Alternative Approaches to Reciprocal Tariff Liberalization

Transcription

1 Alternatve Approaches to Recprocal Tarff Lberalzaton Arvnd Panagarya* *Professor, Department of Economcs, Unversty of Maryland, College Park MD

2 Alternatve Approaches to Recprocal Tarff Lberalzaton Arvnd Panagarya 1 Artcle XXVIII bs of the General Agreement on Tarffs and Trade (GATT) calls upon the WTO members to perodcally sponsor negotatons amed at substantal reductons n the general level of tarffs on a recprocal bass. Paragraph 2(a) of ths artcle states, Negotatons under ths Artcle may be carred out on a selectve product-byproduct bass or by the applcaton of such multlateral procedures as may be accepted by the contractng partes concerned. Such negotatons may be drected towards the reducton of dutes, the bndng of dutes at then exstng levels or undertakng that ndvdual dutes or the average dutes on specfed categores of products shall not exceed specfed levels. The bndng aganst ncrease of low dutes or of duty-free treatment shall, n prncple, be recognzed as a concesson equvalent n value to the reducton of hgh dutes. Thus, GATT gves the member countres substantal flexblty wth respect to how tarffs are lowered. They can engaged n bargans wth ther major tradng partners by sectors as was the case n the earler rounds or may agree on a general formula applcable unformly to all members as was the case n the Kennedy and Tokyo Rounds or a combnaton of the two as n the Uruguay Round. 1 Professor and Co-drector, Center for Internatonal Economcs, Department of Economcs, Unversty of Maryland, College Park, MD Ths paper has benefted greatly from the excellent dscusson of 1

3 Multlateral lberalzaton under the auspces of GATT dffers from unlateral lberalzaton n one key respect: t brngs gans to a country not merely from ts own lberalzaton (assumng the country s small or abstractng from the terms of trade effects) but the lberalzaton of ts partners as well. 2 The more a country s able to get ts tradng partners wth large markets for ts products to lberalze, the more t stands to beneft. From the vewpont of developng countres, ths means optng for an approach that leads to the maxmum openng by developed countres n products of nterest to them. In trade negotatons, countres vew ther own lberalzaton as a cost to be pad for obtanng expanded access to the markets of partner countres. These costs may take at least three forms. Frst, f the country s large n the world markets, ts own lberalzaton may ncrease the world demand for the goods t mports suffcently that these goods become more expensve. For developng countres, whch are generally small n the world markets, ths rsk s qute mnmal. At the aggregate level, the proporton of ncome spent by developed countres on developng country goods s less than 2 percent. Second, trade lberalzaton makes mport competng goods cheaper relatve to exportable goods, whch n turn leads to a redstrbuton of ncome from the owners of factors specalzed n the producton of mport-competng goods to those tarff formulas n Hoekman and Kosteck (2001) as well as from excellent comments by Bernard Hoekman on an earler draft. 2 If a country s suffcently small n the world market, trade lberalzaton by t ncreases ts mports but ths ncrease s too small to have any effect on the world prces. In ths case, the country necessarly benefts from ts own lberalzaton snce t reduces the domestc output of goods n whch t lacks comparatve advantage and expands the output of the goods n whch t enjoys comparatve advantage. In contrast, f a country s a large buyer of ts mports n the world market, ncreased demand due to ts lberalzaton can lead to a rse n the prce of the mports. Ths terms of trade effect s harmful to the country and n prncple may be large enough to outwegh the effcency gans resultng from ncreased specalzaton n the goods of comparatve advantage. 2

4 specalzed n export goods. Ths may be poltcally costly. Fnally, the reallocaton of resources nduced by trade lberalzaton may tself brng short-run real costs wth t. The presence of these costs automatcally gves rse to the noton that tarff reductons should be balanced accordng to some crteron: the cost of one s own lberalzaton must be balanced by the benefts from lberalzaton by ts tradng partners. The obvous crteron for achevng ths balance wth respect to the terms-of-trade cost s to ensure that the terms of trade are unchanged n the post-lberalzaton equlbrum. Ths would effectvely allow each country to beneft from the effcency gans resultng from ts own lberalzaton wthout redstrbuton of ncome across countres. From the vewpont of small developng countres, ths may not be a major ssue snce ther own lberalzaton has lttler mpact on the terms of trade. But for large countres such as the Unted States, European Unon and Japan, t s mportant and plays out n the form of the demands for recprocty. As for the long-run ncome dstrbuton effects, unless one s faced wth the conservatve welfare functon such that any movement away from the exstng equlbrum s seen as harmful, there s no presumpton that the effects of trade are harmful. Indeed, to the extent that most developng countres are exporters of laborntensve goods, trade lberalzaton s lkely to mprove ncome dstrbuton by rasng the return to labor at the expense of captal. Fnally, phasng n lberalzaton over a number of years can mnmze the adjustment costs. Startng wth the Kennedy Round, ths has been the approach under the GATT and WTO. In the case of multlateral lberalzaton, these costs are lkely to be especally low snce the smultaneous 3

5 lberalzaton by partner countres allows the export ndustres to expand rapdly to absorb the resources released by mport-competng ndustres. In the followng, I dscuss the man approaches to trade lberalzaton that have been taken n past multlateral negotatng rounds and ther relatve merts. 1. Sectoral Approaches Conceptually, we can dstngush two types of sectoral approaches. Frst, one or more sectors may be dentfed and all members may lberalze mports wthn those sectors. Ths s sometmes called a zero-for-zero approach. Second, each member may seek lberalzaton from ts major tradng partners n sectors of ts comparatve advantage n return for ts own lberalzaton of sectors of comparatve dsadvantage. The frst approach has guded much of the lberalzaton of trade n servces durng and snce the Uruguay Round. The member countres have dentfed broad sectors such as fnancal servces or telecommuncatons and barganed market access wthn them. The approach has also been appled recently to trade n goods recently durng frst the WTO mnsteral meetng n Sngapore. At ths meetng, a group of countres sgned the Informaton Technology Agreement (ITA) whereby they commtted themselves to complete free trade n a set of technology products. As far as the lberalzaton n goods s concerned, ths sectoral approach s not partcularly attractve. There are two problems. Frst, there s some concern that under ths approach, the sectors that wll be lberalzed frst wll be those n whch economcally powerful natons have export nterests. Sectors n whch developng countres enjoy 4

6 comparatve advantage, such as textles and clothng, wll be taken up last. Ths vew s certanly supported by the fact that, on balance, developed countres are exporters and developng countres mporters of the products covered by the ITA. It s also supported by the observaton that even n servces, negotatons n sectors n whch developng countres have an overwhelmng advantage have not been opened. The second drawback of ths approach s that ntally t s lkely to lead to a lower level of welfare. Under ths approach, the sectors that wll be pcked ntally are lkely to be those wth lower tarffs to begn wth. Economc theory suggests that under plausble assumptons, the elmnaton of relatvely low tarffs wthout a smultaneous reducton n the hgh tarffs leads to reduced welfare globally as well as n ndvdual countres. Such a polcy change leads to a reallocaton of resources from less dstorted to more dstorted sectors. Under the second sectoral approach, all ndustral products are made part of the negotaton but members negotate wth ther partners sector by sector. In prncple, ths approach can be expected to yeld an effcent outcome. Each member wll seek lberalzaton n sectors and countres where ts exports face the hghest barrers. In turn, ts tradng partners wll seek access to ts most protected sectors. Thus, the bargan s based n favor of lowerng the hghest tarffs. The approach may become admnstratvely complex f trade patterns happen to be such that each country exports ts goods to one set of countres but mports them from an entrely dfferent set of countres. For nstance, f Inda s exports go manly to the Unted States whle ts mports come from the European Unon, blateral negotatons between Inda and ts two tradng partners become dffcult. Whle t remans to be 5

7 emprcally verfed, one suspects that under the current structure of trade and tarffs, ths s not a sgnfcant problem. Goods subject to hgh tarffs n developng countres are largely mported from developed countres and evenly dstrbuted over the latter. Conversely, major exports of developng countres face hgh barrers n all developed countres examples nclude agrcultural products and apparel Across-the-Board Approaches and Tarff-Reducton Formulas Rather than negotate on a sector-by-sector bass, member countres may adopt an acrossthe-board approach such that all tarffs are rolled back accordng to a pre-specfed formula. Of course, even n a sector-by-sector approach, some prevously agreed rule must be used to ensure a balance n the bargan (recprocty). Therefore, the formulas to be dscussed below are relevant for the sector-by-sector approach as well. Perhaps the smplest rule to follow s the reducton n all tarffs by a fxed percentage. For example, the member countres may agree to roll back all (or a subset of) exstng tarff by 50 percent. Formally, we can set dt (1) t = b where b s a constant between 0 and 1, t the ntal tarff rate n sector and dt /t the proportonate reducton n t. Ths approach has the advantage that wthn a country t leads to a larger absolute reducton n hgh tarffs and smaller absolute reducton n low 3 See Hoekman, Ng, and Olarreaga (2001) for an analyss of the ncdence of tarff peak protecton n Quad markets, net of preferences, on least developed countres. 6

8 tarffs, whch must broadly promote effcency. More protected sectors are lberalzed more and effectve protecton to any partcular sector s unlkely to rse on account of nput tarffs declnng proportonately more than output tarffs. If the countres engagng n lberalzaton have equal levels of tarffs and are also of equal sze, ths approach wll lead to balanced reductons n tarffs across them. The bargan wll be unbalanced, however, f ether of these condtons s volated. For nstance, f two countres are of equal sze but one has an average tarff of 50% but the other 5%, a 50% reducton n the tarff by the former leads to a 25-percentage ponts reducton but only 5 percentage ponts by the latter. Such lberalzaton leads to a deteroraton of the terms of trade of the former. Lkewse, f both countres have 50% tarffs but one s ten tmes the sze of the other, a 50% reducton by the former leads t to gve greater market access than t receves and hence a deteroraton of ts terms of trade. There are dfferent approaches that attempt to correct for these dfferences n the ntal tarffs and country sze. The smplest approach s to defne lberalzaton n terms of the tarff revenue forgone. For example, we can set the proportonate reducton n tarff equal to dt f f (2) = = t (t.p.m ) (t.v ) where f s a constant, p the border prce, M the quantty of mports, and V the value of ntal mports at the world prce. We can thnk of f as a measure of market access gven by the country undertakng lberalzaton. An advantage of tarff reductons accordng to (2) s that t takes nto account both the level of the ntal tarff rate and the sze of the country n the world market. To acheve the same level of lberalzaton, f, a country that mports larger volumes of good and mposes a hgher ntal tarff on the latter has to 7

9 lberalze proportonately less to acheve the same level of lberalzaton. Alternatvely, f the ntal level of tarff n a sector s low, the credt gven for a gven percentage reducton n the tarff s also low. Tarff reductons accordng to (2) are defensble from the vewpont of balancng the bargan between member countres. But, for a gven level of mports, t mples a lower proportonate reducton n the tarff whenever the ntal tarff s hgh. From an effcency standpont, ths may not be a desrable outcome. For nstance, f fnal goods are subject to hgh tarffs and nputs to low tarffs, ths pattern of tarff reducton may lead to ncreased effectve protecton to certan sectors. 4 Thus, there s necessarly a tenson between the twn objectves of balancng a bargan and preservng effcency across sectors. An alternatve measure that compromses somewhat on the bargan-balancng objectve but s more defensble on effcency grounds s dt f f (3) = = t (p.m ) V 4 The concept of effectve protecton refers to protecton provded to value added n a sector by the entre structure of tarffs. To explan, suppose the cost of auto parts n the world markets s 90 percent of the fully assembled automoble. Under free trade, domestc assembly of the automoble wll be proftable up to the pont that the margnal cost of assembly s $1,000. The ntroducton of a 10 percent duty on the fnal automoble wthout a smlar tarff on the mported nputs rases the domestc prce of the automoble to $11,000 and the domestc assembly ndustry can expand up to the pont that the margnal ts margnal cost s $2,000. Ths s an expanson of the domestc value added per automoble from $1,000 to $2,000 and amounts to a 100 percent protecton on the assembly operaton. Thus, the effectve protecton to automoble ndustry turns out to be ten tmes the nomnal protecton! If we were to also mpose a 10 percent tarff on auto parts, the domestc value added can expand from $1,000 to $1,100 (snce parts now cost $9,900 and the automoble sells for $12,000), whch s exactly 10 percent. 8

10 Ths measure controls for country sze n that the larger the V the smaller the proportonate reducton n the tarff to acheve the gven level of lberalzaton, f. But the measure s ndependent of the ntal level of the tarff. Yet another approach goes a step further towards effcency by rollng back the hgh tarffs proportonately more and lower tarffs proportonately less. From the recprocty angel, the argument made s that lnear cuts leave the countres wth hgh ntal tarffs more protected. Ths argument s clearly based on achevng recprocal concessons n terms of fnal outcomes rather than addtonal market access. Put dfferently, t must rely on the assumpton that tarff levels must be harmonzed across countres. If partcpatng countres share the objectve that the eventual goal s to acheve free trade everywhere, the approach clearly makes sense. Nonetheless, t must be recognzed that f countres are of equal sze but some are ntally more protected than others, n the transton towards zero tarffs, ths approach wll result n the latter experencng deteroraton n ther terms of trade. The explanaton s smlar to that presented earler and s understood smply by assumng that one country has a 50% tarff on ts mports whle the other has only 10% tarffs. Elmnaton of these tarffs wll result n the former openng ts market more at the margn and hence a deteroraton of ts terms of trade. The smplest rule to acheve harmonzaton, suggested by the European Economc Communty (EEC) durng the Tokyo Round, was to lower each tarff by the same percentage as ts ntal ad valorem rate. Thus, f the tarff on a good s 70 percent, t s cut by 70 percent whle a tarff of 20 percent s cut by 20 percent. Formally, we have (3a) dt t = t 9

11 The EEC argued that as a part of the Tokyo Round, countres apply ths formula four tmes, brngng tarff rates n excess of 50% (but less than 100%) down to below 13%. The Unted States suggested that all tarffs equal to or above 6.67 percent be cut by 60 percent whle those below ths number be cut accordng to the formula (3b) dt t = 1.5t Ths formula s less progressve than (3a). Gven that the U.S. tarff structure pror to the Tokyo Round was characterzed by much greater escalaton than the European tarff structure, t s not surprsng that the Unted States wanted a tarff-cuttng formula that was much closer to the lnear cut. Observe that accordng to the U.S. approach, a tarff of.02 (or 2%) would be cut by.53 (or 53%) and a tarff of.06 by 59% whle all tarffs equal to or hgher than 6.67 percent would be cut by 60 percent. There s only a small varaton n cuts accordng to the ntal level of the tarff rate. The formula actually appled n the Tokyo Round was the one suggested by Swtzerland. Accordng to the so-called Swss formula, dt t (3c) = t r+t where r s a postve constant. Ths formula has a progressve element the degree of whch depends on the value of r. The hgher the value of r, the more progressve s tarff reducton n the sense of lowerng the hgher tarffs more. In the Tokyo Round Agreement, the value of r was set between 0.14 and Lettng r = 0.14, tarff rates of 0.14 (or 14 percent) are reduced by 50 percent. Rates exceedng 0.14 are reduced more and those less than 0.14 are reduced less. For nstance, a tarff rate of 0.06 (or 6 percent) s reduced by 30% whle a tarff rate of.36 (or 36 percent) s reduced by 72 percent. 10

12 Clearly, ths formula has much greater potental for harmonzng tarff rates both across sectors wthn a country and across countres wthn the same sector. Fnally, countres may agree to some average tarff reducton wthout specfyng the reductons n specfc, sectoral tarff rates. The average may be smple or weghted. Ths approach can result n both unequal and neffcent tarff reductons ex post. The Uruguay Round Agreement on Agrculture requred developed countres to reduce tarffs by 36 percent on the average, wth the rate on each tem reduced by at least 15 percent. Assumng four tems n all wth 100 percent tarffs on three and 1 percent tarff on the fourth, a 15 percent reducton n the former rate and elmnaton of the latter yelds ( )/4 = percent average reducton. Thus, rules n terms of an overall average reducton can be easly manpulated to mnmze lberalzaton Choosng Among Approaches A choce among varous approaches depends on the underlyng objectve. If the objectve s to acheve maxmum lberalzaton worldwde, an across-the-board approach that lowers hgher tarffs more such as that based on the Swss formula would be the rght choce. The across-the-board approach mnmzes the room for successful lobbyng by poltcal powerful sectors, whch often happen to be the most protected sectors n the frst place. Moreover, a formula that lowers hgh tarffs more reduces the dsperson n tarffs and hence effectve protecton n all sectors. A formula approach also has the advantage 5 Specfyng the reducton n terms of a weghted average does not solve the problem snce the country can acheve a hgh average by lberalzng more where ts mports are already hgh whle leavng the most protected sectors vrtually untouched. 11

13 that t does not te up negotatng resources n a major way as do sector-by-sector negotatons. From an ndvdual country s vewpont, the answer may be dfferent. Assumng welfare maxmzaton to be the objectve, a large country wth hgh tarffs s lkely to prefer equal proportonal tarff reductons whle one wth low tarffs may prefer Swssformula lke reductons. For the hgh-tarff large country, proportonal reductons wll yeld better terms of trade than the Swss-formula lke reductons whle for the low-tarff large country the opposte holds true. Governments that are drven by domestc lobbes are lkely to prefer the sectorby-sector approach snce ths approach lends tself most easly to offer or seek the sectors selectvely for lberalzaton. Import-competng lobbes wth poltcal clout can ensure that ther sectors are not offered for lberalzaton. Lkewse, powerful export lobbes may pressure ther governments to seek market access on ther behalf from ther major mporters. In my judgment, n the present scenaro, from the vewpont of developng countres, a Swss-formula lke approach would make the most sense. Ths s because developed countres apply rather hgh tarffs on goods of export nterest to them ([see Annex 1 of ths Handbook]). Therefore, the Swss-formula lke approach wll nduce large lberalzaton n the products of nterest to them. Of course, gven hgh tarffs n many developng countres themselves, ths approach wll requre them to lberalze more as well. But snce these countres have mnmal market power, ther own lberalzaton s lkely to result prncpally n effcency gans wthout deteroraton of ther terms of trade. 12

14 A smlar argument apples even more to trade n agrculture snce many developng countres are, n fact, potental exporters of agrcultural products. 13