obtained for the countries included in our sam ple, surprisingly enough, do not seem to indicate, on the average, any significant difference in the ratio of foreign re se rv e s to im ports between developed and less developed countries. * The sim ple and m ultiple re g re ssio n resu lts shown in Table 8 indicate that no significant relationship ex ists between the size of foreign re se rv e s and the income grow th rate. If one w as inclined to m ake any inference from the sign of the coefficient of the reg ressio n of incom e grow th ra te on the average 1The average ra tio of foreign liquidity was estim ated sep arately fo r 44 le ss developed countries fo r which inform ation was available and for the 34 of them which a re further used for testin g , as well as for the 18 developed countries. The results follow: A rithm etic Mean ( % ) Median (% ) Range (% ) 44 L ess Developed C ountries 34 Less Developed Countries 18 Developed Countries 42.37 36.82 134.80 43.17 35.64 134.80 45.46 34. 56 126.99 TABLE 8 SUMMARY RESULTS OF INCOME GROWTH, EXPORT INSTABILITY, AND FOREIGN RESERVES REGRESSIONS Equation Dependent Constant Coefficients of Explanatory V ariables i 2 F -T e st V ariable x„ x X „ 2 4 6 D. F. 1. All Countries X1 2.783 -.0015 -.0 1 9 .04 1 (6.738) (. 192) (1,50) 2. All Countries X1 4. 805 - . 258 -.0022 .212 7.86 X (7 .668) (3.959) (. 327) (2,49) 3. All Countries X1 1.570 -. 098 .326 .0015 .578 24.27 X (2.338) (1. 830) (6.593) (.303) (3,48) 4. All Countries X6 46.005 -.2 6 5 -.0 1 9 .04 o (4. 074) (. 195) (1,50) 5. L ess Developed X1 2.167 .0023 -.0 2 9 .07 C ountries X (4. 598) (. 256) (1,32) 6. Less Developed X1 4.680 -.2 7 3 .0009 .246 6.37 C ountries X (5.753) (3.557) (. 123) (2,31) 7. Less Developed X1 2.258 -.1 6 5 .268 .0065 .548 14.35 Countries X (2.768) (2. 579) (4 .666) ( 1.072) - . 029 (1,32) 8. L ess Developed X g 47.594 -. 493 -.0 2 9 .64 C ountries o (2.856) (. 280) (1,32) 00 -<1 TABLE 8—Continued „ .. Dependent Equation . , , V ariable Constant Coefficients of Explanatory V ariables X2 X4 X6 R2 F -T e s t D. F. 9. Developed Countries X1 4.024 -.0009 -. 021 .64 (5.883) (.803) (1,32) 10. Developed Countries X1 3.443 . 116 -.0107 -.0 7 0 .44 1 (2.608) (. 519) (. 837) (2,15) 11. Developed Countries X1 - . 302 .046 .527 -. 0101 .627 10.50 X (. 289) (.350) (5.383) (1.345) (3,14) 12. Developed C ountries X6 35.236 1.932 -.0 5 0 .20 (1.449) (.444) Note: W here: X^, Xg and X4 denote the sam e variab les as before. Xg = average ratio of foreign re se rv e s to im ports. Source: Tables 13 and 14 of Appendix C. 00 00 89 ratio of foreign re se rv e s to im p o rts, one could say m erely that the a p rio ri assum ptions seem to be justified for the le ss developed countries only. For the total sam ple of countries and for the developed countries as well, th ere is a slight indication th a t the relationship between the size of foreign reserv es and the incom e grow th rate seem s to be ra th e r negative. T hese re su lts, specifically the significance of the coefficient of foreign re s e rv e s , do not change with the introduction of the index of export instability and the grow th rate of exports as additional explanatory v ariab les as it can be seen from Equations (3), (7) and (11). One possible reason fo r the lack of any significant influence of foreign re s e rv e s , p articu la rly for the le ss developed countries, m ight be the fact th a t the positive effects of high foreign re se rv e s a re negated by the co st in volved in keeping them . In addition, the above resu lts show no c le a r indication that the countries w ith the higher export instability, generally, hold la rg e r re se rv e s. As one can see from Equations (4), (8), and (12), th e re is no significant relationship between export instability and the ratio of foreign re s e rv e s to im p o rts. 1 Thus fa r we have found no significant evidence of the role of the re la tive adequacy of foreign re se rv e s on the incom e grow th and its relationship to *The rank co rrelatio n coefficient between export instability and the ra tio of foreign re se rv e s to im ports fo r the le ss developed countries is equal to r r = -.2 1 5 and fo r the developed ones r r = . 182. 90 export instability. It m ight be argued that the m ere availability of international liquidity per se is not a stim ulant to econom ic grow th unless it is properly used for counterbalancing the detrim ental effects of export instability. To te st this possibility, we attem pted to use the average y e a r-to -y e a r percentage variation of foreign re se rv e s as a crude m easure of the deliberate policy fo r counterbalancing export fluctuations. The resu lts of single and m ultiple reg ressio n analysis shown in Table 9 clearly indicate that the fluctuations in foreign re se rv e s a re significantly r e lated to export instability and the level of foreign re se rv e s, as expected. 1 How ev er, a m ore careful exam ination of th ese re su lts, especially a com parison of the reg ressio n equations between le ss developed and developed countries shows that only for the form er fluctuations in foreign re se rv e s seem to be directly related to export instability and, in v ersely , to the size of foreign re se rv e s in a significant fashion. F o r the developed countries th e re is no significant r e lationship between export instability and fluctuations in foreign re se rv e s. This j | m ight be a good indication in support of the claim that the developed countries do not suffer dom estically as m uch as the less developed countries from the export instability and th erefo re, it is not necessary for the developed countries to take severe steps for counterbalancing its effects. In addition, they proba bly are in a position to m o re effectively u se pro p er m onetary and fiscal policy ^Coppock on the contrary does not find any relationship between export ; instability and fluctuations in foreign re se rv e s. Coppock, p. 113. TABLE 9 REGRESSIONS OF INCOME GROWTH, EXPORT INSTABILITY AND FLUCTUATIONS IN FOREIGN RESERVES Equation Dependent V ariable Constant Coefficients of the Explanatory V ariables X, R2 F -T e st D. F. 1. All C ountries X7 7.543 (1. 727) 1.494 (2.850) .123 8.12 (1,50) 2. All C ountries X7 15.421 (3.378) 1.449 (3.050) -.1 7 1 (3.456) .280 10.92 (2,49) 3. All C ountries X1 4.755 (8.512) -. 247 (3.519) -. 007 (. 395) .213 7.89 (2,49) 4. All C ountries X1 1.716 (2. 800) -. 086 (1.509) .325 (6.641) -.0 0 9 (. 692) .581 24.60 (3,48) 5. All Countries X1 1.717 (2.395) -.0 8 6 (1.487) .325 (6. 528) -.00002 (. 003) -.0 0 9 (.615) .572 18.06 (4. 47) 6. Less Developed Countries *7 7.362 (. 964) 1.542 (1.911) .074 3.65 (1,32) 7. L ess Developed Countries X7 17.564 (2. 845) 1.437 (1. 979) -.2 1 4 (2.945) .253 6.60 (2 ,3 1 ) 8. L ess Developed Countries X1 4. 813 (6. 586) -.2 5 5 (3.177) -.0 1 2 (. 718) .258 6.72 (2,31) TABLE 9—Continued Equation Dependent V ariable Constant Coefficients of the Explanatory V ariables Xr t xo X 2 4 6 7 R2 F -T e st D. F. 9. Less Developed Countries X1 2.752 (3. 793) - . 150 (2.245) .258 (4. 586) m -.0 1 4 (1.096) .549 14.39 (3,30) 10. Less Developed Countries X1 2.453 (2.806) - . 151 (2. 245) .266 (4. 572) .004 ( .626) -.0 1 0 (.664) .540 10.67 (4,29) 11. Developed Countries X7 11.316 (2.188) .691 (. 747) -.0 2 7 .56 (1,16) 12. Developed Countries X7 14.602 (2. 860) .871 (1.008) -.0 9 3 (1.890) .115 2.11 (2,15) 13. Developed Countries X1 2.413 (1. 718) .055 (. 247) . .058 (.971) -. 054 .56 (2,15) 14. Developed Countries X1 -1. 057 (.976) -.00006 (. 0004) .518 (5.178) .040 (1.116) .613 9.96 (3,14) 15. Developed Countries X1 -.6 1 3 (.514) .026 (. 187) .521 (5.175) - . 008 (. 920) .024 (. 593) .618 7.60 (4,13) Note: W here: X^, X2, X^ and Xg denote the sam e v ariables as before. X7 = average y e a r to y e a r percentage change in foreign re se rv e s. Source: Tables 13 and 14 of Appendix C. 93 m easu res without having to re s o rt to the fluctuation of th e ir foreign reserv es. F o r the le ss developed countries it seem s that th e re is a som ewhat passive o r m echanical response of the fluctuation of foreign re s e rv e s to export instability which is indicative of th e ir inability to neutralize the effects of the la tte r for th e ir dom estic econom ies. Concluding our sta tistic al testing of the effects of export instability on econom ic grow th, we may say that our findings, despite the lim ited num ber of v ariab les investigated, clearly show that th e re is a significant negative effect of export instability on econom ic grow th fo r the le ss developed countries only. Final judgm ent and policy recom m endations derived from the findings of our testin g have been postponed for the concluding ch ap ter of the study, a fte r the exam ination of the im portance of p ric e and quantity fluctuations in the variation of export proceeds. CHAPTER V PRICE STABILIZATION AND ECONOMIC DEVELOPMENT To conclude our em pirical study, we attem pt to determ ine the relative im portance of export p ric e s and quantities in the observed instability of export proceeds. If p o ssible, we will also tra c e any distinctive differences between developed and le ss developed countries on this account. We consider such in form ation to be relevant to the discussions of the s till unsettled issue as to w hether stabilization policy should be adopted and, if so, what p articu la r m easu re would be m o st advantageous. It m ust be said that in what follows we w ill not attem pt to review the p ro and con argum ents fo r the different proposed stabilization schem es, since se v e ra l authors have already done th is. 1 We w ill m erely show som e em pirical evidence which may be easily derived from data employed e a rlie r in this study relating to a p ro d u c e r's p ric e stabilization schem e. The R elative Im portance of the P ric e and Quantity Instability We f ir s t estim ated the instability indexes fo r the export unit p rice and quantity indexes fo r all the countries fo r which such indexes w ere available. *See fo r instance, A. M acBean, Export Instability and Economic D e- velopm ent (Cam bridge: H arvard U niversity P re s s , 1966), P a rts III and IV. Also S ir Sidney Caine, P ric e s fo r P rim ary P ro d u cers, H obart P ap er 24, 2d ed. (London: Institute of Economic A ffairs L td ., 1966). 94 Since, as shown in C hapter n ,1 one of the p ro p erties of our "m echanical" instability index is to rem ain unchanged when each value of the tim e series fo r which the index is estim ated is m ultiplied by a constant, this index is particu larly appropriate for the estim ation of the instability in a tim e s e rie s of index num bers. F o r, in such a case, the value of the instability index is independent of the selected referen ce b ase y ear. In other w ords, the value of the insta bility index estim ated would not be affected by shifting th e referen c e base of th e tim e s e rie s and, consequently, it would not m ake any difference if the instability indexes w ere estim ated from tim e s e rie s with different reference bases. The instability indexes fo r both quantity and unit value of exports are p resented in Table 10. F rom the re su lts p resented in this table, it is esti m ated th at the instability indexes fo r both the quantity and p ric e of exports for the average le ss developed country a re m ore than tw ice the size of the c o rre - 2 sponding indexes of the developed countries. A lso, on the average, export volum e instability is higher than p ric e instability fo r both developed and less 3 developed countries. 1See C hapter n, pp. 31-32. | 2 The average p ric e and quantity instability indexes of the le ss de veloped countries are significantly different from th e corresponding indexes fo r the developed countries a t a level of significance le ss than 1 percent. 3 Less developed countries: A verage Quantity Instability Index,8.67; A verage P rice Instability Index, 7.09; Developed co u n tries: A verage Quantity Instability Index, 4.25; A verage P ric e Instability Index, 3.28. Cf. Mac Bean, "In the case of world exports of individual com m odities, the d egree of in sta - TABLE 10 VALUE, QUANTITY AND PRICE INSTABILITY INDEXES LESS DEVELOPED COUNTRIES DEVELOPED COUNTRIES Instability Indexes Instability Indexes Country Value Quantity P ric e Country Value Quantity P ric e I I I I I I M Q P M Q P A rgentina 6.50 8.00 6.13 A ustralia 9.38 3.78 7.86 Bolivia 14.17 10.14 6.79 A ustria 3.73 3.33 2.17 B razil 7.36 9.16 9.14 Belgium -Lux. 5.70 4.60 3.00 Ceylon 6.43 4. 48 4. 87 Canada 5.20 4.69 1.19 Chile 10.75 9.24 9.87 D enm ark 3.31 1.99 4.14 China (Taiwan) 13.64 12.82 8.78 Finland 4. 58 3.86 4. 01 Colombia 9.32 6.18 8.57 F ran ce 5.26 4.42 2.79 C osta Rica 10. 09 10.66 8.36 Germ any 3.58 2.81 1.57 Cyprus 9.96 5.10 7.68 Iceland 7.19 6.05 9.82 Dominican Republic 14. 27 11. 43 12.96 Italy 6.85 7.16 3.11 Ethiopia 8.63 9.35 6.11 Japan 7.95 8.85 2.42 Ghana 8.65 8.89 15.52 N etherlands 4. 03 3.29 1.78 G reece 6.87 5.67 5.47 New Zealand 6.76 2.80 5.24 G uatem ala 9.12 9.37 11.09 Norway 5.29 4.39 3.34 Honduras 12.57 10.03 5.64 Sweden 3.71 2.71 1.86 India 5.03 5.44 5.60 Sw itzerland 3.76 2.97 1.72 Ireland 8.22 7.54 2.09 United Kingdom 2.51 1.85 1.65 Isra e l 7.40 6.13 7.03 United States 6.49 6.55 1.20 Jam aica 6.09 4.95 5.72 Kenya 6.40 7.65 3.55 C O M alaysia 12.60 3.77 9.39 O i TABLE 10—Continued LESS DEVELOPED COUNTRIES DEVELOPED COUNTRIES Instability Indexes Instability Indexes Country Value Quantity P ric e Country Value Quantity P ric e I , . I I I I I M Q P I £ < © I M auritius 16.02 10.62 7.36 M orocco 4.84 6.28 4.86 N icaragua 12.81 12.80 9.29 N igeria 7.81 6.50 5.12 Pakistan 8.74 15.29 11.82 Panam a 12.13 11.57 5.88 P eru 8.50 6.11 4.94 Philippines 6.74 5.27 4.11 Portugal 6.67 5.12 3.50 Spain 12.23 13.30 4.43 Sudan 15.15 19.10 9.49 Tanganyika 11.29 9.28 5.08 Thailand 9.01 7.45 3.19 Trinidad & Tobago 4.36 5.59 4.28 Tunisia 10.51 11.96 7.37 Source: I. M. F . , International Financial S tatistics (various issues). United Nations, International T rade S tatistics (various issues). CO A c lo se r exam ination of Table 10 data shows that fo r 20 out of the 36 le ss developed countries included in our sam ple—that is fo r about 55 percent of them —the instability index of export proceeds is g re a te r than both p ric e and quantity instability indexes. Thus, fo r m ore than one-half of the le ss developed countries in our sam ple the fluctuations of the p ric e s and the quan titie s exported during the period under consideration did not tend to offset each oth er but ra th e r tended to be reinforcing. N evertheless, the sam e phenomenon is tru e to a g re a te r extent for the developed countries, since fo r 13 out of 18 developed countries the fluctuations in export proceeds w ere la rg e r than both th e ir p rice and volum e of exports fluctuations. This im plies th at fluctuations of export p rice s and quantities tend to reinforce each oth er ra th e r than to off s e t each other in both less developed and m ore developed countries. Next, we attem pted to estim ate the elasticity of export proceeds with reg ard to both export prices and quantities for each less developed country. D espite the highly aggregate c h a ra c te r of the export proceeds (due m ainly to the fact that many le ss developed countries export a very lim ited num ber of p rim a ry product^, we feel that the estim ated e lasticitie s a re not com pletely m eaningless if they a re in terp reted only as tentative indicators. The re su lts presented in Table 11 (columns 1 and 2) have been obtained a y by fitting the exponential function V = A Q P by the OLS method to data from bility of export p ric e s seem s slightly higher on the average than fo r quantities. " Mac Bean, p. 46. 99 TABLE 11 SUMMARY RESULTS OF THE EXPORT PRICE-QUANTITY ELASTICITIES AND THEIR CORRELATIONS (L ess Developed Countries) (1) (2) (3) (4) (5) (6) E xport- E xport P a rtia l C orrelation Z ero -ran k Corr. Country Quantity P ric e Coefficients Coefficients E lasticity E lasticity 2 2 0 Y rV Q .P rV P .A r PQ rVP A rgentina 1.00 1.01 .988 .942 -. 558* -. 134 Bolivia .85 1.11 .883 .949 .030 .833 B razil 1.01 1.00 .961 .973 -.760* .571* Ceylon .83 .8 4 .946 .927 -.627* .269 Chile .79 1.03 .839 .744 .159 .644* China (Taiwan) 1.40 .21 .987 .162 -.407 -. 346 Colom bia .85 .76 .557 .848 -.537* .811* C osta R ica .73 .3 4 .305 .077 -. 845* -. 389 Cyprus 1.12 .79 .962 .884 .234 .619* Dom inican Republic 1.19 .57 . 832 .501 -.2 0 0 .227 Ethiopia .94 .72 .995 .927 -. 495 -. 237 Ghana .88 .88 .922 .882 -. 842* -.083 G reece 1.00 1.01 .693 .897 .539* . 668* G uatem ala .94 .79 .958 .868 -.783* -.362 Honduras .92 .53 .970 .621 . 169 .362 India 1.02 1.09 .928 .961 .551* .912* Ireland 1.00 1.05 .991 .879 .835* .891* Is ra e l .94 .32 .991 .264 .365 .411 Jam aica 1.20 .5 0 . .990 .477 .615* . 668* Kenya 1.00 .62 .980 .358 -. 832* -.791* M alaysia .95 .92 .995 .979 -.102 .365 M auritius .89 1.01 .996 .988 .074 .555* M orocco 1.19 1.22 .889 .952 .787* .964* N icaragua 1.14 .8 4 .989 .821 -.605* -.432 N igeria 1.03 1.45 .874 .397 -. 426 -.122 P akistan .72 .50 .773 .457 .268 .573* Panam a .88 .93 .992 .979 -.239 -.068 P eru 1.53 .58 .966 .679 .425 .603* Philippines .83 .92 .971 .587 .166 .349 Portugal 1.05 . 14 .945 .002 .524 .505 Spain 1.02 .95 .962 .979 .067 -.265 Sudan 1.06 1.08 .982 .865 -. 571* -.263 100 TABLE 11-Continued (1) (2) (3) (4) (5) (6) E xport- Export P a rtia l C orrelation Z ero-rank C o rr. Country Quantity P ric e Coefficients Coefficients E lasticity E lasticity 2 2 3 Y r VQ.P r VP.A r PQ rvp Tanganyika 1.02 .91 .996 .926 -. 233 -.008 Thailand 1.01 .97 .997 .946 .620* .728* Trinidad & Tobago .96 .32 .967 .205 .194 .277 Tunisia .99 .99 .989 .957 .226 .590* Note: *Significant at 5 percent. Source: I. M. F . , International Financial S tatistics (various issues). United Nations, International T rade S tatistics (various issues). 101 36 le ss developed countries fo r the period 1953 to 1966.1 From Table 11 one can see that the proportion of the le ss developed countries fo r which the "aggregate" elasticity of export proceeds with regard to quantity is g re a te r than the "aggregate" elasticity of export proceeds with reg ard to p ric e is m ore than tw ice the proportion of countries fo r which the • * 2 re v e rs e is true. A lso, from the p artial determ ination coefficients shown in columns (3) and (4) of Table 11, it is c le a r that in about every four out of five less de veloped countries the variation in export proceeds is b etter explained by the v ariatio n in the exported quantities ra th e r than in export p ric e s. This, in o th er w ords, m ight im ply that fo r m ost of the le ss developed countries export proceeds would vary m ore if p ric e s w ere held constant and only quantities w ere allowed to vary than if p ric e s w ere allowed to vary by the sam e p ropor tion while quantities w ere held constant. Finally, from columns (5) and (6) of Table 11 which show the zero rank co rrelatio n coefficients between p ric e and quantity and p rice and export p ro ceeds indexes respectively, one can obtain som e tentative ideas about the * W here V is the index fo r export proceeds, Q the quantity of exports and P the export unit p ric e index. 2 3 < Y 20 p ercen t of the countries. 3 > Y 55 p ercen t of the countries 3 approxim ately equal with Y 25 p ercen t of the countries 102 p ric e elasticity of the demand for exports. r r PQ PV P ric e E lasticity of Demand No. of No. of Countries __________________________ C ountries with Sig. Coef. E lastic 12 1 + Inelastic 5 2 + + Indeterm inant. Shifts in d e- 19 m and g re a te r than shifts in supply. ____ 6 Total 36 9 F irs t, for the m ajority of the less developed countries th ere is no significant co rrelatio n eith er between p ric e and quantity of exports o r between export p ric e s and export proceeds. This indicates that shifts in dem and and supply do not rev eal any system atic p attern of change. Second, tw o-thirds of the countries fo r which significant co rrelatio n coefficients w ere found fo r both the p rice-quantity and the p rice -ex p o rt proceeds relationships do not indicate anything about the elasticity of the dem and fo r exports. It is interesting to notice h ere that both countries fo r which the significant co rrelatio n coefficients indicate an inelastic demand fo r exports with reg ard to p ric e a re the two m ain coffee ex p o rters, B razil and Colombia. It should be stre ss e d that a positive co rrelatio n coefficient between p rice and quantity of exports fo r a certain country does not necessarily indicate th a t the country is faced with an inelastic demand for its exports. It m ay very w ell indicate the significance and the relativ e stren g th of the factors lying behind the demand for the p rim ary prod ucts the country is exporting. This can be explained as im plying th at the 103 significant changes in the demand of the p rim ary products a re factors such as w ars, changes in production m ethods, and sudden changes in indu strial output, which a re not usually offset by opposite changes in the output o r supply of the p rim ary producing countries. Consequently, the volume of exports as well as the value of export proceeds move in the sam e d irection as the level of de mand and export p ric e s. Now, in view of th e re su lts in C hapter IV on the effects of export insta bility on econom ic grow th, th e re is little doubt that som e kind of stabilization schem e would be beneficial fo r the less developed countries. B efore we under take th is subject, how ever, th e re a re certain points which should be clarified: F irs t, it should be recalled that our em pirical study has been c a rrie d out in te rm s of instability in the export proceeds. Second, it should be recognized that "stab ilizatio n " can m ean different things. That is, it can m ean stab iliza tion of export pro ceed s, of export p rice s, of p ro d u c e rs' money incom e, o r p ro d u cers' re a l incom e, o r of p ro d u cers' p ric e s. Finally, due to the wide v ariation of production conditions, elasticities of demand and supply, degree of dom estic consum ption of exportables, e t c ., it is im possible to conceive a stabilization schem e which would be appropriate and su ccessfu l for all o r even many of the p rim a ry products exporting countries. To give an idea of how com plicated the problem of stabilization is as well as of the v ariety of the factors which m u st be controlled o r taken into consideration when a stabilization schem e is undertaken to stab ilize eith er export p ric e s o r export p ro ceed s, we w ill u se a sim ple illu stra tiv e exam ple. 104 F irs t, we assum e th at the world demand for a certain p rim ary product exports is of the constant elasticity type, i. e. : - e v A P e (1) v . w here e is a lognorm ally distributed random e r r o r te rm , th at is 2 v e ~ A (0, ). The world supply of the product (Xw ) is equal to the country’ s exports (XD) plus th e exports of all other countries (Xq). So, W XD + X0 T h erefore, in equilibrium it would be: (2) D a P- e eV = XD + Xf (3) from which: A1/e n r v rl/e v/e A X0> e From (4) it can be seen that: 9P 3X D £ (XD+X0 > 1 + 3 X D (4) (5) T herefore, the p ric e elasticity of the demand fo r exports of the country under consideration is: A ssum ing th at ( 3 v / 3 XD) = 0. In other w ords, that the random fac to rs influencing the world p ric e s a re independent of the quantity exported by the country under consideration. From (6) it is c le a r that the p rice elasticity of demand of a country's exports depends: (a) directly on the w orld's p ric e elasticity of demand for the product it exports, (b) inversely on the country's proportion of the world exports and (c) inversely on th e response of the other countries to a change in exports of the country under consideration. 1 If we assum e fu rth e r that the country is faced with a constant p rice elasticity of demand for its exports we can w rite: DD = P ' n eU (7) w here eU is a lognorm ally distributed random e r r o r te rm , i. e . , u 2 e ~ A (0, ou> . Then, assum ing th a t at equilibrium = X ^, we have from Equation (7): - l / n u / X p • e (o) It is also possible to assum e that at le a st ex-post the dom estic output and supply of exportable prim ary products (especially of cash crops which a re This facto r is im portant in the case of an international buffer stock aim ing at the stabilization of world p rice s fo r a c e rta in commodity. U nless the countries involved are secured ag ain st in creases in supply from outside coun tr ie s , the schem e does not have m any hopes fo r success. 106 not consumed dom estically) is to a g re a t extent random. This m eans that in dependently of the factors taken into consideration when planting sta rte d —that is , the factors which determ ine the a re a cultivated, the type of seed, the am ount of fe rtiliz e r, e t c ., due to w eather conditions, p e sts, and other un controllable facto rs—the realized output m ight be com pletely random m ost of the tim e. So, without loss of reality we may define: w D 0 ) w w here e is also a lognorm ally distributed random variable, i. e . , 2 eW ~ A(0, o ^ ) . Combining (8) and (9) we can see that the earnings from exports fo r the country under consideration a re : R u /n ( l- i) W P • XD = e • e i (10) Now from (10) we can see that: V ar (R) VAR [eu / n ] + V a r [ e (1" ~ ) w ] Exp ( a 2/ n 2) Exp ( a 2/ n 2) - 1 u + Explo/<l -M • JexP [a ^ / ( l - )2] ” 1 } • Now from (11) we can se e that fluctuations of a country's export (11) proceeds, that is the V ar(R), in c re ase with the magnitude of the variances 2 2 and a w , which a re the facto rs causing the changes in the world de mand and the country's supply of the product. With regard to the p ric e elasticity 107 of the country's exports one can see that: 2 , (a) as n in c re ases Exp ( ° / i ) d ecreases and fo r n - ► oo, 2 2 Var(R) - ► [Exp (aw)][Exp (aw)-l ] . This m eans that the g re a te r the p rice elasticity for the demand fo r the c a in try 's exports, the m o re the fluctuations of the country's export proceeds depend on the fluctuations of the dom estic production; (b) on the other extrem e, when n - * • 0, Var(R) -* {[Exp ( a ^ / 0)] [ Exp ( ° u/0) “ 1 If - * ■ 00• This im plies that the sm a lle r the p rice elasticity of demand for the country's exports, the la rg e r a re the fluctuations of its ex p o rt proceeds and the m ore dependent they a re upon the random fluctuations of 2 the w orld's demand (a ); u 2 (c) the c lo se r n is to one, the m ore the Var(R) depends on 2 rath e r than on a u The above considerations should indicate that, despite the extrem ely sim ple assum ptions used in our attem pt to determ ine facto rs which affect a country's export p roceeds, th e re a re s till too many such facto rs and that they a re generally out of the individual country's control. The Argum ents About P ric e Stabilization Given that a ce rta in country can at b est control the fluctuations in the output of its exportables by, say, p e st controls, b etter cultivation m ethods, selection of b etter seed s, e t c . , and also that only in ra re ca se s is it in a p osi tion to significantly affect the p rice of its exports by controlling th e ir volum e, 108 it seem s difficult to postulate that a country can take effective u nilateral actions ito reduce fluctuations in its export earnings. On the other hand, as long as ! international m easures to prevent o r dim inish such fluctuations do not seem prom ising, at le a st for the p resen t tim e, it appears that it falls on the less developed countries them selves to take p ro p e r palliative m easu res to m itigate l the dam aging effects of export fluctuations on th e ir econom ies. We, therefo re, think that it is worthwhile to investigate the effects of a u n ilateral action of stabilization in o rd er to see what argum ents may be sup ported by em pirical evidence. Specifically, we will exam ine the effects of a p ro d u cers’ p rice stabilization schem e with reg ard to the country’s economic growth. F irs t, it may appear as m ore reasonable for a governm ent which wants to prom ote dom estic stabilization to concentrate on a p ro d u cers' p rice ra th e r than on a p ro d u cers' incom e stabilization schem e. This seem s m ore reasonable since, on the one hand, there a re m o re facto rs affecting quantities produced and demanded as well as p rice s. T h erefo re, it is m o re difficult, |if not im possible, fo r the governm ent (or its authorized agency) to succeed in stabilizing eith er the individual p ro d u cers' incom e o r the incom e of the p rim ary se c to r in general. On the other hand, it m ight be argued that unless the pro duction is significantly disrupted by m ajor natural d is a s te rs , fluctuations in p ro d u cers' incom es which a re due to crop fluctuations a re not generally con sidered by pro d u cers as discouraging to th e ir efforts for m ore production. In support of this statem ent one m ight claim th at in such cases producers generally I 109 do not ra ise demands for governm ent's in terferen ce or support fo r the extenua tion of the hardship from crop fluctuations. Consequently, it m ight be claim ed J that on the grounds of feasibility of im plem entation as well as on individual fa rm e rs' p reference, a p rice stabilization schem e could be considered as hav ing an edge over an income stabilization schem e. The m ain objection that is raised frequently against a p rice stab iliz a tion schem e of p rim ary exportables is the fe a r that constant p rices received by producers w ill neutralize them from the m ovem ents of the world m a rk e t prices. As a re su lt of such isolation, the p rim ary exporting country will be deprived from the increased foreign exchange earnings which would have been accruing during the boom period because of in creases in output in response to increased p rice s. On the contrary, the dom estic p ric e s would be held higher than o th er w ise during a decline in international p ric e s, encouraging a higher output which will push p rice s fu rth er down. So, the overall re su lt of a stabilization schem e for the p ro d u ers' p rice s would be sm a lle r earnings of foreign exchange. Given the im portance of foreign exchange in the economic growth of the le ss developed I jcountries, as has been shown before, such a schem e would be detrim en tal ra th e r than advantageous fo r the country which w ill im plem ent it. D espite the plausibility of this argum ent, careful exam ination may show that it does not hold generally. At firs t, it should be s tre sse d th at the argum ent, as is usually presented, is valid only in the case of elastic export supply with resp ect to p rice. However, one m ight have som e reserv atio n s with reg ard to the responsiveness of supply of p rim ary products to p rice changes, 110 especially in the sh o rt run. It may also be said that responsiveness would vary jgreatly with the type of farm ing, the prevailing tenure system , and especially with the kind of product. Unfortunately, there is little em pirical evidence on this point and as a re su lt m ost of the argum ents around this point have been based mainly on a p rio ri assum ptions o r casual observations. A lso, in the case of inelastic demand fo r the exports of a country, it seem s doubtful th at the country can earn m ore foreign exchange by allowing its pro d u cers to respond to fluctuating export p ric e s than by arranging fo r relatively m ore stable dom estic p rice s. F u rth er, the argum ent against p rice stabilization loses a g reat deal of its force if one thinks about a p ric e stabilization schem e that does not call for fully pegged p ro d u cers' p ric e s, but which sim ply intends to sm ooth out excessive p ric e fluctuations without losing contact with the trend in the in te r national p rices. If we co nsider that excessive p rice fluctuations in the sh o rt-ru n a re to a g re a t extent the re su lt of significant differences between the long and the ! jshort-run equilibrium p ric e s in cases where supply and demand conditions change d rastically , then a p rice stabilization schem e m ay help the producers to avoid undue hardship during a transition period by allowing fo r the adjust m ents to take place at a slow er pace. A flexible or adjustable p ro d u cers' p rice stabilization schem e m ay also help the country to avoid a cobweb type situation by discouraging over-investm ent during the boom period. Because, as is frequently the case with farm production—especially that involving tre e Ill c ro p s—the increased output m ight come at som e tim e long a fte r the investm ent jtook place, which may very well be a period of declining p rice s. By flexible o r adjustable stabilization schem e, we m ean one which has been designed to operate in such a way as to pay the producers p rice s which i a re kept in line with the trend of world p ric e s , without, on the other hand, I allowing the im m ediate transm ission of the d ra stic and frequent changes in w orld p rices to the dom estic producers. Such flexibility may be achieved if the p rices paid to the producers have been estim ated with a form ula along the lines with the one proposed by B auer and P a ish . 1 This im plies a weighted averaga of p a st p ric e s with the weights as well as the num ber of y ears included in the moving average depending upon the d esirab le flexibility of the schem e and also upon the p a rtic u la r c h a ra c te ristic s of the supply and demand conditions of the product for the p rice of which the stabilization schem e is designed. Another type of a flexible p ric e stabilization schem e is to change the p rice s paid to the producers according to a rate of accum ulation o r reduction of the funds collec ted by the governm ent from the difference between received p ric e s and p rice s paid to the prod u cers. M ore im portant, however, is the fact that a p ric e stabilization schem e m ight help o v erall in the economic growth of a country. F irs t, th e re m ight be som e gains fo r the individual producers in the form of higher incom es due to *P. T. B auer and F. W. P aish, "The Reduction of Fluctuations in the Incomes of P rim a ry P ro d u c e rs," Economic Jo u rn al, LXII, No. 248 (Decem b e r, 1952), 750-780. 112 g re a te r certainty resulting from m ore stable p rice s. | It is not unreasonable to believe that people in general and especially p roducers in the less developed cou n tries, due to poor financial and banking conditions that make borrow ing e ith e r difficult or extrem ely costly, a re very reluctant tow ard risky activities. Consequently, a reduction in the fluctuations of the p rices received by the producers will re su lt in reduction of the risk involved in farm ing, at le a st on the demand side. By enabling producers to get a m ore stab le p rice and to be confident that this p rice will not change d ra s tic a lly at least from y e a r to y e a r, p ric e stabilization would induce producers jto im prove th e ir farm ing and possibly to draw down th e ir unit costs by im prov- i |ing th e ir yields. It is conceivable, th erefo re, that producers induced by g re a te r p rice stability and reduced ris k would end up with low er costs and hence la rg e r net incomes even if the p ric e s received a re on the average lower than otherw ise, depending of course on the size of cost reduction. Second, a gain from the p ro d u cers' p rice stabilization is that the p ro duction plans of fa rm e rs would not be interrupted o r disturbed because of the frequent changes in world p rice s. Consequently, som e rate of capital form a tion would be m aintained even during periods in which the w orld p rice s for p rim ary products are extrem ely low. In the w orst case, even if they do not in c re ase th eir capital form ation, during a slum p in world p ric e s, the fa rm e rs would not be so discouraged as to d isin v e si by leaving tre e s to die without re planting, o r to allow th e ir farm s to d eterio rate o r to shift to other crops m ainly of the subsistence type, as would have been the case in the absence of p rice 113 stability. This suggests that when, subsequently, the world p ric e begins in creasing again, the country would be b e tte r off since it would have a la rg e r i 1 productive capacity than it would otherw ise, and consequently, that it would be able to earn la rg e r am ounts of foreign exchange from its la rg e r exports. I Third, with m o re stable p rice s received by producers of cash crops and, th erefo re, with the export se c to r potentially having a m ore stable o r possibly a la rg e r on the average incom e than otherw ise, it is conceivable that the economy as a whole w ill benefit, if not from anything else, at le a st from I i the m ore stable spending of those in the export se c to r. It does not take too much thinking to understand the im plications of the above statem ent. It we consider a period of booming world p rice s fo r p rim ary com m odities and a country in which th e re is no stabilization of the p rice s received by the export e rs , it is reasonable to expect higher spending on th e ir p a rt which m ight spread inflationary p ric e p re s s u re s throughout the economy unless the govern m ent takes the p ro p er counterbalancing m onetary and fiscal m e asu re s. On the con trary , when the world p ric e boom is over and low export p rice s prevail, ! i ;the e x p o rte rs' incom es and spending will be significantly reduced and it m ight cause sp iral deflationary waves in the economy. Since the governm ent o r the p rice stabilization agency is in a position to intervene and m oderate the excessive p ric e changes before they reac h the individual p roducers, it would be able to act in the sam e fashion and bring the sam e resu lts that a countercyclical fiscal policy ex erc ises on the economy. 114 Given the w eaknesses of fiscal and m onetary policies of the less developed !countries, such built-in stabilization schem es would seem to be highly d e s ira - i jble. i I An external benefit to a p rice stabilization schem e, which m ight not be !altogether negligible, a t le ast as far as the real income of the country is con- I cerned, re su lts from the m ore even disposal over tim e of the accum ulated foreign exchange earnings from exports. If, fo r instance, the in crease in ex p o rt p ric e s is in line with an overall w orld p rice in c re ase, and all foreign exchange earnings received from exports w ere m ade available fo r expenditures on im ports, then these im p o rts, especially of capital goods, would be p u r chased a t the prevailing high p ric e s. If, how ever, the stabilization schem e operates in such fashion as to release foreign exchange in the sam e ra te it com pensates the producers during the period of low w orld p ric e s, then the im p o rts purchased with the foreign exchange m ade available to the economy in excess of the currently earned amount from exports would probably be obtained at low er p rice s. This m eans that the economy may be gaining overall in real ^purchasing pow er fo r its foreign exchange receip ts from exports. The E m pirical Evidence E m pirical support for the above argum ents is provided in Table 12. Although, as is shown in Table 10, fluctuations in the quantity of exports a re found to be la rg e r than those of export p ric e s, it seem s that the la tte r a re m ore detrim ental than the fo rm er, especially fo r the le ss developed countries. 115 TABLE 12 SUMMARY RESULTS OF THE EFFECTS OF EXPORT QUANTITY AND PRICE INSTABILITY ON THE INCOME AND EXPORT GROWTH RATES Equation Dep. V ari able Constant Coef. of Expla natory V ariables ‘Q V * 2 F -T e st D. F. 1 . Less Developed X1 3.118 -.0 8 1 -.0 1 0 .66 Countries i (3. 428) (.811) (1,32) 2 . Less Developed X 1 3.657 -. 177 .047 1.88 C ountries 1 (4. 451) (1.622) (1,32) 3. L ess Developed X4 6.174 -. 122 -.0 1 5 .51 C ountries *± (3.941) (.711) (1,32) 4. L ess Developed X4 8.264 -.4 5 2 . 142 6.44 C ountries (6.167) (2. 538) (1,32) 5. Developed X 1 1.847 .408 ..1 7 1 4. 52 C ountries X (2.091) (2.125) (1,16) 6 . Developed X 1 3.960 -. 118 -.0 3 2 .47 C ountries X (5.748) (.684) (1,16) 7. Developed X4 4.454 .784 .310 8.63 Countries (3.628) (2.938) (1,16) 8 . Developed X4 8.368 -. 183 -.0 3 1 .48 C ountries (7.981) (.696) (1,16) Note: W here: X, and X . denote the sam e v ariables as before. 1 4 Iq = export quantity instability index (m echanical). Ip - export p ric e instability index (m echanical). Source: Table 10 and Table 14 of Appendix C. 116 As it can be seen from Equations (1) through (4) of Table 12, the m ost significant relationship is that between export p ric e instability and the grow th i rate of exports. D espite the fact that the export quantity instability shows a negative effect on the incom e and the export grow th ra te s , in both cases these i effects do not seem to be significant. These resu lts a re indeed in com pliance i with the argum ent p resented above that it is the p rice ra th e r than the quantity instability that is considered as m o re im portant in the decisions of the p rim ary producers. The re su lts of the sam e reg ressio n s fo r the developed countries, shown in Equations (5) through (8), seem to be exactly the opposite to those fo r the less developed countries. Not only is the instability of export prices le ss significant than the instability export quantities, but in addition, the la tte r a re positively related with the grow th rate s of incom e and of exports. This, of cou rse, m ight not be considered very su rp risin g since, on the one hand, export p rice s of the developed countries do not fluctuate very m uch, so that a good p a rt of the fluctuations in export proceeds m ight be due to fluctuations in the ! exported quantities and because, on the other hand, on the basis of the argu m ent of unbalanced grow th, it m ight be expected that such instability would be stim ulating ra th e r than harm ful to growth, given that p ro d u cers in the devel oped countries can easily adjust th e ir production to p ric e s and demand. One apparent question m ight come to mind a fte r the above exposition. Why a re existing p ric e stabilization schem es not considered successful? Many reasons can be cited in answ ering th is question, the m o st im portant of which 117 m ight be found in adm inistrative problem s. P ric e stabilization schem es, despite the fact that they may be theoretically easy to conceive, a re often very difficult to im plem ent. They involve larg e m anagerial discretio n and, conse quently, th e ir successful operation req u ires som e re stra in t on the p a rt of the governm ent. Specifically, th ere is a need fo r the governm ent to re fra in from using funds obtained from the operation of the schem e for o th er goals. Some of the p rice stabilization schem es, such as the m arketing boards, a re com bi nations of buffer funds and buffer stocks and when difficult tim es a rriv e and I I they a re stacked with la rg e unsold quantities, they tend to re s o rt to production restric tio n s and other m easu res which a re difficult to evaluate. Some other i p ric e stabilization schem es such as variable exchange rate s o r variable export duties, a re sim ply altern ativ e form s of taxation, and m ight not even reach the producers if they succeed only in varying the p ro fit m argins of the m iddlem en and m erchants alone. In g eneral, to quote Sir Sidney Caine, "it m ay be that because th o se in terested in p ric e stabilization have tried to achieve too m uch that they have not achieved as much as m ight in tru th be both feasible and de- ! i ; s ira b le ." Conclusion In conclusion, we feel that the evidence presented h ere supports the a p rio ri argum ents th at export instability is generally la rg e r in the less 1S ir Sidney Caine, "C om m ent," in F irs t Symposium, K yklos, X I, F asc. 2 (1958), 187-193. 118 developed countries than in the m ore developed countries and that this in sta bility is detrim ental to economic growth. More significant, however, is the ! evidence that the harm ful effects of export instability a re inv ersely related to i 1 !the level of economic development. I I The policy im plications which can be derived from the study may be ! I b e tte r understood, if one takes into consideration the distinction m ade p re viously about the different m echanism s through which the effects of export jinstability a re expressed. As was stated before, 1 m ost of the detrim ental ieffects of export instability on the "m icro -lev el, " e. g . , the im position of highei production and consum ption risk s fo r producers of export p rim a ry com m odities and thereby a m isallocation of re so u rc e s away from these m ore profitable export com m odities to less profitable subsistence crops, can be attributed to fluctuations in the pro d u cer p rice s. On the other hand, m o st of the harm ful effects of export instability on the "m acro -lev el, " a re due to the fluctuations in export proceeds which reduce the effectiveness of long-run economic plan ning and in the absence of adequate stabilization policies force the economy to j joperate below its optimum growth rate and lead to chronic inflationary condi tions. As was argued above, a country can conceivably act unilaterally in m itigating the "m icro -ty p e" effects of export instability by im plem enting an 1See C hapter I, pp. 4-5. 119 appropriate p rice stabilization schem e. U nfortunately, the fact that the ipreviously mentioned em pirical studies have attem pted to dem onstrate that i export instability was not a serio u s problem in the le ss developed countries together with the very lim ited su ccess with which existing stabilization schem es I have apparently enjoyed, have lim ited the attention given to dom estic p rice stabilization policies in the rece n t y e a rs. Since the em pirical evidence of our study suggests th at export insta bility is indeed harm ful to the econom ic grow th of the less developed countries, it would seem reasonable and appropriate that the effects of the fluctuations in the p rices of exported com m odities should not be allowed to be superim posed on the "m acro -e ffe cts” stem m ing from the instability which the sam e less developed countries experience in export proceeds. Although it is not possible to define o r apply an optimum p rice stabilization schem e that would be suitable fo r all circu m stan ces, it would seem app ro p riate to recom m end that m uch m o re attention should be paid to the evaluation and the choice of alternative p ric e stabilization schem es th at would fit the conditions of p a rtic u la r le ss i Ideveloped countries. Especially if a country's social costs of export in sta bility a re significantly la rg e r than the p rivate co sts to the exporters affected, then governm ent sponsored stabilizatio n schem es a re indeed justified. i With regard to the problem of instability in export proceeds, it is obvious that effective solutions depend on the cooperation of other countries. T herefore, unless international cooperation is im proved so as to m ake feasible the effective im plem entation of international stabilization schem es, th e re can be little hope that the less developed countries w ill be successful in treatin g the problem of instability in th e ir export proceeds. However, as B rainard and Cooper pointed out, a carefully designed and im plem ented diversification program might be one effective step tow ard the solution o r m itigation of the effects of instability in export p ro cee d s. 1 l W . B rainard and R. Cooper, pp. 269-273. B IB L IO G R A P H Y 121 BIBLIOGRAPHY Books A itchison, J. and Brown, J. A .C. The Lognorm al D istribution. London: Cam bridge U niversity P re s s, 1957. | Blau, G. and M usic, D. A. A gricultural Commodity T rade and Development P ro sp e cts, P roblem s and P o lic ie s. Rome: F. A. O ., 1964. Bowman, M ary Jean, ed. Expectations, U ncertainty, and B usiness Behavior. New York: Social Science R esearch Council, 1958. Caine, S ir Sidney. P ric e s for P rim a ry P ro d u cers. 2nd ed. H obart P aper 24. London: Institute of Economic A ffairs L td ., 1966. C airn cro ss, A. K. F acto rs in Economic D evelopm ent. New York: P raeg er, 1962. Coppock, Joseph D. International Economic Instability. New York: M cGraw- Hill Book Company, In c ., 1962. De V ries, B arend A. The Export Experience of Developing C ountries. B altim ore, Md. : Johns Hopkins P re s s , 1967. E llis, Howard S ., ed. Economic Developm ent for Latin A m erica. London: M acm illan & Co. , 1961. Friedm an, M ilton. A Theory of the Consumption Function. New York: N ational Bureau of Economic R esearch, 1957. H aberler, G ottfried. International T rade and Economic Developm ent. C airo: National Bank of Egypt, 1959. H aberler, G ., and Stern, R .M ., eds. Equilibrium and Growth in the World Econom y. Economic E ssays by R agnar N urkse. C am bridge, M a s s.: H arvard U niversity P re ss, 1961. H arrod, S ir Roy F . , and Hague, Douglas, eds. International T rade Theory in a Developing W orld. New York: St. M artin 's P re s s , 1963. 122 123 H irschm an, A lbert O. The Strategy of Econom ic Development. New Haven: Y ale U niversity P re s s, 1958. International M onetary Fund, International Financial S tatistics. V arious issu es. Johnson, H arry G. International T rade and Economic Growth. Cam bridge: H arvard U niversity P re s s , 1958. _________. Economic Policies Tow ards L ess Developed Countries. W ashing ton, D. C. : The Brookings Institution, 1967. Johnston, J. Econom etric M ethods. New Y ork: M cGraw-Hill Book Company, 1960. K indleberger, C harles P. Foreign T rad e and the National Economy. New Haven: Yale U niversity P re s s , 1962. __________. Economic Development. 2nd ed. New York: M cG raw-Hill Book Company, 1965. Koyck, L. M. D istributed Lags and Investm ent A nalysis. A m sterdam : N orth-H olland Publishing Company, 1954. Levin, Jonathan V. The Export E conom ics. Cam bridge: H arvard U niversity P re s s , 1960. Linder, Staffan B. T rade and T rade Policy fo r Development. New York: F re d e ric A. P raeg er, P u b lish ers, 1967. MacBean, A lasd air I. Export Instability and Economic Development. Cam bridge: H arvard U niversity P re s s , 1966. McKinnon, R. I. "Foreign Exchange C o nstraints in Economic Developm ent and Efficient Aid Allocation. " Econom ic Jo u rn al, LXXIV, No. 294. (June, 1964), 388-409. M aizels, A lfred. Exports and Economic G rowth of Developing C ountries. Cam bridge: Cam bridge U niversity P re s s , 1968. M alinvaud, E. Statistical Methods of E conom etrics. Chicago: Rand McNally & C o ., 1966. M eier, G erald M. International T rade and D evelopm ent. New York: _______ H arper & Row. 1963.__________________________________________________ 124 M eier, G erald M. Leading Issues in Development Econom ics. New York: Oxford U niversity P re s s , 1964. _________. The International Econom ics of Development. New York: H arper & Row, P u b lish ers, 1968. M ichaely, M ichael. C oncentration in International T rad e. A m sterdam : N orth-H olland Publishing C o ., 1962. IMyint, Hla. The Econom ics of the Developing C ountries. London: Hutchison & C o ., 1964. M yrdal, Gunnar. An International Economy. New York: H arper & B rothers, P u b lish ers, 1956. |__________ Economic Theory and Under-Developed R egions. London: G erald Duckworth & C o ., L td ., 1957. N erlove, M. "D istributed Lags and Demand A nalysis fo r A gricultural and O ther Com m odities. " A gricultural Handbook. No. 141. Washington, D. C. : U .S. D epartm ent of A griculture, 1958. O rganization of A m erican S tates, G eneral S ecretariat. Instability in World M arkets fo r P rim a ry P ro d u cts. W ashington, D. C .: Pan A m erican Union, 1958. O rganization fo r Economic Cooperation and Development. N ational Accounts of Less Developed C ountries. 1950-1966. P a ris : O. E. C. D, 1968. j ^ a o , C .R . L inear S tatistical Inference and Its A pplications. New York: John Wiley & Sons, Inc. , 1967. Rowe, J. W. F. P rim a ry Com m odities in International T rade. Cam bridge: Cam bridge U niversity P re s s , 1965. Sachs, Ignacy. Foreign T rade and Economic D evelopm ent of Underdeveloped C ountries. New York: A sia Publishing House, 1965. Sargan, J. D. "Wages and P ric e s in the United Kingdom: A Study in E conom etric Methodology. " Econom etric A nalysis fo r National Economic Planning. Edited by P. E. H art, G. M ills, and J. K. W hitaker. London: B utterw orths, 1964. Pp. 24-54. 125 United N ations, Dept, of Economic A ffairs. R elative P ric e s of Exports and Im ports of Under-Developed C ountries. Lake Success: United N ations, 1949. _________. M easures fo r International Economic Stability. New York: United N ations, 1959. (E/2156, ST/ECA/13) _________. Instability in Export M arkets of Under-Developed C ountries. New York: United N ations, 1952. (1952. II. A. 1) _________. D epartm ent of Economic and Social A ffairs. International Com pensation for Fluctuations in Commodity Trade. New Y ork: United N ations, 1961. (E/CN. 13/40) _________. S tatistical Office. Y earbook of International T rade S tatistic s. (various issues) _________. Yearbook of N ational Accounts S tatistic s, (various issues) W allich, Henry C. "Stabilization of P roceeds from Raw M aterial Exports. " Economic Development fo r Latin A m erica. Edited by Howard S. E llis. London: M acm illan & C o ., 1961. I^>.. 342-361. Wold, H erm an. A Study in the A nalysis of Stationary T im e S e rie s . 2nd ed. Stockholm: A lm qvist and W iksell, 1954. A rticles and P erio d icals A ldons, M. "T rade in A gricultural Com m odities Among Developing C ountries." Q uarterly Review of A gricultural Econom ics, XVHI, No. 3 (July, 1964) 131-139. B auer, P. T. and P aish, F. W. "The Reduction of Fluctuations in the Incom es of P rim ary P roducers. " Economic Jo u rn al. LXII, No. 248 (D ecem ber, 1952), 750-780. _________ "The Reduction of Fluctuation in the Incomes of P rim a ry P roducers: F u rth er Comment. " Economic Jo u rn al. LXIV, No. 256 (D ecem ber, 1954), 704-729. Beckford, G. L. F. "S ecular Fluctuations in the Growth of T ropical A gricul tu ra l Trade. " Economic Developm ent and C ultural Change. XIII, No. 1 (October, 1964), 80-94. 126 B ernstein, E. M. and P atel, I. G. "Inflation in Relation to Economic Develop m ent. " IMF Staff P a p e rs . II, No. 3 (November, 1952), 363-398. B e rrill, K. "International T rad e and the R ate of Economic Growth. " The Economic H istory Review, Second S eries, XII (1960), 351-359. Bhagwati, Jagdish. "Review, " of Coppock's International Economic Instability. Economic Jo u rn al. LXXIII, No. 291 (Septem ber, 1963), 523-525. _________ "The P ure Theory of International Trade: A Survey. " Economic Journal. LXXIV, No. 213 (M arch, 1964), 1-84. B rainard, W illiam C. and Cooper, R ichard N. "U ncertainty and D iversifica tion in International T rade. " Studies in A gricultural Econom ics. T rade, and Developm ent. VIII, No. 3 (1968), 256-285. Cagan, P. "The M onetary Dynamics of Hyperinflation. " in Studies in the Quantity T heory. Edited by Milton Friedm an. Chicago: U niversity of Chicago P re s s , 1956. Caine, S ir Sidney. "Instability of P rim a ry Product P ric e s : A P ro te st and a P roposal. " Economic Jo u rn al. LXTV, No. 255 (Septem ber, 1954), 610-614. ________ . "Comment, " F irs t Symposium. Kvklos. XI, F asc. 2 (1958), 187-193. C airn cro ss, A. K. "International T rade and Economic Developm ent. " Kvklos. Xm , F asc. 4 (1960), 549-558. Cochrane, D. and O rcutt, G. H. "Application of L east Squares R egression to R elationships Containing A utocorrelated E rro r T erm s. " Journal of the A m erican S tatistical A ssociation, XLIV, No. 245 (M arch, 1949) 32-61. Em ery, R. F. "The Relation of Exports and Economic Growth. " K vklos. XX, F asc. 2 (1967), 470-486. Falcon, W alter P. "F a rm e r R esponse to P ric e in a Subsistence Economy: The Case of W est Pakistan. " A m erican Economic Review . P apers and Proceedings, LIV, No. 3 (May, 1964), 580-591. Flem ing, M arcus J. and Lovasy, G ertrude. "Fund P olicies and P ro ced u res in Relation to the C om pensatory Financing of Commodity Fluctuations. " IMF Staff P a p e rs. VIII, No. 1 (November, 1960), 1-76. 127 Friedm an, Milton. "The Reduction of Fluctuations in the Incomes of P rim ary P roducers: A C ritical Comment. " Economic Jo u rn al. LXIV, No. 256 (D ecem ber, 1954), 698-703. Hawkins, R obert G ., Epstein, Jo se, and G onzales, Joaquin. "Stabilization of Export R eceipts and Economic Development. " The B ulletin. Institute of Finance, New Y ork U niversity, No. 40 (November, 1966). Hill, Polly. "Fluctuations in Incomes of P rim a ry P roducers. " Economic Jo u rn al. LXIII, No. 250 (June, 1953), 468-471. Hultman, C harles W. "E xports and Econom ic Growth: A Survey. " Land Econom ics. XLIII, No. 2 (May, 1967), 148-157. K anesathasan, S. "E xport Instability and C ontracyclical F iscal Policy in U nderdeveloped Export Econom ies: A C ase Study of Ceylon Since 1948. " IMF Staff P a p e rs. VII, No. 1 (A pril, 195fy 46-74. K risna, Ray. "F arm Supply R esponse in India-P akistan: A C ase Study of the Runjab Region. " Econom ic Jo u rn al. LXXIII, No. 291 (Septem ber, 1963), 477-487. Lovasy, G. "Inflation and Exports in P rim a ry Producing C ountries. " IMF Staff P a p e rs. IX, No. 1 (M arch, 1962), 37-69. _________. "Survey and A ppraisal of Proposed Schem es of Com pensatory Financing. " IMF Staff P a p e rs. XII, No. 2 (July, 1965), 189-223. M acBean, A lasd air I. "P roblem s of Stabilization Policy in Underdeveloped C ountries. " Oxford Economic P a p e rs . XIV, No. 3 (October, 1962), 251-266. _________. "C auses of Successive Fluctuations in Export P roceeds of Under developed C ountries. " Bulletin of the Oxford U niversity Institute of S ta tistic s. XXVI, No. 4 (November, 1964), 323-340. M assell, Benton F. "Export Concentration and Fluctuations in Export Earnings: A C ross-S ection A nalysis. " A m erican Economic Review. LIV, No. 2 (M arch, 1964), 47-63. _________ "E xports, Capital Im ports and Economic Growth. " Kvklos. XVII, Fasc. 4 (1964), 627-635. 128 M aynard, G. "Inflation and Growth: Some L essons to be Drawn from L atin- A m erican Experience. Oxford Economic P a p e rs. XIII, No. 2 (June, 1961), 184-202. Myint, Hla. "The 'C lassica l Theory' of International T rade and the Under developed C o u n trie s." Economic Jo u rn al. LXVIII, No. 270 (June, 1958) 317-337. N iculescu, B. M. "Fluctuations in Incomes of P rim a ry P ro d u cers: F u rth er Comment. " Economic Jo u rn al, LXIV, No. 256 (D ecem ber, 1954) 730-743. N urkse, R agnar. "T rade Fluctuations and B uffer P o licies of Low-Income C o u n trie s." Kvklos, XI, F asc. 2 (1958), 141-154. P o rte r, R .S. " C om m ent" (F irs t Symposium). Kvklos. XI, F asc. 2 (1958) 231-243. P ra is , S. J. "E conom etric R esearch in International T rade. " K yklos, XV, F asc. 3 (1962), 560-574. Reynolds, C lark W. "D om estic Consequences of Export Instability. " A m eri can Economic Review . P ap ers and P roceedings, L in, No. 3 (May, 1963), 93-102. R ichter, J. H. "Tow ards an International Policy on A g ricultural T rade. " Kvklos. XVI, F asc. 2 (1963), 203-225. S eers, Dudley. "A Theory of Inflation and Growth in U nder-Developed C ountries. " Oxford Economic P a p e rs. XTV, No. 2 (June, 1962) 173-195. Simon, M. "R eview ," of Coppock's International Economic Instability. A m erican Economic Review. Lin, No. 3 (June, 1963), 521-523. Singer, H. W. "Introductory Statem ent, " (Second Symposium). Kvklos. XII, Fasc. 3 (1959), 271-282. Srinivisan, T. N. "F oreign T rad e and Economic Developm ent. " M etroeco- nom ics. XVII, No. 1-2 (January-A ugust, 1965), 83-98. Swerling, B oris C. "P roblem s of International Commodity S tabilization." A m erican Economic Review. P apers and P roceedings, Xffl, No. 3 (May, 1963), 64-74. 129 Yang, Shu-Chin. "F oreign T rade Problem s in Economic Development. " Scottish Journal of Political Economy, XI, No. 2 (June, 1964), 116-135. A P P E N D IX E S 130 APPENDIX A INDEPENDENCE OF THE INSTABILITY INDEX INVOLVING EXPECTATIONS FROM TREND APPENDIX A | It is easily shown that our index, based on expectations, is independent of the type of trend and that it gives zero instability fo r any tim e se rie s having constant trend. I I. Exponential Trend Let Xq be the startin g y e a r's exports and r = (1+i), where i is the 'annual grow th rate of exports; th e re fo re X = X r*. A ccording to our I o definition, (1) T herefore, X r* - |- X r t_3 (r^+r+1) o 3 o \ X r t_3 (r-1 ) (3I-2 + 2 r + 1) o O (2 ) Also, X r o (3) From (2) and (3) we have: (Xt - X*) (Xt_1 - Xt_2) = | X^ r 2t" 5( r - l) 2 (3r2 + 2 r + 1) (4) Summing up (4) we get: 132 133 X > t - Xt><Xt - l - Xt - 2> = I X ^ ( r - l ) V +2 r + l , r * ” , 2 4 2n-6 < 5 > (r + r4 + . . . + r ) (Squaring (3) and summing up we get: (Xt-1 - Xt_ / = X2( r - 1 )2 r 2 (r2 + r 4 + . . . + r2”"6) (6) T herefore, b = E (X, - < ) (Xt-1 - Xt _2) E < Xt - l - X«-2>2 1 J2 .. 2 2 2 4 2n-6 f a (r-1) (3r + 2 r + 1) r ( r + r 4 + . . . + r ) o 0 2 2 2 2 4 2n-6 X (r-1) r (r + r + . . . + r ) o 3 r2 + 2 r + 1 3 r (7) In the case of a perfect exponential trend in exports, th erefo re, by v irtu e of j(l) and (7) X f = X* + b(XA - XA 0) t t ' t-1 t-2 2 |[V-V + r + 1 ) ] + x r t-3 [ i f i l i l + ( 3 A g .H -iH .S ;U r1 , r t-3 3 = r t o L 3 3 r J o 1 o ^ 134 E From which it is obvious that = 0 fo r every t, and consequently the instability index of the se rie s would be zero. II. lin e a r Trend Let X be the startin g y e a r's exports and r the annual in crease in o exports; th erefo re, XA = X + rt. (8) to A ccording to ou r definition: Xt = I <V l + Xt -2 + Xt-3> = V r < ‘- 2> < 9) From (8) and (9) follows that: Also, $ X - X = X + r t - X - r(t-2) = 2 r (10) t t o o X - X = X + r( t- l) - X - r(t-2) = r (11) t-1 t -2 o o From (10) and (11) we have: £ < Xt - x t> < x . - l - Xt - 2» = <"-4)2r2 (12) t=4 Now, squaring (11) and sum m ing up we get: E < x t - l - X. - 2>2 = < - 4> r2 t=4 T herefore, by v irtu e of (12) and (13) we have: (13) E according to (9), (11), (14), and (8). T herefore, Xt - Xt = 0 for all t (t=4,5, ..., n) and consequently the instability index fo r the s e rie s will be APPENDIX B A BRIEF EXPOSITION OF THE RATIONALE BEHIND THE COCHRANE- ORCUTT ITERATIVE METHOD FOR CORRECTING FOR AUTOCORRELATION IN THE ERROR TERM 136 APPENDIX B Here we will attem pt to show the reaso n for the iterativ e solution of Equation (1).* To obtain the le a st squares estim ates of a , 3 and p , and fo r that m a tte r th e ir m axim um likelihood estim ates in the case w here is norm ally distributed, we m ust m inim ize the sum of squared resid u als, i. e . , 2 m inim ize E(e^) . F ro m (8) we have: 2 2 Q = 1 ( et ) = 2 [ Yt - (a ' + PYt _ j + B X j. - 3pXt _1) ] (i) ! D ifferentiating now (i) with resp ect to a ' , 3 and p and setting the firs t deriva- I tives equal to zero we get: = 2 E( Y - pY - 8 X - + 8 P X^ ) = 0 (ii) 3 a t t —1 t t —1 ^ = 2 E ( pX - X )(Y - a' - pY - 8X +8pX ) = 0 (iii) 3 g t-1 t t t -1 t t -1 1 7 = 2 E< B w - Yt-il<Y t - “' - pYt-i - sxt + epXt-1» = 0 <lv> C arrying out the sum m ations in (ii) through (iv) and rearran g in g the term s we have: Z Yt = n a’+ PZY + f3i:Xt " epzXt_i (v) 1 See C hapter III, p. 48. 137 E xpressing the variables in te rm s of deviations from th e ir respective m eans, the system of equations (v) through (vii) becom es: £Vt - p < E Vt-i + £Vt-i,- p 2r x t-iy t-i + + 6( IX j - 2 p l x t x + 0 ; x t-l> iy t yw = e < £ x ty w + £ y ^ ) - 62 £ ^ V i + + p ( ^y2 . ! " 2 eEyt- i x t- i * 62 £ x t . i (tx) w here the lower case le tte rs denote deviations from the m eans of the respective v ariab les. Since the estim ation procedure we are following has the property that the fitted equations a re satisfied by the sam ple m ean s, an estim ate of a can be obtained from (v) by dividing both sides by n, and using the estim ates of 3 and p found from (viii) and (ix). Solving (viii) with reg ard to 3 we get: Now, (x) gives 6 as the ratio of two quadratic equations in P . Consequently, if we substitute the value of 6 from (x) into (ix), a fte r clearing the fractions I we will have a fifth o rd e r polynom ial in p , w ith sum s of squares and c ro s s - products of the v ariables as coefficients. The p ro p e r root of this polynom ial is the acceptable root which yields the sm allest sum of sq uares in (i). N evertheless, since the solution of a fifth degree polynom ial is not an easy operation, we have followed the iterativ e approach which, as has been shown by Sargan, w ill always lead to estim ates of 3 and P which m inim ize <! )■ I "''Sargan, p. 30. APPENDIX C STATISTICAL APPENDIX 140 TABLE 13 INSTABILITY INDEXES M echanical Involving M echanical Involving Country I Expectations I Country T Expectations T M E M E L ess Developed Countries L ess Developed Countries Argentina 6.50 9.70 Indonesia 9.24 8.47 Bolivia 14. 17 16.85 Iraq 7.40 15.94 B razil 7.36 8.38 Ireland 8.22 8.97 Burm a 8.62 8.44 Isra e l 7.40 17.13 Ceylon 6.43 5.75 Jam aica 6.09 1 1 . 0 1 Chile 10.75 12.30 Kenya 6.40 10.41 China (Taiwan) 13.64 21.53 Libya 41.62 47.21 Colombia 9.32 9.85 M alaysia 12.60 12.88 Costa Rica 10.09 12.13 M auritius 16.02 15.85 Cyprus 9.96 9.58 Mexico 5.15 5.45 Dominican Republic 14.27 11.36 M orocco 4.84 6.28 Ecuador 6.88 10.90 N icaragua 12.81 17.64 Egypt 9.63 9.17 N igeria 7.81 10.88 El Salvador 7.28 10.90 8.-74 12.75 Ethiopia 8.63 16.50 Panam a 12.13 18.76 Ghana 8.65 8.67 Paraguay 10.36 10.78 G reece 6. 87 9.06 Peru 8.75 11.80 Guatem ala 9.12 9.68 Philippines 6.74 11.91 Honduras 12.57 15.49 Portugal 6.67 8.33 India 5.03 5.35 Spain 12.23 12.37 TABLE 13—Continued Country M echanical *M Involving Expectations ' e M echanical Country *M Involving Expectations Less Developed Countries Developed Countries Sudan 15.15 14. 48 A ustralia 9.38 12. 82 Syria0 10.92 12.46 A ustria 3.73 6.58 Tanganyika 11.29 12.43 Belgium 5.70 7.49 Thailand 9.01 13.08 Canada 5.20 7.49 Trinidad & Tobago 4.36 8.08 D enm ark 3.31 3.32 Tunisia 10.51 9.70 Finland 4.58 9.34 Turkey 11.77 13.75 F rance 5.26 10.54 Uruguay 17.23 16.51 Germany 3. 58 4 4.83 Venezuela 3.97 5.03 Iceland 7.19 12.91 Y ugoslavia 6.88 20.43 Italy 6.85 10.70 Japan 7.95 10.66 N etherlands 4.03 7.14 New Zealand 6.76 6.92 Norway 5.29 8.84 Sweden 3.71 7.35 Switzerland 3.76 4.48 United Kingdom 2.51 3.49 United States of A m erica 6.94 9.44 & Both indexes estim ated fo r period 1953-1966. ^1953-1963 C1953-1965 M Source: International M onetary Fund. International Financial Statistics (various issues).__________________________ 143 TABLE 14 I VALUES OF VARIABLES USED FOR STATISTICAL ANALYSIS OF EXPORT | | INSTABILITY AND ECONOMIC GROWTH (1953-1966) 1 — i i 1 j Country I i 1 Income Growth R ate v a X1 Export Growth Rate X4 Av. Ratio of Exports to GDP X5 Av. R atio of Foreign R eserv es to Im ports x6 Av. Y r. to Yr. Chng. of Foreign R eserves X7 * Argentina 1.46 3.75 -------- 29.33 48.66 A ustralia 2.36 4.89 13. 49 55.60 21.09 A ustria 4.72 7.75 17.44 58.67 21.26 Belgium 3.15 8.31 33.01 39.62 10.19 Bolivia 1.17 1.57 32.00 14.67 72.32 B razil 2.38 -.2 8 6.47 31.36 24.87 Burma* 2.21 -.4 1 — ----- -------- -------- Canada 1.83 5.69 15. 75 ,'i5. 85 5. 85 Ceylon 1.02 .43 28.21. 36.52 23.83 Chile 1.74 5.45 11. 99^ 18.63 28.80 China (Taiwan) 4.42 13.24 11.06 49.89 23.32 Colombia 1.13 -1 .7 4 10.14 28.91 18.67 Costa R ica .82 3.69 19.65 14.98 28.95 Cyprus 2.23 3.21 20.23 46.01° 27. 09° Denmark 3.89 7.73 24.04 14.64 20.84 Dominican Republic .48 2.34 18.91 28.44 32.67 Ecuador 1.11 3.96 15.87 38.44 11.02 Egypt 3. 65b 2.82 11.94 62.59 12.05 El Salvador 1.90 5.26 21. 76 34.58 12.67 Ethiopia 2. 93c 4.44 7.9 4 1 68.94 12.26 Finland 4.29 6.81 18.23 25.12 17.96 F rance 3.77 7.88 9.90 33.08 9.57 Germany 4.55 10.26 15.40 58.30 22.83 Ghana 1.34 1.18 23. 25m 81.06 13.87 G reece 5.44 6.90 6.43 37.14 13.73 Guatemala* 1.81 5.48 -------- 39.69 15.81 Honduras 1.03 6.55 20.01 22.28 16.68 Iceland 2.57 8.40 24. 33n 25.71 27.25 India 1.42 3.01 4.57 57.27 14.17 Iraq 3. 57d 6.41 38.37 81. 10 21.82 Ireland 2 . 62e 6.66 23.30 52.94 8.14 Israel 5.80 13.85 9.14 17.90 33.55 Italy 4. 80 12.52 10.33 55.60 17.70 Jam aica 4. 28d 10.23 24.12 34.76 9.31 O 144 TABLE 14— Continued 1 — ■ : Income Export Av. Ratio Av. R atio Av. Y r. to , Country Growth Growth of Exports of Foreign Y r. Chng. Rate Rate to GDP R eserves of Foreign Xa x x to Im ports R eserv es 1 4 5 x 6 _ * Z _ Japan 8. 90f 13.46 8.85 33.27 17.02 Kenya .17 7.12 17. 63m -------- M alaysia 2. 00S 4. 72h 42. 08m 7.18 13. 46 Mexico 2.72 4.69 7.51 38.59 18.39 N etherlands 3.32 8.27 34.42 36.97 12.99 New Zealand 2. 36® 3.57 11.45 22.98 21.68 N icaragua 1.95 8.07 9. 90m 21.52 31.03 N igeria* 2.33d 6.09 16.69 47.70° 16.44° Norway 3.17 7.47 10.09 26.99 7.51 {Pakistan 1.69 3.35 4.753 54.74 12.84 Panam a* 3.72 9.40 9.62 31.08 23.03 Paraguay .63 3.80 10.59 16.74 66.95 P eru 2. 80f 9.54 18.01 19.82 24.77 Philippines .91 6.09 10.88 19.77 27.11 Portugal 4.43 7.40 13.91 141.98 6.9 4 Sudan* 2.04 3.89 1 6 .13m 72.98 19.91 Sweden 3.55 8.01 19.80 24.25 9. 59 Sw itzerland 2.81 7.85. 21.40 107.26 5. 16 Syria* 1.76 4.9 1 1 -------- 21.80 29.58 Tanganyika . 54d 6.16 2 4 .24n -------- Thailand 2.94 6.31 15.76 87.95 11.31 Trinidad & Tobago 5.83 8.23 56.30 8.62 7.21 T unisia 2.17 .48 13. 34° 29. 47k 24. 81 Turkey 1.73 2.67 4.5 4 41.02 12.48 United Kingdom 2.41 4.90 14. 96 23.83 16.36 United States 1.84 6 . 04 3.78 141. 63 4.71 Uruguay 0.82 -2 .2 9 14. 07m 112.26 7.51 Venezuela 2.02 3.86 31.54 59.57 20.06 ♦Excluded from testing due to lack of dependence of im ports on exports. Note: aR eal p e r capita GDP a t m ark et p ric e s; ^NDP a t factor cost; °GDP at facto r co st 1957-1966; dGDP a t factor cost; eGDP 1953-57/1958-66; GNP; Sq d P 1955-56; h1953-63; *1953-65; i 1959-66; k1958-66; m 1955-66; n1954-66; °1960-66. 145 TABLE 14—Continued iSources: U. N .. National Income Accounts S tatistical Yearbook (various issu es). O. E. C. D ., N ational Accounts of Less Developed | C ountries 1950-1966. I U. N ., Demographic Y earbook (various issues). | X4 : I. M. F . , International Financial S tatistics (various | issues). I X : U. N ., International T rade S tatistics (various issues). U. N .. National Income Accounts S tatistical Yearbook (various issu es). i X : I. M. F . , International Financial S tatististics (various issues). 1 X7: I. M. F . , International Financial S tatistics (various I issues). O

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The Economics Of Sugar Quotas

PDF

Exchange Rate Policy And Export Performance

PDF

Experimentation with a simple aggregate model for the Greek economy

PDF

On The Dynamics Of Planned Economy: General Theory With Empirical Analysis Of The Hungarian Experience

PDF

The Anomaly Of External Effects: A History Of Concept Development

PDF

Interstate Natural Gas Supply And Intrastate Market Behavior

PDF

The Determinants Of Growth Differentials And Regional Concentration: A Theoretical And Empirical Investigation

PDF

Production Functions In Korean Manufacturing: An Analysis Of Trade Policyand Its Effect On Technological Change

PDF

On The Theory Of Value And Market Syndicalism

PDF

Differential Game Theory Approach To Modeling Dynamic Imperfect Market Processes

PDF

A Theory Of Regional Economic Growth: Growth Poles And Development Axes

PDF

Approaches To Economic Policy For Stabilization And Growth

PDF

Applications Of Stochastic Processes To Economic Development

PDF

Applications Of Stochastic Processes To Indian Agriculture

PDF

Economic Aspects In The Evolution Of The Great Lakes Freighter

PDF

The Economics Of A Non Profit Enterprise In The Dental Health Care Field

PDF

Some Effects Of Monetary And Fiscal Policy On The Distribution Of Wealth

PDF

A Theoretical Analysis Of The Dynamic Aspects Of Economic Integration Of Less Developed Countries

PDF

Valuation Adjustment In Input-Output Tables, Economic Structure, And Economic Systems: A Comparative Empirical Analysis

PDF

European Economic Integration And African States

Asset Metadata

Creator Glezakos, Constantine (author)

Core Title Export Instability And Economic Development: A Statistical Verification

Degree Doctor of Philosophy

Degree Program Economics

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Economics, General,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Tintner, Gerhard (committee chair), Nugent, Jeffrey B. (committee member), Pounders, Cedric J. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-455046

Unique identifier UC11362153

Identifier 7112387.pdf (filename),usctheses-c18-455046 (legacy record id)

Legacy Identifier 7112387.pdf

Dmrecord 455046

Document Type Dissertation

Rights Glezakos, Constantine

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA