Further Thoughts on Strategic Trade Policy under Asymmetric Information

Transcription

1 Modern Eonomy, 03, 4, Publshed Onlne November 03 ( Further Thoughts on Strateg Trade Poly under Asymmetr Informaton Chung Yuan Fu *, Shrley J o Department of Eonoms Natonal Chengh Unversty, Tawan Emal: * fuhungyuan@gmalom, sho@nuedutw Reeved July 3, 03; revsed August, 03; aepted August 9, 03 Copyrght 03 Chung Yuan Fu, Shrley J o Ths s an open aess artle dstrbuted under the Creatve Commons Attrbuton ense, whh permts unrestrted use, dstrbuton, and reproduton n any medum, provded the orgnal work s properly ted ABSTRACT We study the nformatonal mpats of multlateral voluntary export restrants (heneforth VERs) n an nternatonal trade model wth dfferentated produts [] We frst show that wth ompetng mehansms, the two frms lyng ntentons are strateg omplements and wll nrease wth the degree of produt dfferentaton Next, we show that eah government wll desgn ther VERs menus to allow for only partal revelaton Contrary to the sngle nterventon ase [], a separatng eulbrum where eah ountry s domest frm truthfully reveals ts prvate nformaton does not exst under multlateral poly nterventons Fnally, we demonstrate that trade retalaton, when the two governments VERs are postvely related, wll happen when the government beleves that ts domest frm s more lkely to be neffent Keywords: Strateg Trade Poly; Voluntary Export Restrants; Partal Informaton Revelaton Introduton It s now well known that government nterventon an shft rents by provdng a strateg advantage to the domest frm In partular, the poneerng work by Brander and Spener [3] showed that under Cournot ompetton, an export subsdy enables the domest frm to be a Stakelberg leader, and thus t nreases the domest welfare at the expense of the foregn frm Subseuently, Eaton and Grossman [4] demonstrated that the optmal poly wll be an export tax, when the domest frm ompetes wth a foregn frm n pres Ths ndates that the type of export poly s senstve to the form of ompetton In addton, the lterature has also noted that nformaton an also rual n determnng the approprate poly For example, Wong [5] demonstrated that n the Brander-Spener model wth asymmetr nformaton (about ost), the optmal export subsdy sheme derved from the full nformaton ase s no longer nentve ompatble Colle and vd [6] and * Correspondng author De Meza [7] onsdered ost asymmetry between frms and shows that the ountres wth the lowest osts provde the hghest export subsdes Allowng a soal ost of publ funds that exeeds unty, Neary [8] smlarly found that non-onavty of demand s a suffent ondton for the government to provde more subsdes to the more ost ompettve frm Bandyopadhyay [9] found the onventonal result n De Meza [7] and Neary [8] s reversed for nelast demand Qu [] examned the use of strateg trade poly as a sgnallng deve when the nformatonal asymmetry s between domest and foregn frms Wth nomplete nformaton, the Prnpal-Agent model an best desrbe the leader-follower relaton between unnformed government and nformed frm By applyng the Revelaton Prnple, the unnformed government an adopt a dret mehansm by offerng a menu of poles and lettng the nformed domest frm self-selet the ntended poly When domest frm s ompetng wth foregn frms, ths self-seleton proess beomes more nformatve In Qu s [] words, t s a mx of sreenng and sgnallng problems ; by hoosng among the menu of poles, the domest frm also sgnals ts prvate nformaton to the rval frms [6] The task of the government s then to trade off between the neffeny from asymmetr nformaton and the strateg advantage from trade poly; whether to have the nformed domest frm truly reveal ts nformaton, or to hde and enoy the strateg beneft from asymmetr nformaton? Qu [] showed that under Cournot ompetton the unformed government wll hoose a poly menu to truly reveal ts ost nformaton (separatng eulbrum) Whle the analyss on unlateral poly nterventon has proved to be powerful to desrbe the dlemma enountered by the unnformed government, n realty we often

2 697 see blateral or multlateral poly nterventons nstead of unlateral poly If we adopt a dret mehansm for these multple ontratng ases, wll there be a separatng eulbrum? If not, how muh nformaton an be revealed? ow does the degree of nformaton revelaton relate to the market struture? Wll there be trade retalaton? In ths paper, we provde answers to these uestons by studyng an nomplete nformaton VERs game n an ntra-ndustry trade model The model we onsder s an nternatonal trade model wth dfferentated produts [] There are two ountres and eah ountry has only one produer, whh sells ts produts to two ountres In order to apture ntra-ndustry trade of ths knd, we employ the dfferentated produt framework by Dxt [0] and Sngh and Vves [] There are three reasons to study an nomplete nformaton VERs game of ths form Frst, there have been many dsussons on the nformatonal mpats of taxes, subsdes or tarffs; we hope to omplement the lterature by nvestgatng the nformatonal mpats of VERs, whh are seen as eually mportant strateg tools Seond, although we are aware that both the hoe of poly nstrument and the degree of nformaton revelaton an be senstve to the form of ompetton (Cournot or Bertrand), sne our fous s on the nformaton aspets of multlateral nterventons, we adopt a uantty ompetton setup to better handle the mpats from VERs Thrd and most nterestngly, we wll show that n the omplete nformaton benhmark ase, eah government s VER deson s as effent as n the sngle government ase There s no dret nteraton between the two governments, so the poly effeny an be retaned even wth blateral nterventons The ueston of onern s: now that there s no game between the two governments, an we apply the revelaton prnple dretly and look for a truth-tellng dret mehansm n the nomplete nformaton VERs game? Unfortunately, the answer s no, as we wll demonstrate that the sgnalng effet of menu seleton wll hange the rval frm s perepton about domest frm s prvate nformaton Conseuently, eah frm s ntenton of nformaton revelaton wll be related to the rval frm s ntenton, and thus the two governments are no longer ndependent from eah other Our paper starts wth the omplete nformaton benhmark ase, where we show that, by usng VER eah government gans a frst mover advantage n the rval ountry [-4] owever, when onsderng nomplete nformaton, we frst demonstrate that wth ompetng mehansms, eah frm s ntenton to le s postvely related to the rval frm s lyng ntenton The two frms lyng ntentons are strateg omplements and wll nrease wth the degree of produt dfferentaton Importantly, we show that eah government wll desgn ther VERs menus to allow for only partal revelaton Contrary to the sngle nterventon ase [], a separatng eulbrum where eah ountry s domest frm truthfully reveals ts prvate nformaton does not exst wth multlateral poly nterventons Fnally, we demonstrate that trade retalaton, when the two governments VERs are postvely related, wll happen when the government beleves that ts domest frm s more lkely to be neffent Ths result partly reflets the result by Martna and Vergoted [5], who dsussed the role of retalaton n trade agreements and showed that retalaton s a neessary feature of any effent eulbrum The ssues on ompetng mehansms have reeved many dsussons Pek [6] and Martmort and Stole [7] frst llustrated apparent falures of the standard revelaton prnple wth ompetng mehansms Sne there s no obvous way to deal wth these problems, the lterature has responded by mposng ad ho restrtons on the set of mehansms from whh Prnpals an hoose [8-] Ths s the reason why we stk to dret mehansms n ths urrent paper Next, our paper nvestgates ex-post nformaton revelaton under blateral government nterventons Ths s dfferent from ex-ante nformaton revelaton suh as Creane and Myagwa [3], where duopoly frms make ther revelaton desons before they observe ther own prvate nformaton From the nformaton ontent, our model s losed to Colle and vd [6] and Qu [], but they manly assumed unlateral poly desgn Branard and Martmort [] onsdered blateral government nterventons but restrted to truth-tellng eulbra Fnally, our results onlude that trade retalaton happens when the government beleves that ts domest frm s more lkely to be neffent Ths partly ondes wth Martna and Vergoted s [5] results They showed that n the presene of prvate nformaton, retalaton an always be used to nrease the welfare derved from suh agreements by the partpatng governments In partular, t s shown that retalaton s a neessary feature of any effent eulbrum The remander of the paper s organzed as follows Seton dsusses the omplete nformaton VERs game as a benhmark of omparson We demonstrate that, by usng VER, eah government gans a frst mover advantage n the rval ountry Seton 3 haraterzes the eulbrum n the nomplete nformaton VERs game In eulbrum, eah government wll only mplement partal revelaton from ts domest frm, and there an be trade retalaton when the government beleves that ts Branard and Martmort [] restrted to truth revelaton eulbrum, beause Myerson [4] showed that n a blateral prnpal-agent struture, truth revelaton wll be an eulbrum, f eah agent s assoated unuely wth one prnpal In our model, eah frm wll be related to both governments, and hene truthful revelaton does not onsttute an eulbrum

3 698 domest frm s more lkely to be neffent Seton 4 onludes the paper wth some suggestons on further researh The Model Spefally, let the supersrpt k,, ndex the two ountres Country k s demand for frm and frm s produts are gven by k k k p, k k k p, k k where subsrpt,,, ndexes frm and denote frm s pre and output n ountry k The oeffents and denote the own pre effet and ross pre effets, respetvely We assume to reflet that the own pre effet s hgher than the ross effets Note that the level of an be seen as a measure for the substtuton between the two produts; When the two produts are almost homogenous; when 0 the two produts are almost dfferentated The reason we have onsdered a uantty nstead of pre ompetton model s beause we an handle the uantty restrants easly We are aware that the form of ompettonmght hanges the poly nsghts [4] Sne our fous s on the nformatonal mpats of VERs, we wll stk to ths smple framework for a neat presentaton Fnally, sne our fous s on the nformatonal mpats of government nterventons, we assume zero transportaton ost for smplfaton Eah frm s margnal produton s assumed to be Eah frm s proft s thus gven by π π R R,,, where R p and R p denote frm s revenue n domest and foregn ountres, respetvely We onsder that the government of eah ountry wll hoose a VER,, to maxmze soal welfare SW Eah ountry s soal welfare s the overall utlty deduted by foregn frm s revenue, added by domest frm s revenue n the foregn ountry, and mnus the domest frm s total produton ost That s, SW U R R, for,,and Ths setup s frstly gven by Sngh and Vves [], whh assumed that U, The advantage of ths setup s: sne n dfferentated produt models, we annot use the area under demand funton to measure the onsumer surplus The setup of an help us easly measure the onsumer surplus U p Also, we an easly derve the ountry s demand funton by partal dfferentatng wth respet to U Complete Informaton VERs Game As a benhmark of omparson, Seton dsusses the omplete nformaton VERs game where both frms produton osts are publly known We wll show that, under omplete nformaton, eah government s VER deson wll be as effent as n the sngle government ase owever, ths poly effeny wll dsappear when we onsder nomplete nformaton n Seton 3 Not only beause there s nformaton rent n the VERs ontrat, but also beause the two governments ompetng mehansms wll nrease frms strateg nentves to le The omplete nformaton VERs game proeeds as follows Frst, government and government set ther VERs,, and smultaneously After observng the VER desons, both frm and frm ompete n the produt markets of the two ountres By bakward nduton, we frst solve the market eulbrum, gven the two governments VER desons, and then determne eah government s optmal VER Market Eulbrum Gven the two governments VER desons, and frm and frm maxmze ther profts smultaneously max π, R R max π R R Eah frm s best reply to the rval government s VER s gven by and () Optmal VERs Now, gven the two frms best reples n (), eah government hooses ts VER to maxmze ts soal welfare In the ase of ountry, the strutural form of SW s: SW From the frst order ondton of maxmzaton: SW 0 3 Therefore 3 The seond order ondton of maxmzaton s satsfed

4 699 () Smlarly, we an alulate government s optmal VER: We wll later refer and as the effent VERs, as eah government s VER deson s as effent as n the sngle government ase Aordngly, substtute,, nto frms best reples n (), we have and ere we make two remarks on these effent VERs Frst, we an ompare to the outputs n the free trade ase, where eah frm hoose both and,, to maxmze The free trade outputs n domest and foregn markets are and 3, respetvely A dret omparson 3 shows that The VER s hgher than the foregn output n the free trade ase The advantage of usng VER s that, by pre-ommttng to, the government gans the poston of a leader n the rval ountry, and n output ompetton, there wll be a frst mover advantage Ths strateg advantage s frst mentoned by arrs [], who analyzed the mpats of VERs n a Bertrand model Rosendorff [5] explaned why governments prefer VERs to tarff n a Cournot model Also, Ishkawa [6] studed the effet of VERs on proft, market share, onsumer surplus and welfare n a Cournot model Berry, et al [7] evaluated VERs that was ntally plaed on automobles exports from Japan n 980 s They found that VERs had nreased both pres and the profts of domest frms, whle leavng onsumer welfare worse off Feenstra and ews [8] onsdered a domest government wth poltal pressure to negotate over the volume of trade and the transfer of rents They haraterzed the globally optmal, nentve-ompatble trade poles, n whh the domest government has no nentve to overstate (or understate) the pressure for proteton De Sants [4] studed the mpat of VERs on exportng ountres e showed that VERs at the free-trade level would favour the onentraton of ndustry, and rase the pre mark-up n the domest market owever, the mpat on welfare s ndetermnate dependng upon the effet on global effeny Seond, n () tells us somethng about the motvaton of mmkng Note that s dereasng n, and ths ndates that a more effent frm wll have a hgher VER In partular, f has two possble values: and wth, then we have Ths, however, does not mply that the less effent frm wll mm the effent frm Atually, sne s onave n, and by the defnton of maxmzaton, the less effent frm s better off hoosng than hoosng In the words of Spene [9], the neffent frm does not envy the effent frm Ths means that when we onsder nomplete nformaton n a unlateral nterventon ase, a separatng eulbrum where eah type of hooses ts ntended VER mght exst ene, our model would suggest the same result as Qu [] n the unlateral nterventon ase In the next seton, we wll show that wth multple mehansms, a separatng eulbrum does not exst 3 Inomplete Informaton VERs Game Seton 3 dsusses the nomplete nformaton VERs game where s only prvately known by frm Nether government nor government or frm knows ths value To smplfy the analyss, we assume a bnary type set,, wth We have shown that wth omplete nformaton, eah government s VER deson s as effent as n the sngle government ase There s no dret nteraton between the two governments, so the poly effeny an be retaned even wth blateral nterventons Wth nomplete nformaton, we wll show that the sgnalng effet of menu seleton wll hange the rval frm s perepton about domest frm s prvate nformaton Conseuently, eah frm s ntenton of nformaton revelaton wll be related to the rval frm s ntenton, and the two governments are no longer ndependent from eah other The nomplete nformaton VERs game proeeds as follows Frst, eah government announes a VERs, ndependently Seond, eah menu frm self selets a VER from the menu, and ths hoe s publly observed Thrd, aordng to the observed poly hoes, the two frms update ther belefs about the rval frm s produton ost, and then ompete n the produt markets of the two ountres By bakward nduton, we frst solve the market eulbrum gven the two frms menu seleton Then we haraterze the two frms menu seleton eulbrum Fnally, we determne eah government s VERs menu Before proeedng wth the dervaton of market eulbrum, we defne more notatons for the pror and posteror belefs on Frst, as sad, we assume that, s only prvately known by frm All other players (nludng government, government, and frm ) have ommon pror belefs that the prob-

5 700 ablty for s, 0 and the probablty for s After observng the rval frm s menu seleton strategy, eah frm an update ts belef on, et, denote frm s seleton strategy from the menu, It s assumed that,, for That s, n, the seleton strategy for type of frm s, and for type of frm, t s In partular,, 0 denotes the separatng strategy where eah type of frm selets ts ntended VER As another example,, denotes the hybrd strategy where type of frm selets ts ntended VER, whle type of frm randomly selets be- wth a probablty tween and 3 Market Eulbrum Belef Updatng We now solve the market eulbrum gven the two frms menu seleton After observng frm s seleton strategy, frm an update ts belef on That s, aordng to the Bayes rule, gven the observaton, the on-eulbrum path belef s gven by: For smplfaton, we assume that the off-eulbrum path belef wll be the same as the pror If then t an be alulated that In partular, for the separatng strategy, 0, we have As another example, for the hybrd strategy,, we have and Fnally, gven the posteror belef, let E denote frm s posteror expeted ost Market Competton et denote ountry s VER assoated wth seleton strategy Sne eah frm s proft wll be affeted by ther seleton strateges, we rewrte the profts as π,,,,, where π,, E, (3) (4) E π,, In the ase of π,,, there s an expeted term E The reason for the expetaton form s beause frm annot observe, and by observng, frm wll guess frm sver as: So the ex- E Gven, of frm takes s type of frm takes s peted probablty of takng s (5), the probablty that type, and the probablty that The expeted probablty of takng an be explaned smlarly In the produt market, eah frm hooses to maxmze π,,, gven the seleton strategy, Note that,, wll be determned by the menu seleton strategy By the frst order ondton of maxmzaton 4, we have E E and (6) Ths s smlar to () n the omplete nformaton ase, exept that the expeted VERs wll be determned n the menu seleton game Substtute the two best reples n (6) to (4) and (5), we π,, as: an rewrte π,, π,, E E E E E E 4 The seond order ondton of maxmzaton s satsfed, (7) (8)

6 70 In the ase of,,, note that there s a posteror expeted ost term E From (6), we know E that frm s best reply s But sne frm annot observe, t an only use the observaton of to update ts belef to be E The smlar argument apples to frm 3 Menu Seleton Frst, reall that n the omplete nformaton ase, we onlude that type s better off hoosng than In the unlateral nterventon ase, a separatng eulbrum where eah type of hooses ts ntended VER mght exst ene, to smplfy the dsusson, we restrt the seleton strategy, to be, That s, we assume that type wll take the ntended VER, whle type mght mm type by takng a mxed strategy Wth ths restrton, 0 wll ndate the separatng strategy where eah type hooses ts ntended VER, and denotes the poolng strategy where both types and hoose (7) and (8) an be rewrtten aordngly E n be a wrtten as and E, E an be rewrtten as E (9) (0) emma onludes a prelmnary result whh wll be useful for eulbrum haraterzaton E emma 0 E Proof Sne 0 0, we have E 0 and Next, we derve the eulbrum seleton strateges,, gven the best reples n (6) That s, for, and,, frm maxmzes,, wth respet to In the ase of frm, after replang,,,, n (7) s rewrtten as (see Euaton () ): In,, type wll take the ntended VER, and type wll mm type by takng a mxed strategy So n (), we have replaed n (7) by Also n (), we have replaed n (7) by Also, we have replaed E and E wth the defntons gven n (9) and (0) The smlar argument apples to frm Seleton Eulbrum Gven the announed VERs,, wll onsttute a menu, Bayesan eulbrum ff for,, max,, for, We need to show that for,, and an maxmze,, for,, respetvely Frst, from (), f we reure to be best reply for type, the ntended VER must be set at the followng level π,, E E E, E π,, E E () ()

7 70 E, (3) whh s obtaned by dfferentatng () wth respet to s set to be (3), then In other words, f wll be best reply Next, for type, needs to be best reply to, That s, gven, the frst order ondton of maxmzaton for () s: π,, E et B, frm s best reply s: E (4) B Smlarly, we an alulate frm s best reply to, : E, B where B (5) Note that s related to through the term E The eulbrum, needs to smultaneously satsfy (4) and (5) owever, sne the struture forms of eulbrum, are omplated, we derve the followng propertes on the menu seleton eulbrum Frst, from (4), t an be alulated that E B 0 B A smlar argument on also shows that 0 Ths ndates that and are strateg omplements Sne the level of denotes the degree that type of frm wll mm type the more that type of frm les wll trgger type of the rval frm to le more Next, reall that the level of measures the degree of substtuton, and the smaller ndates a hgher degree of produt dfferentaton If we take the partal dfferentaton of wth respet to, we know that the degree of strateg omplements s postvely related to the degree of produt dfferentaton emma and are strateg omplements, and s dereasng n Seond, t s nterestng to know f the separatng eulbrum exsts That s, n the ase of frm, we ask f 0 s a best reply to 0 n (4) To fnd out, substtute 0 n (4), so we have E and frm s best reply n (4) be omes The only hane for B 0 s to let Compare ths level to the effent VER We an onlude that f, these two values are dental, but f, s hgher In other words, the effen t VERs an ndeed ndue a truth-tellng eulbrum but on ly when More nterestngly, f we onsder an arbtrary postve level of, the VER for (4) to be zero s: E from (3) We therefore have the m, whh wll ndue eah type of frm to tell th e truth, no matter what the rval wll do In other words, gven the menu,, the domnant strategy for eah type of frm s to tell the truth and pk the ntended VER emma 3 ) Gven the menu,, the domnant strategy for eah type of frm s to tell the truth and pk the ntended VER ) The effent VERs an ndeed ndue a truth-tellng eulbrum but only when In the nomplete nformaton VERs game, the sg- Also reall enu,

8 703 nalng effet of menu seleton wll hange the rval frm s perepton about domest frm s prvate nformaton Conseuently, eah frm s ntenton of nformaton revelaton wll be related to the rval frm s ntenton The ueston of onern s whether the government wll fnd t optmal to ndue truthtellng, no matter how the rval frms mght le Wll t be better off for the government to ndue a ertan degree of lyng n eulbrum? ow wll ths degree relate to the market struture? We wll provde answers to these uestons shortly n next subseton Fnally, t an be alulated from (4) for the followng prelmnary results emma 4 0 and 0 As llustrated n Fgure, nreasng nward When wll shft the best reply of nreases, the menu seleton eulbrum wll move from E 0 to E Sne and are strateg omple- ments, ths ndates that both and wll de- rease Next, 0 ndates that th e smaller s, the m ore the best reply of wll shft outward In other words, when the two produts beome more dfferent- ated, the more lkely that type wll le Intutvely, when the two produts are more dfferentated, eah frm wll hope to nrease ther atual export to ounterat the demand reduton from a smaller Sne we are restrtng, type of eah frm has more ntenton to le and hene beomes hgher 33 Eulbrum VERs Gven the market eulbrum n (6) and the menu seleby (4) and (5), we ton eulbrum determned now λ 0 E 0, 0 0 Fgure Inreasng shfts the best reples leftward, E 0 λ determne eah gover nment s menu of VERs As sad, we have smplfed the dsusson by restrt the seleton strategy, to be, In the ase of frm, aordng to the dsusson on menu seleton, f we reure to be best reply for type, the ntended VER must be set at the followng level E (6) The same argument an apply to frm ene we are left wth the determnaton on As mentoned n emma 3 that the menu, (from (3) and (6)) an ndue 0 for any level of The ueston of onern s whether the government wll fnd t optmal to ndue truth-tellng from ts own frm, no matter how the rval frms mght le To answer the ueston, we frst alulate the frst or EW der ondton of maxmzaton: 0 In the SW from Seton : ase of frm, reall R RR SW U R The expeted soal welfare s: EW SW SW Sne the strutural forms are omplated, we omtted for the detaled der- EW 5 vaton of Then, by applyng the mplt funton theorem, we an alulate, the sgn of whh an tell us wether the two poles are strateg omplements or substtutes Fnally, we hek whether t s domnant for government to hoose,, suh that 0 The same argument an apply to frm Proposton 5 and are strateg * omplements for, and they are strateg substtutes for * Reall from emma and 4 that and are strateg omplements and 0 Proposton 5 says that the key to udge the strateg relaton between and s the sze of pror belef In the ase of government, f she thnks that the domest * frm s more lkely to be effent (e, ), then, the best reply of when government nreases 5 Detaled dervatons are avalable upon reuest

9 704 wll shft nward Sne and are strateg omplements, the eulbrum, wll both de- Now, f dereases, then the best reply rease of wll shft outward, whh ould ause the eulbrum to nrease rather than derease Contrarly, f government thnks the domest frm s * more lkely to have (e, ), then t s better to allure the type of the rval frm to le less To do so, when government nreases and move the best reply of nward, government must also nrease and move the best reply of nward to derease both of and n eulbrum Martna and Vergoted [5] dsussed the role of retalaton n trade agre ements They showed that n the presene of prvate nformaton, retalaton an always be used to nrease the welfare derved from suh agreements by the partpatng governments In partular, t s shown that retalaton s a neessary feature of any effent eulbrum Our results show that retalaton an * only happen when Proposton 6 Gven any level of, t s optmal for the government to mplement 0 In the ase of frm, to see wheth er t s optmal for government to hoose, (from (3) and (5)) to mplement 0, we substtute EW and nto the frst order ondton In the Appendx, we show that, EW 0 under the 6 VERs menu Due to the onavty of EW, we an onlude that the optmal must be smaller than By emma, sne 0, we an onlude that gvenany level of, t s optmal for government to mplement 4 Conludng Remarksss We study the nformatonal mpats of multlateral vol- s verson of untary export restrants n a heterogenou Brander and Krugman s [30] nternatonal trade model Smlar to the unlateral nterventon ase suh as Qu [] and Colle and vd [6], we ask f separatng eulbrum where eah ountry s domest frm truthfully reveals ts prvate nformaton stll exsts If not, how muh nformaton an be revealed? ow does nformaton revelaton relate to market struture? Wll there be trade retalaton? 6 Detaled dervatons are avalable upon reuest To these uestons, we frst showed that wth ompetng mehansms, the two frms lyng ntentons are strateg omplements and wll nrease wth the degree of produt dfferentaton Next, we showed that eah government wll desgn ther VERs menus to allow for only partal revelaton Contrary to the sngle nterventon ase [], a separatng eulbrum where eah ountry s domest frm truthfully reveals ts prvate nformaton does not exst wth multlateral nterventons Fnally, we demonstrated that trade retalaton, when the two governments VERs are postvely related, wll happen when the government beleves that ts domest frm s more lkely to be neffent As mentoned, we have restrted our dsusson on the VERs game to an ntra-ndustry trade model, for both analytal onvenene and for omplementaton to the lterature It s also nterestng to extend our model to dsuss other poly nstruments suh as subsdes or taxes, and n another form of ompetton suh as Cournot and Bertrand models The lterature has shown that the type of export poly s senstve to the form of ompetton In the ase of subsdes or taxes, the nformatonal mpats wll not be as surprsng as n VERs game, as the two governments are already related to eah other wth omplete nformaton We wll leave these nterestng ssues for further researh REFERENCES [] Wang, C Peng and Wu, Tarffaton and Welfare n a Dfferentated Duopoly, The World Eonomy, Vol 36, No 7, 03, pp [] D Qu, Optmal Strateg Trade Poly under Asym- Vol 36, No 3-4, 994, pp metr Informaton, Journal of Internatonal Eonoms, [3] J A Brander and B J Spener, Export Subsdes and Internatonal Market Share Rvalry, Journal of Internatonal Eonoms, Vol 8, No -, 985, pp [4] J Eaton, and G Grossman, Optmal Trade and Industral Poly under Olgopoly, Quarterly Journal of Eonoms, Vol 0, No, 986, pp [5] K Wong, Inentve Inompatble, Immserzng Export Subsdes, Unversty of Washngton Dsusson Paper No 900, 990 [6] D R Colle and M vd, Export Subsdes as Sgnals of Compettveness, Sandnavan Journal of Eonoms, Vol 95, No 3, 993, pp [7] D De Meza, Export Subsdes and gh Produtvty: Cause or Effet, Canadan Journal of Eonoms, Vol 9, No, 986, pp [8] J P Neary, Cost Asymmetres n Internatonal Subsdy

10 705 Games: Should Governments el p Wnners or osers? Journal of Internatonal Eonoms, Vol 37, No 3-4, 994, pp [9] S Bandyopadhyay, Demand Elasttes, Asymmetry and Strateg Trade Poly, Journal of Internatonal Eonoms, Vol 4, No -, 997, pp [0] A Dxt, A Model of Duopoly Suggestng a Theory of Entry Barrers, Bell Journal of Eonoms, Vol 0, No, 979, pp [] N Sngh and X Vves, Pre and Quantty Competton n a Dfferentated Duopoly, The RAND Journal of Eonoms, Vol 5, No 4, 984, pp [] R arrs, Why Voluntary Export Restrants Are Volun- Vol 8, No tary, The Canadan Journal of Eonoms, 4, 985, pp [3] C C Ma and Wang, Why Voluntary Export Re- The Canadan Jou- strants Are Voluntary: An Extenson, rnal of Eonoms, Vol, No 4, 988, pp [4] R A De Sants, Why Exportng Countres Agree to Voluntary Export Restrants: The Olgopolst Power of the Foregn Suppler, Sottsh Journal of Poltal Eonomy, Vol 50, No 3, 003, pp [5] A Martna and W Vergoted, On the Role of Retalaton n Trade Agreements, Journal of Internatonal Eonoms, Vol 76, No, 008, pp [6] J Pek, A Note on Competng Mehansms and the Re- 995 [7] D Martmort and Stole, Communaton Spaes, Eu- velaton Prnple, Oho State Unversty, Mmeo, lbra Sets and the Revelaton Prnple under Common Ageny, Unversty of Chago Graduate Shool of Busness, Dsusson Paper STE09, 997 [8] P MAfee, Mehansm Desgn by Competng Sellers, Eonometra, Vol 6, No 6, 993, pp [9] M Peters, On the Euvalene of Walrasan and Non- Walrasan Eulbra n Contrat Markets, Revew of Eonom Studes, Vol 64, No, 997, pp [0] M Peters and S Severnov, Competton among Sellers Who Offer Autons Instead of Pres, Journal of Eonom Theory, Vol 75, No, 997, pp [] S Branard and D Martmort, Strateg Trade Poly wth Inomplete Informed Polymakers, Journal of Internatonal Eonoms, Vol 4, No -, 997, pp [] G Epsten and M Peters, A Revelaton Prnple for Competng Mehansms, Journal of Eonom Theory, Vol 88, No, 999, pp [3] A Creane and K Myagwa, Informaton and Dslosure n Strateg Trade Poly, Journal of Internatonal Eonoms, Vol 75, No, 008, pp [4] R Myerson, Optmal Coordnaton Mehansms n Generalzed Prnpal-Agent Problems, Journal of Mathematal Eonoms, Vol 0, No, 98, pp [5] B P Rosendorff, Voluntary Export Restrants, Antdumpng Proedure, and Domest Polts, Ameran Eonom Revew, Vol 86, No 3, 996, pp [6] J Ishkawa, Who Benefts from Voluntary Export Restrants? Revew of Internatonal Eonoms, Vol 6, No, 998, pp / [7] S Berry, J evnsohn and A Pakes, Voluntary Export Restrants on Automobles: Evaluatng a Trade Poly, Ameran Eonom Revew, Vol 89, No 3, 999, pp [8] R C Feenstra and T R, ews, Negotated Trade Reg/0307/ strtons wth Prvate Poltal Pressure, Quarterly Journal of Eonoms, Vol 06, No 4, 99, pp [9] M Spene, Job Market Sgnalng, Quarterly Journal of Eonoms, Vol 87, No 3, 973, pp [30] J A Brander and P Krugman, A Reproal Dumpng Model of Internatonal Trade, Journal of Internatonal Eonoms, Vol 5, No 3-4, 983, pp