‘Trade Policy and Market Structure Ethanan Helpman and Paul R. Krugman ‘The MIT Press Cambridge, Massachusetts London, EnglandContents Preface xi 1 Introduction 1 IT Trade Policy and Market Power 4 12 Strategic Bets 9 13 Production Efficiency 7 114 The Political Economy ofthe New Theory. 8 115 Organization 9 References 10 2 Trade Policy under Perfect Competition 11 21 Tas Analysis 12 22 Import Quotas 14 23 Export Subsidies and Taxes 15, 24 Optimal Trade Taxes 17 25 ANNote on General Equilibrium 19 26 Domestic Distortions and Trade Policy 20 27" Algebra for Small Policy Changes 22 References 25 {3 Protection and Domestic Market Power 27 31 The Case of an ImportCompeting Monopolist 27 532 Noncooperative Domestic Oligopoly 38 33 Collusive Domestic Oligopoly 41 Reterences 46 4 Foreign Market Power 49 41 Optinl Policies agit a Foregn Monoplst 50 42 Trade Taxes 2 - -Contents Preface x | 1 Introduction 1 | 11 Trade Poley and Market Power 4 | 12 Strategic Effects 5 1.3 Production Eficiency 7 i 14 The Political Economy of the New Theory 8 | 1L5 Organization 9 | References 10 2 Trade Policy under Perfect Competition 11 2. Tarif Analysis 12 2.2 Import Quotas 14 23 Export Subsidies and Taxes. 15 24 Optimal Trade Taxes 17 25 A\Note on General Equilibrium 19 2.6 Domestic Distortions and Trade Policy 20 | 2.7 Algebra for Small Policy Changes 22 References 25 5 Protection and Domestic Market Power 27 31 The Caze ofan Import-Competing Monopoist. 27 53.2 Noncooperative Domestic Oligopoly 38 ‘33 Collusive Domest Oligopoly 41 References 46 4 Foreign Market Power 49 4.1 Optimal Policies against a Foreign Monopolist 50 42 Trade Taxes 53empresas Contents vi Contents 43 Quotas 56 9 Conclusions 181 444° Optimal Trade Taxes 58 9.1 Methodology 182 45 Kinked Demand 60 922 Surprising Efets of Policy 184 46 Cournot Oligopoly 61 9.3 The Welfare Effects of Trade Policy 185, 447 Bertrand Oligopoly 63 9.4 Where Do We Go from Here? 186 448 Ad Valorem Trade Taxes 65 References 188 49 Quotas under Oligopoly 69 4:10" Appendix to Section 497 Index 189 References 80 Strategic Export Policy $9 5.1 Competitive Foreign Conduct $4 522 Profit Shifting 88 53 Price Competition 97 5.4 Entry with Increasing Returns 102 5.5 Resource Constraints 104 156 Two-Way Export Policies 108 5.7 Consumption Effects 112 References 115, 6 Strategic Import Policy 117 {64 Tals: The Cournot Case 117 62 Tariffs: The Bertrand Case 123 63 Import Quotas 125, References 131 7 Intraindustry Trade 133 7:1 A Basic Monopolstic Competition Model 134 72 Tariffs and the Terms of Trade 137 | 753 The Production Efficiency Effect 140 74 Tans That Lower Prices: The Home Market Eifect 145 75 Tas and Segmented Markets 149 References 153, 8 Quantification 155, 811 Problems of Quantifiation 155, 832 Models with Fixed Numbers of Firms 165, {83 Models with Free Entry 169 ‘84 General Equilibrium Modeling 177 5 Conclusions 178 References 1791 Introduction Relatively few markets for industrial products and services meet the tssumptions of perfect competition, Except in markets fora limited shumber of standardized commodities, firms do not usually view themselves as pure price-takere. Indeed in many if not most markets there are only a limited number of impartant competitors, and these competitors see aware ofthe interdependence among thelr ations. ‘What i tre of the economy’ a5 a whole is true of international trade es well. More than hall of world trade i ie manufactured goods, where markets are often oigopolisic rather than competitive. Markets for minerals are also often oigopolste (or olgopsonistic, ‘where the procesing stage is highly concentated). Even in agicul> tural products the pervasive roles of such institutions as marketing boards mean that the relevant players are not always prcetakers, In other words, the study of international trade should be in part a ‘study of international industrial organization. ‘Once sated, this observation seems obvious. Yet until a decade ago the theory of international trade was almost completely domi- ‘ated by models in which perfect competion was assumed to prevail fn all markets. This was true both ofthe postive theory of interna tional trde—the explanation of why tade happens and what effect, it has—and of the analysis of trade policy. With only a few excep tions, the international economists tolbox contained models where all firms took prices as given and set their own prices equal to ‘marginal cost. ‘Recently allthis has changed. International wade theory has gone through 2 quiet but fundamental evolution in which anew approach has supplemented and, in some cases, supplanted the traditional analysis The change began with a new theory of trade in which Increasing retume played coequal ole with comparative advantage(Chapter 1 > In giving rise to trade, Since increasing returns are normally incon: sistent with perfect competition, this approach necessrly modeled ‘markets as imperfectly competitive. Traditionally trade theorists had avoided dealing with imperfect competition because they expected it to prove intractable and barren of useful insights. As work by « number of authors showed, however, this need not be the coer drawing on new developments in industrial organization theosy, it Proved possble to develop models of international trade inthe pres. fence of increasing returns and imperfect competion that stale ‘orthodox models in their elegance and suggestiveness, Furthermore it became clear over time that what had at fist appeared te be proliferation of inconsistent models could be subsumed under s com, ‘mon framework that showed a fundamental commonality of concep, Weccan nov talk with more oles justification ofthe theory of trade under increasing returns and imperfect competion as» coherent body of ideas. Indeed, this theory has now received widespread scceptance and has become par ofthe standaed ew Although the postive theory of trade under imperfect competition ‘has now reached a certain maturity and acceptance, the same cannot. be said of the theory of trade policy under imperfect competition, It {clear that changing one’s view of why trade happens’ and howe lnfernational markets work, ought to change one's view of what Kine of tade policy is eppropriate. To oversimplify slightly, traditional competitive trade models provided support fr free tate a par of the general case forthe optimality of laissez-faire ina competitive conomy. Allow that markets for traded goods and services ore not ‘yplalyperfecy competitive and one opens the door for a variety ‘of anguments for government intervention. That much fs eas, Be beyond this very lite has been clear. As model of trade potcy under imperfect competition have proliferated, the Geld has begun {0 fake on some of theless atractive attbutes of industal organ zation theory, Instead of a convergence around some basic set of Principles, there seems to be ever increasing diversity; cynics have been heard to remark that nowadays a bright graduate stadent con construct model to yield any desired policy conclusion, Parly aso result the policy implications of nev trade theory have ot received anything like the wide acceptance given tothe descriptive se Our purpose in this book isto achieve some progres in this state of affairs by bringing atleast some order to the variety of models {hat have appeared in the analysis of international tae poly in rnrodston ere competitive mates Ina previous book (Helpman and Filgman 156) feeds syethess Othe new pov theory of ade that we Boeve played nfl sole both in siping the ring conene in that area and in communicating the devel Cf dr ude nt Wat ne see sfheve hres the sume thing for the ply sie. Ade, tak shrew The aly nde ly der pro tonpelion canna be punged ofall ambigutyandncertaty crue mary ofthe condicting rents of diferent modes ae rooted Infundamental amg in thundering econome. The wee les of sy,» tan levi aaa a forcign monopoly ‘Spor indy oan oligopoly areal nce Beane iy lg sao mn ae nd wp tha comet be determined «prio. What we cn do, however toate the man eect of tae poy In inperacycompetive Indes where thee fics conde, we wllotleat kw prey ther the sure of ancora es im In presenting this nag we hve chosen foc arly shore runin te aaj ade ply ae psy fm the prevalence of imperectcompettion. This bool na ened tees on tide palcy Thre ta huge at Rghy sophie Morte tat analy the imptons or trade poly of external tconoms of fcr market ditortne of ran (a opposed 1 invidunt power tde, anno on. We Bey review ome ot base impcton of his etre in chapter 2, but oly fra bctop. In ly tae ply fo ny, the aca indy Prob inveves jn about every consideration that has ever been enone, wheter in tadonal rade theory ors newer ina Zon etemaiis, wage dierent anal make owe, non ‘conomic objectives rent seling oat international aly, and Algona. We donot yt get al In but fun on what aay cally new and different—namely, the role of imperfect competi ‘tag es th ak el on pal on sd Ta ati don Tee ou eho equim facts coud sharply ae enacasons we poi ox cut andor onan danensivs theme whe Gir bens seca. Fr thon pt, homered na th ie wre npr 9 mt wnt poe 2 iy of how tds poey works In pacts sere Daneer Chapter 1 : {invoke all the extra machinery needed to embed our analysis in a ‘general equilibrium framework. What, then, are the main themes inthe economics of trade policy under imperfect competition? To some extent the themes are best expressed in terms of models rather than words, bt it say be use to state three basic issues that recur in much of the new tee bi ature, and in our book as wel, These are the effets of wade poy cn market power the stiategic eet of trade ply on competion and the effect of trade policy on consumer chore, 11 Trade Policy and Market Power The defining feature of imperfect competion is that fms do not take prices as given, As a result they do not regard the sole efec of seling another unt of output on ther revere as being the pce of that unit they have some conjechue about what eflet selling wane {or cutting their price in order to sell more) will have onthe revenue they get from inframargnal sales. The end result is that price docs ‘not equal marginal cos, and normally exceeds ‘The ratio of price to marginal costs one measure of market power, A distinctive feature of trade policy under imperect competion something that cannot happen in perfectly competitive merkets is that a trade policy may ater the markup of price over marginal cost, in ways that are either beneficial or harmful fo the county thet Initiates the policy The most familiar argument here is that protection of domestic industries is anticompetitive, allowing domestic firms to Increase their markups at the expense of domestic consumers, Economists have also long argued thatthe extent ofthis anticompetitive eect depends on the form of protection as well sits level Ataf makes imports more costly, ut they can sil increase. A domeste indo that raises its prices too much will still find customers turing fe imported substitutes. An import quota, on the other hand, prevents ‘uch substitution. Thus there i a familar argument that gots ae in some sense more anticompetitive than tails (ee Bhagwat 196), At formal level, however, this argument has been made only for the case ofa pure monopolist confronted by import competition tn chapter 3 we review the argument that protection eeates market Power and that quotas are worse than tafe, and we examine how tection vl the argument holds up when the domestic industry is a ok ‘ater than a mony. eta uncer aten te fle f proton om the markt wer of frign finns sling into the domestic markt. An oe Elnaty popula argument but rf ett they wl be ey Sunred through deine nfregn markups rather than passed Siocon" fore sth al” On he te ad Comore have offen warmed that import quote sly cre feeupn prec, Inducing ther ose rcs and acy ene fang foreigners at domett expense, Chapter 4 tet Doth anes, fskng in parca how the rela depend onthe market structare of foreign indy {A partly intereating but problematic question show trade pelt) afc competion when both foreign and domestic pple Ie imperfecy competitive. On oe sige, such potecton may in ‘fet help domestic an foreign Sms form sxe exploiting emetic consumer. On the ther hand, protection may seve at 4 Stele policy, ofthe kind dacussed below, hat shifts the game tence reign and dometc ms tothe domes ms advantage. Orit may reallocate consumption in a socially desinble way fom the point of view of he deve county (lwo see below). These tute are treated in chapter “Gouly,there are say nse invlvng che fc of tc plcy on market power. One thread tat ts themn together Is that of nai method What we shove hat» sping wide variety of probles involving mht power canbe syed by osing on ocd magne reece increase in seven that a fm expects {orecive by prod one more unit, which aays es than he price (bce ofthe ee! om intrmarginl sles) ut may ceed the tte mapa revere that would preva ifthe induaey acted in concert. As we il se, a spl appratosthat compares the de trad carve withthe perceived and acl marginal evense cares fav appcabty to many topes: 12 Strategie Effects One ofthe most widely noticed and controversial aspects ofthe new literature on tode policy under imperfect competition is the poss Dilty that interventionist trade policies may have benefical “stra- tegi” effets. A strategic move a tis defined in modern industrialChapter 1 organization theory, isan action that iS not profitable viewed in Tolation but that alters the terms of subsequent competition to a firm's bent, For example, afr may invest in exces capacity that it doesnot intend to se, but whose presence deters potential com pettors from entering the market. An implication of the new trade theory is that government trade policies may serve the same kind of role ‘The now-famous example i the so-called Boeing-Airbus case (any resemblance tothe real firms with those names may be purely co. lncdenta). Suppose that there isan aircraft that either Boeing or Airbus could produce profitably but that—owing to the fied costs of developing the plane—if both enter the market, both will lose ‘money. Then there isa game of mutual deterrence between the firms in which each tries to convince the other that itis committed to produce; if it succeeds in making the commitment credible, firm will deter its rival from entering and reap the profits. Firms may, however, lack any credible way to make a commitment to produce | this situation government policy can make the diference. I Eu- ropean governments make a promise to pay Alsbus a subsidy for production that is large enough to induce entry even if Boeing enters a8 well, then Boeing wil find entry unprofitable and leave the field to Airbus. The result wil be profits to Airbus that include not nly the subsidy but also the profits from sole possesion of the world ‘market. Ignoring consumer costs, this will raise European national income at American expense (see Dit and Kyle 1985), Since it was first enunciated by Spencer and Brander (1983), the strategic trade policy argument has received wide attention. ‘The Bocing-Airbus story ems at first sight to provide exactly the kind ‘of argument for aggressive national trade policies that protctionists have always wanted, and it has been picked up in some variant in semipopular discussion. Yet from the start economists have been ‘cautious. Does the strategic argument realy offer « presumption in favor of aggressive policies or is this just a special case? Ate there likely to be offsetting effets in the kind of industry to which the strategic trade argument might apply? How likely iit theta govern ‘ment will be able to have the information necessary to conduct @ suocessful strategic policy? Inchapter5 we ofer a synthesis of the strategic trade policy debate ‘it has evolved over time, Interestingly, itis both possible and tnrodution useful 0 apply to this debate the analytical method of contrasting ved and actual marginal revenue; this helps us to offer afaiy Eompoct and straightiorwacd review of the main arguments. The bse result reinforce the skepticism about the general applic bility of Bosing- Airbus type examples: on balance, the ease for ag> gressive trade policies is relatively weak "As we have noted, strategic issues also arise in the context of protection when both domestic and foreign firms have market power ‘Thus chapter 6, which studies sch two-sided situations, aso touches con strategic effets. 113 Production Efficiency “There is a quotation often attsbuted (we hope wrongly) to Abraham Lincoln, tothe effect that when an American purchases a coat fom 4 foreigner, the American gets the coat but the foreigner gets the Toney, when the cost is purchased from an American, the buyer stil gets the cost but an American gets the money. The usual re- sponse of economists is that there is no such thing 88 ree coat: to ‘make an all-American coat requires using resources that might more productively have been used to make something else (say, Boeing). “Once one recognizes the importance of imperfect competition, the ‘usual economists response loses something of is force. Making a cat is costly to be sure; but if its price exceeds marginal cost, then the resources sed to make it might not have an equally productive use elsewhere. In principle at leat, polices that induce consumers to buy domestic oods whose price exceeds marginal cost may raise ational income. In this book we emphasize this theme in the contest of inra- Industry trade, where countries are exchanging goods in which net ther has a comparative advantage. But we teat i also in other eases ‘We show thatthe argument has some validity, Perhaps even more surprising, protection can under some circumstances Induce an in- ‘crease in domestic production that actually lowers prices to consum cers. However, the analysis also makes clear that this is not fundamentally an argument about trade policy, but rather one about how to handle the general problem of pricing under imperfect com- petition, When we turn to quantitative analysis of trade policy inChapter . ‘chapter 8, tis notable that as more policy instruments are added the ‘optimal policy typically looks less and les ike a trade policy. 414 The Political Economy of the New Theory ‘This is a book about theory and methods, and not abot policy. Yet ‘one must realize that in economics theory will be used to advocate policy, whatever the theoris’s intention. Stateic trade policy ar- ‘guments have already appeared in support of views none of the oncept’s originators hold. $0 we need fo say something here about the politcal economy of the theory we discus. Pethaps the most important poin to make i that arguments based ‘on imperfect competition are not the only or even pethaps the most ‘powerful professionally respectable arguments against free trade “Arguments based, in particular, on external economies and factor ‘market distortions are oflang standing and have substantial empirical support. Thus the new theory should not be viewed as somehow shatteringa monolithic view in which free trade was the only possible Position (se, for example, Bhagwati 1971; Corden 1974) ‘Admittedly, there is a philosophical diference between the new arguments and the ol. In traditional trade policy analysis, distor tions that could justify government intervention were in effect si perimposed on a theoretical structure whoee basic loge was that of ‘ficient, competitive equiliium. In the new theory the imperfox tions are bul nto the structure from the beginning. So the arg ments for interventionist policies are deeper in some logical sense Yet in practical terms its not clea how much diffrence ths makes, (One can imagine a competitive market without extemal economies ‘hough one eannot imagine an oligopoly where price equals marginal cst. But the odds may be that extemal economies are a more int Portant distortion of real-world economies than monopoly power. ‘Meanwhile there are a number of reasons for treating trade policy ‘conclusions cautiously. One is uncertainty: a6 we show, the effects ‘ofa given policy may depend erucally onthe details ofthe market ‘Another is domestic poitial economy’ there are many people eager to appropriate the results of new trade theories to support highly dubious political uses, and international economists do not want tobe apologists for exude protectionism. Finally, there is the problem of international rivalry: @ policy that benefits one country acting tntadstion naterally may be harmful if everyone does it Since quantitative nalyses sem to suggest thatthe gains fom even optimal interven- tion are smal (se chapter 8), many economists have suggested that free trade remains a usefl rule of thumb, even though it is rarely optimal in modern trade models. 115 Organization We develop the arguments gradually beginning with market steuc- tures that exhibit one-sided market power and proceeding to market sructures that exhibit tosided market power. After providing back- {round on rade policy in competitive enviconments in chapter 2, in ‘chapter 3 we discuss import protection by a country with a domestic ‘monopoly or oligopoly. There we assume that foreign exporting firms ate competitive, thereby isolating the role of domestic market power, Inchapter 4 we deal with import protection ofan economy that faces «foreign monopolist ora foreign oligopoly, while domestic supply Js noneuistent or competitive, This market structure enables us t0 faolate the pure effects of foreign market power. in al of these chap- ters there are no strategic interactions. ‘We bringin strategic interactions in chapter 5 there we deal with export markets in which domestic fms with monopoly power com- pete with foreign firms with monopoly power. Our inital working hypothesis is that exportables are not used domestically. This way swe need not worry about consumption effects, and we can isolate the pure effects of strategic interactions. At the end of the chapter sve dacuse domestic consumption. In chapter 6 we explore the role of strategic interactions in a domestic market in which domestic firms with market power compete with foreign firms that possess market power This structure contains many elements of previous market Structures as well as new ones. We concentrate on the later (Chapter 7 is devoted to trade policy in the presence of two-way tmade. This may arise from monopolistic competition in differentiated products or for strategic reasons, In ether case there are new features {hat bear on the effectiveness of trade policy. Finally, in chapter 8 we review the recent iterature that tries 10 quantify the effects of tade poly in noncompetitive environments, We discuss the new meth- ‘dlogy aswell as the numerical resus. In chapter 9 we Sum up ‘our findings and take stock of what has been learned.Chapeet References ‘Bhogwati Jagdish N. (1965). “On the equivalence of ais and quotas tn Rober E, Baldwin etal. (s), Td. Growth snd th Bloc of Payments says Honor of Gli Hater. Chicago: Rand McNay. ‘Bhagwat, Jagdish N. (1971), “The generalized theory of distros and ‘wefan Jagdish N. Bhagwat et eds), Trade, Bale o Payments, end Grout: Papers Iteration! Heme Hono of Cars P. Kinder ‘Amtech North old, orden Max W970 Trae yan mo Hire, Oxford Credo Din, Avinash, and Kye, A, S (188). “The wseof potion and subsides forestry promotion and deterence.” Anson Enon Rena 75 199-182, epman, Ehanan and Krugman, Paul. (195). Merk Strate ond Foreign Tre Cambridge, MA: MIE Pres Spence, Barbora, and Brander, James (198), “Inteational RD rvlry and industrial statey.” Review of oom Sadi 50 700-22. ‘rade Policy under Perfect 2 Competition [Although this book is conceened with the theory of trade policy in Smperfeclly competitive markets, i is useful to begin with a beet exposition of some main results from the theory of rade policy under fonditions of perfect competition. For one thing, this traditional the- fr important as a backdrop, t help us see what is new about the resus presented later (and equally important, what isnot). Also a presentation of waditional theory can serve as a convenient place to Introduce some concepts and techniques that will recur in our later analysis. ‘The version of traditional theory we present inthis chapter is of ‘course highly ebbreviatedthere isan enormous literature on trade policy under perfect competition, too large to do justice in a single ‘chapter. In particular, we wil focus primal ona partial equilibrium ‘model of a single market, thereby neglecting important general equi> lirium issues. Also this chapter sin no sense a lterature survey— too many people have contributed to the vast edifice that traditional trade theory even to start giving credit where credits due, We note simply that a more extensive survey in the same split as that ofthis ‘haptcr may be found in Corden (1974), that a modem textbook teealment tha follows a similar approsch may be found in Krugman and Obsteld (198), and that some ofthe more subtle issues in trade policy analysis are treated in Dist and Norman (1980) and Woodland as. Tn this chapter, then, we briefly review the analysis of trade policies in a competitive market. We bogin by considering three basic wade Policies: tariffs, import quotas, and export subsidiestaxes. We then tum tothe calculation of optimal polices, with a brit digression into ‘general equilibrium considerations, and fllow tis with a discussion ‘ofthe case for government intervention in competitive markets, Wechapter 2 2 ‘conclude with some welfare algebra that will find repeated use in later chapters 24 Tariff Analysis Figure 21 strats the determination of trae vlune and pres inasinge competitive market. The quay mrereents the ports ffs county ta we wil ll Home, and ao the expanse oust that we wl cll Forign. Te curve D represents Home's import Amand cre shows the exes of dameste dear ver Spy teach prc. The carve $tpresents Fore’ eet Spl oe, ‘inet the exes of omen uppy over demand lech pce Inthe absence of nyt or angporation cet elm wl est pat heehee cos Ate pcp Home thes demand im and sos Prelsexes supply us for the Word ana whole excess demand vata Now conser inposing a ri. aa perfec compete market it doesnot mater whether the tail i pei an per py tt or ana alr tif evi asa poporton of he po we ‘Simpy by asuming aspce taf tet The elt of Such tax must be fo diva wedge beeen the pio the good info ann Ho ne 21 i shown ing the intra rc in Home op whe loweng the pc in Fre {opts The diference p= fms fet equal the afte gu 21 ‘iad Ply under Peet Competition Bb Clearly, the tariff makes producers and the government in Home etter ff and Home consumers worse of. To say more than this it TE necessary t0 make some further assumptions, The standard as- fumptions are the flowing. Fist, we assume that income distribu ‘on snot an issue: a marginal dollar of government revenae has the fave socal value a8 a marginal dollar of producer or consumer furplus Second, we assume that the economy is efficient in all ‘haties other than this one, 20 that consumer surplus and producer Surplus are accurate measures of wellae. The firs assumption may be defended either by the asertion thatthe government manages foome distribution using some other tools or simply as a working fssumption that puts dstbutional sues on one side. The second Sssumption wil be relaxed in section 26 ‘Given these assumptions, the welfare effects of the aif can be measured by the effects on consumer plus producer surplus together ‘wth government revenue. In Home, the combined effect on con- ‘mer and producer surplus may be represented by the reduction in the area under the import demand curve; that s, by the area 12345 in figure 21. Meanwtile, the tariff generates revenue equal to the tarif rate multiplied by the posta volume of imports that is, the tea 2067, "The overall welfare effet i the diference between these two. [Neting out the area of overlap, we have a loss represented by the pa the price of competing Imports wil cease tobe binding on the firm. The im i fe to charge its profit maximizing monopoly price, and changes inthe tanif rate ‘within this range will have no effect ‘This three-part analysis may be usefully summarized by figure 3.2, which shoves how the domestic fim’s output varies with the tare inclusive price of competing imports. For low tariffs the frm matches ‘anf increases with is ovin pic increases and sides up its marginal cost curve. For tarif increases beyond the prohibitive level, the do- rmestc price continues t0 rise one for one, but the firm now slides bck along the domestic demand curve, Finally, at sufcenty high tans domestic price and oupat remain unchanged atthe closed ‘economy monopoly level Effects ofan Import Quota Figure 33 shows the effects of a quota that limits imports to some ‘maximum quantity I the domestic firm charges a price below pw, this limitation will not be binding, since there will be no incentive to import For any price above pu, however, m units willbe imported, and the domestic frm will satisfy the residual demand. The result ‘will therefore be to present the domestic fim with a new demand gure 3.2 Prolecion and Domes Market Power a JEN LOE curve D’ that has three segment (epresented inthe figure by the Heavy fines): for prices above pu, itis the domestic market demand shied let bythe quota amount for prices below px este market demand: and there isa horizontal segment at Px Corresponding the new demand curve is anew marginal revenue by the broken Une, also with the three atthe horizontal segment marginal revenue 1d the demand curve coincide) In situations where the country ‘would have imported the good under fee trade, the rightmost scg- ‘ment of MR’ fs never relevant, $0 we drop i in Iter diagrams. ‘nthe situation illustrated in figure 3.3, the quota is seta a Hevel smaller than the tree ade level of mpors. The domes fir mak Imizes pros by choosing the level of output where marginal cost equals this new marginal revenue, shown in the figure as Xa with the coresponding price fo. ‘An impart quota may, however, still have an effect even when i fn st larger than the fee trade import lve. Tis case i ilsrated in figure 34 tis immediately apparent that in this ease the marginal cst curve crosses the marginal revenie curve in two places—that i,Chapters 2 ° there are two lolly proft-maximicing levels of output. One ofthese, at pont 1 isthe ree trade output the other, at potnt 2. i position In which the monopolist takes advantage ofthe quota To reduce ‘output and aise the price above its fre trade lve. To detemmine which ofthese i the gata! profit maximum, consider ihe ght prin ofl reg Ot Fos pn | point 2. At rst, marginal revenue excueds margin! cost ‘Thus it reducing output othe evel Xe whch heats an the frm wil reduce profits by an amount that is measured by the shaded tangle witha vertex at point 1 As output is further reduced however, marginal revenue seb than marginal cost and 90 he fal in outpst tothe monopoly eve produces arise in profs, measured by the area of the shaded triangle with a vertex at point 2. The decision on whether to produce a the fe tae level or retreat Inside Wie quota therefore depends on which of these Tangle Se See eae What i immediately clear is that when the quota exceeds the fee trad import evel the more restcive iis the dose itis othe ree Protection and Domestic Matet Power a trade level of imports, the more likly iis that the fim wil choose to reduce its output. In particular, a quota tha limite impor to precisely their fee trade level (which would not be binding under perfect competition) wil definitely lead to a reduction in output and 2 se i the price ‘There is no correspondence inthe case ofan import quota to the case of a tai that raises price without creating monopoly powes Any binding quota will have the eet of leading the frm fo behave ina monopolist fashion, Compating Quotas and Tatts We can now turn to the most famous result of this analysis that a {ae and an import quota that lead tothe stme level of imports have tering effects, with the quota leading to a higher domes pice and a lower domestic output than the "equivalent taf" This may be seen immediately for the comparison of prohibitive ‘acs with a total ban on imports. As we have seen eatiey there is 4 range of tailfs for which imports are zero, but for which the domestic price i below pu and the domestic vvtput larger than Yor (On the other hand, a quantitative restriction that eliminates mpots limmediately leave the domestic frm fee to move tothe mano ‘ly price and output. The point is of eouree that in the ange where uw + #< pus the threat of imports constrains the domestic pce ven though there are no imports infact an import quota eliminates trea, cresting more market power and thus a highes domestic awe. ‘The same is true when positive levels of imports are allowed, as igure 3.5 shows. The technique used sto construc an import quote that Limits imports to the same level as an arbiteay fanth, We tose find the level of imports corresponding to some nonprohibitive tan and then compare its effects with a quota set atthe name level of imports. The postquota demand curve D’ facing the fra passes through the tri production)price point 1, Is coresponding mer, ‘inal revenue curve therefore lies below, leading to quota pliChapter 3 s gue 93. ‘ium witha lower output Xo and higher price po. This stems from the fact that the tarif-constrained demand curves more elsticthen the quota-consrained demand curve for output X A special case ofthis results where the fai to which we ind ota equivalent zeo—that is, where imports are rested only ‘ot to exceed ther fee trade level. Clealy, the reult is aie inthe rice and a fallin the output of the protected firm This example also shows thatthe effect ofan import quota on domestic output may be perverse. Here limiting imports lad toa fallin domestic output, instead ofthe vise that would wccur under Perfect competion. The reason is of course that the import quota reales monopoly power that is used o restrict output and raise the rice The moral of this story i therfore lear an import quota cre tore domestic monopoly power than a laff that resis inane by the same amount and-it therefore Id Batt lower domestic uipat and a higher domeste price. Although Bath the aif and the {quota reduce welfare inthis model, the quota obviously reduces it How robust are these conclusions? We fist consider some exten: slons ofthe case ofa domestic monopolist, then tum o more complex market structures, es © pnction an Domestic Market Power Decining Marginal Costs ers Te pnt fair solu, However, he case of domestic mo 2 yah an be poe od cy ee ee a eee ae eee 2 Selec reeteniet uptime eae eee re Sores Seen ee ee ae oe ee ogee “distension See eee Bee iV em cold ep Wag bale to eng poiy we sume itCaper 3 a Aispaces all imports (see Corden 1974, ch. 8). Any further increase in the tariff above this “sientic” level will allow the frm to increas ‘price and reduce its output unt the monopoly positions reachea With downvard-soping marginal cost then, the upward sloping section ofthe curve in figure 3:2 is eliminated The nonequivaence of tarifs and quotas cesrly remains. As in the ‘se of upward-sloping marginal cost, there ia tange of prohibiting taf that raise prices less and lead to higher output than a simple Drohibition of imports. For les restrictive import quotas, there we ‘no equivalent taf, since taf either eliminete imports or fal te establish the domestic industry. Nevertheless, the basic bout the role of protection in creating markt power andthe gear monopolizing effect of quotas seems to come through essentially inact Impericet Substitutes ‘We now consider the case of a domestic monopolist tht faces com petiively supplied imports that are imperfect substitutes for ite own Production, When imports are not a perfect substitute for domestic Production, the single domestic firm has some monopaly power even lnder free tade; thus protection can no longer have the effect of treating monopoly power when none existed before, However the ‘monopoly power ofthe firm may sil be increased by protecion, We ‘want fo se whether it continues tobe tue that quotas aze protective (oa greater extent than tan, Consider, then, a fin that isthe only producer of a good but faces ‘competition from imports that are impertect substitutes for is prod ‘cts. The firm will always have some monopoly power: even under free trade it wil face a downward-sloping demand curve. We denote by PQS pits demand function, where X represents output and p represents the domestic consumer price of imports, that, inclusive of tariffs or quota rents, ‘A tariff (here again the tariff can be specific oF ad vaocem) will hhave one definite efect: by raising the price of competing imports from p= pf to p= pf + f, it will increase the demand for the domestic firm's output at any given price p. However, the domestic fim might take advantage of this increase in demand by holding te Dice constant and allowing its sales t rs; in this case we woud ‘ot register any increase in monopoly power as measured by the "potcton and Domestic Markt Power y “ato price to marginal cost (or marginal even) 36 lang as mat conare cont Or fray ie both sls and pres then we Set ge: nce in menopely power Wich care apple Stn the shit othe maria vena cree tas may, ‘cmpaion wth got to mga he nore Bidet that quote Genes oe memoply poe thas a ‘SNe tn we ned Wo thnk about how a quota wrt nthe inp setts case The higher he price ofthe domes, (Rego wilt he demand fo temps tat subtte s wee ft pot iy ing ta ane ines, Ind het pre mst ners the ust eer eat se oder temper ave Tb eae tt wh te “ct fm spo by an por quot know at when frac to pte the pes of tape done cones ferne a wel fet that tl enact ning ts nd ce “oem 3.7 shows what this implies for the comparison of tariffs. snd uta Te ce Brepeete the demand feng the SomesCaper 3 = firm when protected by some particular tai rte, MRe isthe come sponding marginal revenue curve, and hence X; i the proftmasi mizing output and p the resulting price. Now suppose thatthe tani were replaced by a quota limiting imports to their level under the laf Ifthe price charged by the domestic firm were unchanged, the quota rent would equal the previous tariff, and there would be no change in the demand for domestic output! Thus the demand curve Dg corresponding to the quota passes through the orginal price! production point 1. However, the new demand curve is seq The ‘reason ip that a rise in the price ofthe domestic good will pl the Price of competing imports to consumers up with f, dampening the ‘egative eect on sles” Because the demand curve with « quota is stocper, the comesponding, marginal revenue curve MRe lies below (MR, The results that the output under quota protection Xo slower, and the price jo higher, than under the tariff that yields the same level of imports. The imperfect substitutes case, then, appears to confirm broadly the results ofthe perfect substitutes ease 3.2 Noncooperative Domestic Oligopoly ‘The case of a single domestic firm whose only competition comes {om imports is, as we have just seen, one whose simplicity allows ts to derive strong results with a minizum of technical dificalty, However, as @ guide to real-world policy the monopoly analysis presents probloms, since what one typically finds in practice is 8 ‘domestic oligopoly: a small numberof frms, but nt one, that receive protection inthe form of tariffs and quotas. The question is therefore ‘whether insights from studying the monopoly case hold up when there are several domestic firms. “There unfortanately no general theory of oligopoly behavior. We therefore need to make some additonal assumptions about how firms behave in order to complete our model. In ou Hist pass we ‘assume that firms behave noncaopeatcly, that is that they take one another's actions as given. We discuss tacit collusion in the next 5am he dc dsl ripen gre by he vere demand untn IMO, fhe pe of mete output unde the tl eas Xp ond Inpstly deed by p= Xp. ll aner the unt, Hence ap othe ‘sn domand En gp andthe sop the qt cond ded ‘Si unl ptt~ pp The ere ger (slate a) nda th xonae ‘ssunpton pin = prtction and Domest Matt Power 2 Once we have decided t9 model the behavior of firms as being oncooperstive, there romains the question of what itis tat firms ‘ake as given-the fue of static trial. Does a frm assume that {tscompetitors wil choose a price and adjust output to meet demand that pice, or that they wil choose levels of output and adjust their = 30 a8 to be abl to sell ll their outpat, or that they set some Tore complex trade-off between price and output? The two simple ‘remes of price-setting and quantity-seting ae the famous cases ‘of Bertrand and Cournot competition. As it turns ou, che qualitative ‘esults are pretty much the same, The analysis in the Cournot case fs somewhat simpler, however, and so we focus on that eas. ‘We suppose that there are » domestic firms producing the same good that i, however, an imperfect substitute for an imparted good, ‘As before the demand function fr the domestic good may be waiten SS H&, pi. We assume thatthe domestic firms all have the same, fonstant marginal cost c this implies that any equilibrium wall be Synmcricthat is, each domestic firm will produce a quantity x = Min. This assumption of symmetry is a very useful simplification, allowing us to draw a direct parallel between equiibsium under ‘ligopoly and monopoly ‘rece trade equim is illustrated in figure 3.8, which looks a rest deal lke gure 37. There is a dovenward-sloping demand curve curtesponding to an import price equal tthe world price pt, and culitrum is at point 1, where marginal revenue equals marginal However the marginal revenue that i equated to marginal cost is sot the marginal revenue that corresponds to the demand curve Instead, itis the marginal revenue pereved by firms, defined by x, wy = pew + (8) ms po lnreslts rom the fact that every fim calculates its marginal reverie, taking as given not only the price of foreign imports but also the output level ofits domestic rivals. Namely its marginal revenue equals the change in p(X" + xm) when x changes by one unit and X'—the output of domestic rvale—does not change. In a ymmet ‘eal equiitrium x = Xin, and perceived marginal revenue i638 de- Sexbed above. ‘This perceived marginal revenue curve is a weighted average of the trae marginal revenue curve and the demand curve:chapters 0 Figure snowy = (2) see + (1-2) poe, here MCX, p) ~ pC, pi) + XpuX, p)- Bauilibrim obtains where perceived marginal revenue crosses marginal cost, at point 1 Bifets of 8 Quota [Now suppose that tariffs replaced with an import quota that limits Jmport othe same level, Then just asin the case of single domestic firm, each of the competing. domestic firms will now believe that ‘when it restricts its output, the attempts of consumers to shit thelr demand to imports willbe frustrated by arise in the quota rent. Each firm stil takes the output of ather firme as given, so tha the demand curve it perceives stl takes into account the loss of sales to other ‘rms, but the elimination ofthe substitution into imports makes this perceived demand curve steeper than before. Thus percived mar ‘inal revenue falls, and firme are indice to reduce thee output and protection ad Domestic Market Power a increase ther prices. A quota then, creates more monopoly power for the domestic Indust than an equivalent tart. ‘Our examination of the noncooperative oligopoly case seems deatly to confiem the intuitions from the monopoly case. Tariffs increase monopoly power, quotas increase them even more. Our reat step is To conser the more difficult question of what happens When the domestic industry may act collusively rather’ than roncooperatively 433 Collusive Domestic Oligopoly Ie firs were to collude perfectly the study of oligopoly would be no different from that of monopoly. What gives collusion its interest land also its dfcuty isthe likelihood that firms cannot monitor and ‘enforce agreements perfectly, and thatthe extent to which they can ‘exploit theie potential joint monopoly power depends on the envi= ronment—including the trade policy environment. In reality the abi iy of firms to collude no doubt depends on a variety of factors, ‘including anitust law, information issues, and even the social cohe- son of the industry's executives. In recent work i industrial orga- ration, however, intresting insights have been gained through an {approach that strips the problem of collusion down to only one issue ‘whether the expected costs resulting from a breakdown ofa cartel oF tact collision are sufficient to deter cheating, We assume that even {in cases of explicit collusion the agreements have tobe seltenforcing: otherwise, we have a monopolyike situation. ‘Thebasic ideas that at any point in time a frm can always increase fs profits by breaking cartel discipline and undercuting is vals! price. However, it does so, the cartel is lkely to collapse, reducing, the subsequent profits ofall participants. If the cartel going to be sustainable, the costs to each member frm of its collapse must out- ‘weigh the gains from cheating. Under some circumstances the need. for sustainability wll constrain the ability of the cartel to raise prices above their noncooperative level: if too high a price makes cheating attractive, the cartel must keep the price just below the level that provides crical level of temptation (ee Tirole 1988, ch. 6) ‘When this ides is applied to trade policy, ae was done by Rotem berg and Seloner (198), a Seemingly paradoxical result emerges: a ‘quota may raise prices less than a tart, precisely because raises Profits in the absence of a care. If profit: willbe higher in the