POLICY RESEARCH WORKING PAPER 18 Regionail Integration wrJi i o. >.Whhr0oa trade , -'' as Diplomacy agrfteeme h~elp redUce securt tensions between neighbors, Maurice Schiff L. Alan Winters The World Bank Development Research Group August 1997 PPOLICY RESEARCH WORKING PAPER 1801 Summary findings Regional integration agreements (RIAs) are ex imples of * The optimnumi tariffs on imports from the rest of the second best and have an ambiguot]s impact on welfare, world are likely to decline over time. contend Schiff and Winters. * Deep integration implies lower optimum external They build a model in which RIAs unambiguously raise tariffs if it is exogenou]s. welfare by correcting for externalities. It assumes that * But if deep integration is endogenous, it implies trade between neighboring countries increases trust higher optimum external tariffs before it occurs and between them and reduces the likelihood of conflict. lower ones thereafter. The optimum intervention in that case is a subsidy on Fnlargement of a bloc (in terms of symmetric imports from the neighbor. The authors show that an countries) has an ambiguous impact on external tariffs equivalent solution is for the neighboring countries io hutt improves welfare, and some form of domino effect tax imports from the rest of the world -- that is, to form exists in the sense that enlargement increases the an RIA - together with imposing some domeitic taxes. incenitive for nonniembers to seek accession. In fact, security threats have moved neighbcring Although externalities associated with security matters countries to form RIAs. Examples include the creation of imply that an RIA may maximize welfare, this model the European Coal and Steel Community (1991) and the suggests that the RIA is a transitory arrangement in the European Economic Community (1957) to reduce- the sense that optimumtin trade preferences are highest at the threat of war in Europe, as well as various RIAs among time the RIA is formed (when securitv is low) and tend developing countries. to decline ovcr time. In other words, the RIA's external Schiff and Winters show, among other things, that: trade policy becomes increasingly open over time (as well as following deep integration). This paper - a product of the Development Research Group - is part of a larger research program on regionalism and development (directed by the authors). Copies of the paper are availahle free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Jennifer Ngaine, room N5-056, telephone 202-473-7947, fax 202-522-1159, Internet address jngaine@'worldbank.org. August 1997. (37 pages) The Policy Research Working Paper Series dissetninates the findinigs of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than full), polished. The papers carry the names of the authors and shoul I be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do niot necessarily represent the iview of the World Bank, its Executive Drrectors, or the countries they represent. Produced by the Policy Research Dissemination Center Regional Integration as Diplomacy* Maurice Schiff and L. Al-an Winters We rld Bank *We would like to thank partcipants at seminaLrs at the World Bank and at the 1997 AEA meetings for useful conmments. ABSTRACT Regional integration agreements (RIAs) are examples of seccnc uesi and hiave an ambiguous impact on welfare. This paper builds a model where RiAs unambigjously raise welfare by correcting for externalities. It assumes that trade between neighboring countries raises trust between them and reduice the likelihood of conflict. Thie eptimum intervention in that case is a subsidy on imports from the neighbor. 'The psDer Sh'oWS that an equivalent solution is for the neighboring countries to tax imports fronm the rest of the world, i.e., to form a RIA, together with some domestic taxes. In fact. security threats have moved neighlboring countries to form RJAs. Examples include the creation of the European Coal and Steel Community, ('B7CSC, i7151) and the European Economic Community (EEC, 1957) to reduce the threat o0 wvar in Europe, as well as various RIAs among developing countries. The paper shows: i) that the optimu m tariffs on imports from the rest Gf tile wiorLd are likely to decline over time; ii) deep integration implies lower optimum3 exterrnal tariffs if deep integration is exogenous; iii) on the cther hand, deep integration implies higher optimum external tariffs before deep integration and lower ones thereaater if Ireep integration is endogenous; iv) enlargement of bloc size (in terms of symmetric ountries, has an ambiguous impact on external tariff; but raises welfare, and some form or QOJ2ifO effect exists in the sense that, starting from a two-member RIA, both the ouTtside country and the RIA members want to enlarge the RIA. Foreword As regional tading arrangents (RTAs) have spread, enlarged and deepened over the last decade, they have posed chalenges to economists on both intellecftal and policy levels. On the former, do RTAs stimulate growth and investment, fiicilitate technology transfer, shift compaative advantage towards high value-added activities, provide credibility to reform programs, or induce political stability and cooperation? Or do they, on the other hand, divert trade in inefficient directions and undermine the multilatal tading system? The answer is probably "all of these things, in diffent proportions according to the particular circmstances of each RTA." This then poses the policy challenge of how best to manage RTAs in order to get the best balance of benefits and costs. For example, should technical standards be harmonized and, if so, how; do direct or indirect taxes need to be equalized; how should RTAs manage their interational trade policies in an outward-looking fashion? Addressing these issues is one important focus of the research program of the International Trade Division of the World Bank. It has produced a number of methodological innovations in the traditional area of trade effects of RTAs and tackled four new areas of research: the dynanics of regionalism (e.g., convergence, growth, investment, industrial location and migration), deep integration (standards, tax harmonization), regionalism and the rest of the world (including its effects on the multilateral trading system), and certain political economy dimensions of regionalism (e.g., credibility and the use of RTAs as tools of diplomacy). In addition to thematic work, the program includes a number of studies of specific regional arrangements, conducted in collaboration with the Regional Vice Presidencies of the Bank. Several EU-Mediterranean Association Agreements have been studied and ajoint program with the staff of the Latin American and Caribbean Region entitled "MakNg the Most of Mercosur" is under way. Future work is planned on African and Asian regional integration schemes. Regionalism and Development findings have been and will, in fure, be released in a number of outets. Recent World Bank Policy Research Worldng Papers conceming these issues include: Glenn Harrison, Tom Rutherford and David Tarr, "Economic Implications for Turkey of a Customs Union with the European Union," (WPS 1599, May 1996). Maurice Schiff, "Small is Beautifll, Preferential Trade Agreements and the Impact of Country Size, Market Share, Efficiency and Trade Policy," (WPS 1668, October 1996). L. Alan Winters, "Regionalism versus Multilateralism," (WPS 1687, November 1996). Magnus Blomstrdm and Ari Kokko, "How Foreign Investment Affects Host Counties" (WPS1745, March 1997) ii Magns Blomstrdm and An Kokko, "Regional Integrtion and Fori Direct Investment: A Concptual Frmework and Three Cases" (WPS1750, April 1997) Eric Bond, "Using Tariff Indices to Evaluate Preferential Trading Arngements: An Application to Chile (WPS 1751, April 1997) Pier Carlo Padoan, "Technology Accuznulation and Diffuson: Is There a Regional Dimension?' (WPS1781, June 1997) Won Chang and L. Alan Winters, "Regional Integrtion and the Prices of Imports: An Empirical Invesigation" (WPS 1782, June 1997) "Glenn Harison, Thomas Rutherford and David Tarr, "Trade Policy Options for Chile: A Quantitative Evaluation" (WPS1783, June 1997) Anthony Venables and Diego Puga, "'Trding Arrangemens and Indusirial Developmen (forthcorming) Plamned future issues in this series include: Sherry Stephenson, "Stndardst Conformity Assessnents and Developing Countries" Valeria De Bonis, "Regional Integration and Factor Income Taxation" and "Regional Integration and Commodity Tax Harnonization Other papers on regionalism produced by ECIT include: Ahmed Galal and Bemard Hodeman (eds), Regional Parters in Global Maret: Limits and Possibilities of the Euro-Med Initiative. CEPR 1997. Berad Hoeluan and Simeon Djankov, "hIports of Inputs, Foreign Investment and Reorientation of East European Trade," World Bank Economic Review (fortmming) Bernard Hoekman and Simeon Djankov, "The EU's Medit Free Trade Initiative," World Economy Benard Hoekman and Simeon Djankov, 'Bffective Proectdon in Jordan and Egypt in the Transition to Free Trade with Eurpe," World Developmnent. Bartlomiej Kaminsi, "Establishing Economic Foundations for a Viable State of Bosnia and Hercegovina: Issues and Policies". iii In addition, Making the Most of Mercosur issued the following papers: Alexander J. Yeats, "Does Mercosur's Trade Perfrnnance Raise Concerns About the Effects of Regional Trade Arrangements?' (WPS 1729, Febnry 1997)) Azita Amiadi and L. Alan Winters, "Transport Costs and 'Natual' Ilngation in Macosur" (WPS 1742, March 1997) Claudio Fischtak, Danny M. Leipziger and John F. Nomand, "Industial Policy in Mercosur: Issues and Lessons" Sam Laird (WTO), "Mercosur Trade Policy: Towards Greater Integrtion" Maret Miller and Jerry Caprio, "Empirical Evidence on the Role of Credit for SME Exports in Mercosure Malcom Rowat, "Competition Policy within Mercosue, For copies of these papers or ination about these programs contact Mauice Schiff, The World Bank, 1S18 H Steet NW, Washington, D.C. 20433. An additional major outlet for World Bank-sponsored research on regionalismn will be the Anual Bank Confrence on Development in Latin America, 1997, Montevideo, June 30-July 2, 1997, orgnized by the Office of the Chief Economist and the Technical Deartment for Latin America and the Caibbean Region, with the support of the International Trade Division and the Economic Development Institute. Masood Abmed Director Intemational Economics Department [g,iRj: | | t t lI | * i FA Sr|| 3Z1lw~ ; f XX]6ig1 X,,,IC '1l, ',I' p-ically. And even if they do occur, they need not result in welfare gains. For instance, increased FDI may result in immiserization if it leads to the expansion of protected sectors. One case in which RIAs may in theory generate unambiguous welfare pins is if they correct eramities For istance, it is frequendy claimed that a developing country which is in the procas of refrming its trade ,>- other poUCili can benfit from a RIA with a large, developed country or region (e.g., the US or EU) because this binds its reform in an intrnational treaty, weakens the groups who stand to lose fom and oppose the refoms, and raises credibility m their sustaiabily.3 And even though the standard static welfae impct of such a RIA may be negative for the reforming country, the latter is likely to gain once the benefits of the enhanced credibility of the refms are taken into account. A second ldnd of extmality arises if a RIA genates improved security for its member countres. There are basically three types of situations where RIAs may generate such positive extlrnalities. First, there may bdomestic security tbreats such as civil disption or civil war. For instance the Egyptian government has been conced with the spead of fndimentalismand those of Morocco and Tunisia have been concerned with the possibility of contagion from fundamentalism in neighboring Algeia, with the associated risk of civil strife. These issues, which have also been of concern to the EU, have provided one of the motivations for agreements betwe these Mediternean countries and the EU. Second, counies may respond to third-country security theat by forming a regonal arrangement. For instnce, the Southern Afican Development Coordination Conference (SADCC), which evenuy developed trading aangements under the Southen African 3See Fernandez (1997) for an insightful discussion of these arguments. 2 Development Community (SADC), was formed to provide a united fiont aganst South Afica. The Gulf Cooperation Council (GCC) was created in part in rsponse to the potential threat of regional powers such as Iran and Iaq. And a major motive of Cental and Eastem European coitries for applying for membership to the EU is as protection against the perceived threat from Russia Third, security threats between neighboring countries may move them to form RIAs. Examples of this motivation are said to include the creation of the Euopean Coal and Steel Community (ECSC, 1951) and the Eupean Economic Community (EEC, 1957) to reduce the thra of war in Europe (see the quote of Jean Monnet at the start of the paper), ASEAN to reduce tesion betweer Inonesia and Malaysia, and MERCOSUR to reduce tensions between Argentina and Brazi (see Bastian, 1996). Page (1996) suggests that this element is also found in the fiomation of APEC and the CACM which include potential political/military opponents. This paper explores this last case-the use of regional trade agrements as a route to rapprochemett between antagostc staes. Our pupose is not to advance this argument, although we shall brefly consider the case that has been made for it, but rther to subject it to a measure of fiomal scutiny. Thus we shall take at fice value the claim that an RIA reduces the tenion between potenial enemies and ask wat implications this has for the form and evolution of the agreement Our purpose is twofold. First, if we accept the hypothess-sy, on the basis of statmts by politicians-we would be able to make predictions about the development of regional integration mad thus possibly to comment on issues such as whether or not it is a stepping stone towards multiatealsm. Second, and more inteestngly, the implications we derive provide a set of testable predictions which would potentially allow us to test whethe or not particular RIAs have stemmed 3 fiom this form of security motive. We say "potentially" because we recognize that in practice RIAs are likely to arise and evolve from a combinaton of forces which my offset the effects of the political motivation. Nonetheless, if our predictions are rejected we may at least refuite the assertion that security was the only or even the main motivation behind an RIA. The rmainder of the paper is organized as follows. Section 2 explores the argment that integation fosters security and in Section 3 this is transated into a formal model. The solution is provided in Section 4 (under symmetry in Secdon 4.1 and under asymmetry in Section 4.2). The relationsip between deep intgrtion and optimm extmal trade policy is examined-in Section 5. Exogenous deep integration is analyzed in Section 5.1 and endogenous deep integration in Section 5.2. Dynamic aspects are examined in Section 6. Bloc enlargment and domino effects are examined in Section 7. Section 8 concludes and Section 9 discusses possible extensions. The model provided in this paper abstcts from defense as an altemative meas of providing security. This issue is on our reseach agenda, however, see Section 9 for a discussion of this and other possible extmsions. 2. Trade amd Security4 The notion of trade as a civilizing influence is an old one-see Hirschman (1982). The notion that international trade is a means of diffing tension and bnngng nations together is also venerable. The nineteeth century British politician Richard Cobden peristently advocated that Britain should trade freely with her neighbors as a means of convincing them of the advantages of free trade and also as a means of locking them more fully into the community of nations. In the 4 We ae grateftul to Doug Irwin for suggestions on this section, but he should not be held responsible for its shortcomings. 4 twentieth century Cordell HulL US Secretay of St (1933-44) and in some wy the ahtct of the post-war interntonal tadig order, advocad this view trughout his publc hfe. IHs auiogra contai homely stories of trade reconn waing neighbors along with discussions of inenatnl strtegy-Hull (1948, e.g., p. 364 and p. 84, respectively). While tade as a reconcile" seems a neceay npining for "rgioalim as security," it is not really sufficet Both Cobden and Hull aso eyxpssd a stog fith inin Thus, while the Cobden-Chevaler Treaty of 1860 bdween Frce and Britin was bilateral, Cobden's concepdon was muilateal and the Tty itself contined an unconditional MFN clase. Hull ws explicit that trade dion bred tnions rathr ta diffiised them (Hull, 1948, p. 81, p. 363).' To obtain an advocacy for regionaism on secty grounds one must tun to continenta Europe. Wilfiedo Pareto appaeny agued in 1889 th -customs unions ...[were]... a meas to better political lions and eveul pi on" (Machlup, 1977, p. 41) and Robert Schuman and Jean Monnet-the founding fh of the EEC-were explicit that the ECSC was to make Franco- Geman war not only "unhinkable, but matrially impossible" (see Swamn, 1992, p. 6). Monet also agued that while Britain, the United SUates and tie USSR coud witdraw into thdr own speres, France and Genmany wee inexticably linkod and had no alteative than to solve the "Eurpean problem." Thus his more regional focus was, pe ul Lat echoes of bnnt's views are common. For eaample, Dr. Walter Hallsoen, a fomer preident of the EC Commission, put it clearly when he stated We are not in buiness at all; we 'Hull advocated the principle of unconditional mfn, which in US usage implies offering mfh triffs to any country offering mfin to the United States. That is, it continues to permit discrimination against those "outside the system." 5 are in politics (see Swann, 1992, p. ix) And Jones (1993), referring to Frace and Geamany, states that "Some trading blocs may be advocated prmarily to avoid military conflicts' (p. 83). 6 7 Hirschman (1981, Chapte 12) makes a similaw point, though he argues that the EC may have arrived a bit late. Security and other "non-economic" aspects of RIAs also seem to have played a role in the Soutern Cone. Argentina and Brazil signed nuclear cooperation and econmic ageemes (e.g., in the area of capital goods and automobiles) in the mid-1980s, quite possibly hoping that the removal of extera tensions would allow each to reduce the power of the military and stengthen its fagle democacy. The creaton of Mercosur in 1991 cornimned this process and bound smaLer neighboring counines into it The political effctiveness of Mercosur is sometimes held to be proven by recent rumors that a posable coup in Paraguay was laid to rest following a pointed reaffimation by the presidents of the four member counties - based on a clause in the Treaty of Asuncion establishing Mercosur - that democracy was a necessary condition for membersbip in the bloc (Survey on MERCOSUR, The Economist, October 12, 1996). While this appears persusive evidence of the link between RIAs and politics, it is not beyond dispute that such conct action depended wholly on the exstence of Mercosur, or that expion from the RIA rather than some 6Jones (1993) also argues that, since open markets imply a loss of national control over the economy, countries may prefer to form RIAs with 'like-minded' neighbors. In other words, the issue of cultual externalities may influence the decision on forming a RIA as well as th choice of parter. For instance, one reason that is sometimes advocated for why a number of North African countries have chosen to form RIAs with the EU rather than liberalize unilaterally is that the latter would subject them to open competition with Asia and possibly threaten their lifestyle. The analysis provied in this paer applies to cultural externalities as well. 7Nowe that Switzerland, which found an alternative way to generate security, namely through neutlity, has so far been reluctant to join the EU. 6 other snction was the critical thrat Thus at least a measure of caution should be exercised in interpreting this expeene. Srinivasan (1994) argues that integration might contribute to reducing tensons in South Asia. He states that "It is conceivable that promoting freer movement of goods, services, people and capital in the regon might also facilitae the resolution of ;olitical and tenntcrial disputes" (p. 7). Political scientists have also discussed the use of trade diplomacy within a regional ont, including whether and what tpe of RIAs might raise welfine of the member coes through "intra-mural conflict avoidance and mas " Bastian (1996) argues that trade negoiations between leaders of neighboring countries are likely to result in a higher degree of trust between them. He mentions that RIAs may enable "...political and/or economic elites to form coalitions for subsequent collaboration and cons _sual action", and that they may a... carry their own language and discourse, thus being able to sociaize the partcipants, e.g., when talkdng about a 'region'." Mansfield (1992) also recognizes the importance of security extemalities in trade relations. However, he argues that countries will lower barris with those who belong to the same alliance (e.g., NATO) because the inreased trade will raise incomes which can then be used to raise defense expenditures; the latter is only beneficial if done between allies, not with adversares. Contrary to Mansfield, we assume that tho very action of trading generates security benefits, irrespective of any income effects. Moreover, Mansfield's analysis is weakeed by the a Note also that the preservadon of democracy in a member country is a somewhat different politl payoff from the alleviation of iter-member coiflict that we are primarily inest in, thou democradc regmies may contbue to scurity by reducing i-member conflict tough inceased tade (Polachdek 1996). 7 assumption that lowering trade barriers towards allies generates income gains, since, in fact, the impact of trade prefrences on income is ambiguous (second best). Evidence on the impat of trade on the likelihood of conflict between any pair of countries is limited. A large number of studies in the political science literature have confirmed the results of Chan (1984) that conflict is less prevalent between any two counties if both are democratic. Polachek (1992, 1996) explains this finding through the effect of democracy on trade. He finds that democraces trade mote with each othr than other countries, and - using detailed data from the Conflict and Peace Data Bank - finds a significant and negative impact of trade on conflict He obtains an elasticity of his measure of conflict with respect to trade of -0. 15 to -0.19. One question is the direcdon of causality, i.e., wheter confict reduces trade or trade reduces conflict Polacek toste this by means of three-stage least squares, and found that the tade variable remained statistically significant and became empirically more important (elasticity of -0.30) in the conflict equation while the conflict variable was not significant in the trade equation. The causality results were confimed by Granger causaity tets (Gasiorowsi and Polachek 1982). Before turig to the modeL we should note that this paper refers lgely, if not exclusively, to RIAs in a geogrphic sense, or to what Ethier (1996) has caLled "regional regionalism". This definition would seem to cover most RIAs since they are geneally formed among neighborig countrie Examples include the EU, Mercour, the Andean Pact, the CACM, NAFTA, ASEAN's AFIA, SAFMA, the GCC, ECOWAS and SADC. Our analysis is thus of potental relevance for the mority of RIAs. 8 3. The Model Assume three countries (1, 2 and ROW) and three nomnal goods (A, B and R). Country 1 produces good A, Country 2 produces good B, and ROW (the rest of the world) produces goods A, B and R. The production, consumption and trade patterns are shown in Table 1. We abstract from optimal taxation issues related to economic power on the world market in order to focus exclusively on matters of security and trade diplomacy.9 Hence, we assume that Counties 1 and 2 are small and that the world prices PA, PB* and Pe (of goods A, B and R, respecdvely) they face are determined in ROW. Table 1. Production, Trade and Consumption Pattern Production Export Import Consumption Country 1A A B, R A, B, R Country 2 B B A, R A, B, R ROW A, B,R R A, B A, B, R Without loss of generlity, units of A, B and R are selected such that PA* = PB* = PR - 1. Denote consumer prices of goods A, B and R in Country i ( i = 1, 2) by PA', PB' and Pr, respectively. Assume a representative consumer Ii in Country i, whose consumption of goods A, B and R is denoted by Al, B, and R1, and with utility 9 See Krugman (1991) for an analysis of custom unions' optimal common extenal tariffs associated witi market power. 9 (1) U, = X(A;, B;, Ri )+ Z; (SIQ; i = 1, 2, XN,A'>O, X2,Zj"O, withpercapitaincomeyl= a and Y2= b. We refer below to Countries 1 and 2 as partners whether or not they are formally members of a RIA. Assume that an increase in home country imports from the partner raises the home country's knowledge of and trust in the parter and reduces the level of insecurity in the home county. In other words, importng from the partner country increases intction with individuals in the parter country, raises the level of information about them, and increases the level of trust and security in the home country.10 Equivalently, larger home country imports from the partner raises the imporunce of the home country as a market for partner exports and, given the inreased cost of klling the goose that lays the golden egg, lowers the likelihood of a security threat by the partner country. Assune that the level of security in the home country falls as the partner's relative income increases becuse it implies a greater relative power for the partner and thus a larger potential security threat for the home country. Assume also that the higher the partner's relative income, the greater the home country's security gains from additional imports from the partner. Thus, (3) SKI = SK (B1l, Y2/Y1); SK2 = SK (A2, YN/Y2)11, 9 Note that in this setup, the level of home coutry imports of the partner's good equals the level of home county consumption of the partner's good (since it is not produced in the home country). 11 In fact, SK depends on total (rather than on I 's per capita) imports from the partner, so that the term Li should be included in the equation. Excluding it does not affect the model's solution (though it matters in the case of population changes, as noted below) and is done for notational simplicity. 11 with SKI twice continuously differentiable, SKI (SK2) ineing in B, (A2) and deceasing in Y2/YI (Y,/Y2), stricty concave, and with aSK,'/8(Y2/Yl) > 0, aSK2'/a(Y,Y2) > 0 (where SKI' (SK2') - 8SK,/8B1 (MSK2/aA2)). The latter implies that the security beefits fiom import from the part incease as the partner's relative income increases. 4. Solution Initviduals I, maxmim U; with emspect to Ai, B1 and R subject to the budget constaint Y, = PA,AI + PB' B, + PR' R, and taking SK; as given exogenously (i =1, 2). The first-order conditions are (4) XA'/XR = PA/ PR', XB'IXR = PB'/ PR'; XjauoX/8 (i - 1,2; j= A, Bi, RP. n te absence of domestic ortrade xes in Countries I and 2, Pji = p* = I (i = 1, 2; j = A, B3, RJ), and individual I,'s utility maximization implies (5) XAIXR' = XB = I . Even though Country 1 (2) is small, the absence of domestic or trade taxes does not maximize welfare because of the exmralities associated with the impact of imports B, (A2) from Country 2 (1) on SKI (SK2). A policy of zero trade and domestic taxes maximizes national income L1X-, i.e., it maximizes the part of weLfar W; = LI.U, which depends directly on the consumption 12 of goods and services. However, welfare depends on national income L3.X1 as well as on the benefits from security L1.71. Individual utility maximization is based on the assumption that oX/8j = MU/oj for Vj, while social welfare maximization recognizes that 8X/8J < U/A forj = B1, A2, with the home country's social gain from imports from the partner larger than the private gain by the impact on security. It is well known that welfare maximization requires distortions to be attacked at the source (Bhagwati 1987), and policy should result in the intemalization of the extemalities. In this case, since SK is a function of imports, welfare W; = L1.U, in the home country is maimized with a subsidy Si on imports from the patner country (assumed to be financed thomugh lump sum taxation) equal to (6) Si = Li.Zi'.SKi'/. = Li.Zi'.SKi'/XR'; i = 1, 2, where A; is the marginal utility of income. Note that = X/P* = X3', the maginal utility of R Note also that the optimum subsidy Si is equal to the optimum subsidy rate defined as a share of the world price (since PA* = PB = 1). Note also that Si need not increase with Li because Y, ixcreases with Li (equation (2)), and SK,' falls with an increase in Y; (equation (3)). Since Countries 1 and 2 continue to trade with the ROW following the import subsidy Si (see equation (1)), producer prices in Countries 1 and 2 are not affected by the import subsidy. The only effect is to lower the consumer price of imports from the partner. This is shown for Country I in Figure 1 where all quantities are in per capita terms. (DB5) measures the private (social) value of good B for Country 1. The demand price on the 13 Figure 1. DsA PB~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~B (P2 )OPT T--= --- 0 BP B p B4s 'social' demand curve DBS (i.e., the social value) is the sum of the private value and the value of the extraity L,.Z,'.SKI'/XR' = L,.Z,'.SK1'/X,. Following the strict concavity assumptions made in equations (1) and (3), the vertical distance between the two demand curves falls with Bl. Private equilibrium in the absence of interventions is at point E, where XB'/XR' = PB*/Pt* = 1, and imports equal to BIP. The social optinum is at point F, where (XBe + LI.Z,'.SK')/XR1 = PB*IPR* = 1, and imports equal to B,S. This solution is obtained with a subsidy on imports of good B from Country 2 equal to S, = Ll.Z,'.SK,'/XR' = L,.Z,'.SK,'/X,. This is shown in Figure 1 by the line PBI, the consumer price of good B in Country 1 - obtined as the world price PB* minus the import subsidy (or PB* - S, = I - S,) - and which intersects the private demand curve D8r at point G, with subsidy S, = FG and imports B,S. The relative consumer prices following imposition of the subsidy rate S, are: 1 x - y _ 2/P 2 -2/P 2-1 (7) PA /PR1 = , PBI/PR'= I 1, PA PR =1 - S2, PB PR As can easily be veified, the same relative price configurtion as the one obtained with the subsidy rate Si onpartner imports can be obtinedwith ataxrate T; = S-/(l - S) on imports from the ROW and on consumption of the home good (with the tax returned in lump sum fashion). Thus, welfare of Country i (i = 1, 2) is maximized either with a subsidy rate S, on imports from the partner, or altenatively, by Countries 1 and 2 forming a RIA with a tax rate on imports from the ROW of T, = S-/(I - S;) and an equal tax rate on consumption of the home good. Note that even if the tax is imposed exclusively on imports from the ROW and not on consumption of the home good, there is a positive tax smaller than T; at which welfare is higher 14 than in the absence of interventions, though welfare is lower than with the optimim subsidy Si or with the optimum tax T; applied to both imports from the ROW and to consumption of home goods. The impact of the optimum intervention on per capita income and welfiLre can be seen in Figure 1 (measured in currency units). Benefits from security Z, increase by area EGFK, income XI falls by area EFG, and welfire U1 rises by area EFK. Note that since the security extenality falls as imports increase, area EFK is larger than area EFG. The income loss for Country 1 from the RIA can be appr ma by S12.BB.ss/2PBI, whee B is the elasticity of demand for good B in Country 1. Thus, S12.BI.sBn2PB1, the income loss measured as the loss from the RIA relative to a free trade situation, provides a lower bound of the welfiare gain EFK generated by the optimum intervention. 4.1. Symeryn Under symmetry between Countries I and 2, LA = LB L, a = b, YA = YB. and Zl = Z2 Z-. Since U is symmetric in A and B, B = A2 , SKA = SKB m SK, and UA = UB = XXM, KR) + Z[SK(M 1)], with (8) SI = S2 a S = L.Z'.(oSK/MY/XR= L.Z'.(8SK/8M)/;. Thus, under symmetry, welfare of Countries 1 and 2 is maximized either with a subsidy rate equal to S on imports from the partner, or with the formation of a customs union (CU) between 15 Countries 1 and 2 with common extemnal tariff rate CET = S/(1-S) and with a tax rate T = CET on consumption of the home good. Note also that in the symmetric case, trade between the partners is balanced and so is their trade with the ROW. This is not necessarily the case under asymmetry. 4.2. Asymnetry Assume symmetry between Countries 1 and 2, but with one exception: for any SK1 = SK2, Z2' > Zl'. n other words, increasing security is more important in Country 2 an in Country 1. This may be due to the fact that Country 1 has a stronger tendency to resort to aggression to resolve disagreements or that Country 2 suffered more in previous wars han Counry 1.12 The equlibrium in tie previous subsection would no longer be sustainable because Country 2 now desires a subsidy S2 > SI (see equation (6)). In other words, the optimum subsidy for Country 2 - the country with the stronger security concerns - is larger than that of Country 1. Altematively, Countries 1 and 2 can maximize welfare by forming a FTA with a tax rate T1 = S-/(I - S.) on imports from the ROW and on consumption of the home good, and with T2 > TI.'3 Thus, optumum taiffs on imports from the ROW are higher for the country with the stronger security concems. Note that with S2 > SI, B, < A2. Thus, at the social optimum, Country 1 runs a trade surplus with Country 2 and a deficit with the ROW (and vice versa for Country 2). In other words, at the social optimum, the country with stronger security concems imports more from its partner. 12 For instance, France may have been more concerned with security than Germany in the first half of the 20th century while the opposite may have been true in Napoleonic times. 13 The FTA requires rules of origin to prevent deflection of R from Country 1 to Country 2. 16 This is shown in Figure 1. Country 2's private demand curve DAP is identical to DBP. However, the 'social' demand curve for Country 2 is DAS > DBs, and the price fimction PA2 (equal to the world price minus the optimum subsidy function S2) intersects the demand curve at point J. The subsidy S2= HJ > SI = FG, and B1 < A2. Alternatively, assumne symmetry between Countries 1 and 2 except for Y, > Y2. This may be due either because a > b, LI > L2 or both (equation (2)). Assume a > b and LI = L2. From equation (3), 82SK1/aB,a(Y2/Y1) > 0 and 82SK2/8A28(Y/Y2) > 0. Thus, SK2' > SKI' for A2 = B1 (i.e., for the same level of imports from the partner). In other words, the marginal impact of imports from the partner on home country security is larger for Country 2 than for Country 1. This would imply S2 > SI. On the other hand, the marginal utility of income X1 < X2 since yi > y2 (i.e., the demand for security increases with per capita income and is thus larger in Country I than in Country 2), implying S2 < Sl. Thus, whether S2 is larger or smaller than SI depends on wheher the first or second effect is larger. If Y, > Y2 is due to the fact that LI > L2 (with a = b), thethe first reason for S2 > SI mentioned in the previous paragraph continues to holds, but the second one does not since Yi = y2. Thus, S2> SI in this case. Note that in both the symmetric and asymmetric cases, the welfare of both Countries 1 and 2 would also be maximized if one of them applied the optimum subsidy on imports from the partner while the other applied the optimum tax on imports from the ROW and on consumption of the home good. The formation of a RIA (CU or FTA) would require coordination to ensure that both Countries 1 and 2 applied the tax. If we added to the model the assumption that policy coordination provided additional security benefits by improving trust and understanding among the 17 leaders (and negotiators) of both countries (see the quote from Bastian 1996 in Section 2), then a RIA (accompanied by appropriate domestic taxes) would be superior to subsidies on imports from the partner, and welfare maximizing governments in Countries 1 and 2 would coordinate their policies. On cooperative solutions, see Section 8 (Extensions). What has been shown so far is that, in the presence of security extemalities, the formation of a RIA provides an optimum (though not the only one). In what follows, we examine the implications of the formation of such an optimum RIA for the evolution of extemal tarffis over time as well as following deep integration and enlargement. The issue of whether Countries 1 and 2 are symmetric or asymmetrc has no significant imnpact on results in the renainder of the paper. Unless noted otherwise, symmetry is assumed below. 5. Trade Preferences and Deep Integration It is often argued that the focus on the static second-best aspects of regional integation may be misplaced and that gains can be expected from deep integration, including hrmonimtion of technical standards and of investment codes, and general facilitation of movement of goods and factors. Deep integration is expected to result in lower real trading costs and welfare gains for member countries."4 A recent example is EC-92 which resulted in lower barriers on intra-EC trade. The impact of EC-92 has been estimated by Smith and Venables (1988), Baldwin (1989), Smith (1992), and Harrison et al. (1996a). 24 However, welfare gains are by no means certain if RIAs are not optimal since lowering internal trading costs may in principle result in ixmriserization. This cannot occur in our model. 18 An unresolved question is: Are regional trade preferences necessary for or supportive of deep integation on a regional basis? The relationship between deep integration and trade prefces is examined in the case of exogenous deep integration in Section 5.1 and in the case of endogenous deep integration in Section 5.2. 5.1. Exogenous Deep Integation Assume initially that trade with the ROW involves the same trading costs C as intra-bloc trade. Following deep integration, the trading cost for intra-bloc trade falls to C0 < C while trading costs with the ROW remain unchanged at C. Since trade with the ROW continues following deep integration (see equation (1)), producer prices in Countries 1 and 2 are not affected by deep integration, but consumer prices in the home country for goods imported from the partner (BI in Country 1 and A2 in Country 2) fal by an amount C - C0. Consequety, consumption (= imports) of B1 and A2 ises. Given that Z", SK' < 0 (equations (1) and (3)), the optimum subsidy Si falls (equation (6)). A reduction in trading costs means that the PB* line in Figure 1 shifts downwards and the intersection point E (of the PB line and the private demand curve D) shifts to the right, and similarly for the intersection point F (of the PB line and the 'social' demand cuve DB3). Since the distance FG = SI falls as B1 increases, the optimum subsidy falls. Equivalently, lower tading costs on intra-bloc trade imply lower optimum taxes on trade with the ROW and on consumption of home goods. Thus, in the case of exogenous deep integration within the region, trade preferences and deep integration are substitutes: lower intra-bloc 19 trading costs imply lower optimum trade preferences. 5.2. Endogenous Deep Integration Altematively, assume that lowenng tade barriers on inta-bloc trade is costly and that the cost is negatively related to the level of tust or security capital SK. Since trust is positively related to imports from the partner, a larger volume of such imnports reduces the cost of deep intgrtion and is thus likely to result in more deep integration and in lower unit trading costs on trade with the partner. Haas (1958), saw increased trade among members of the EC-6 as having a positive impact both in the area of international politics and of deep integration. He wrote (p. 311) that "... it is inconceivable that the liberaliztion not only of trade but of the conditions governing trade can go on for long without '"armoniation of general economic policies" spilling over into the fields of curency and credit ...The spill-over may make a political community of Europe ..." We examine two altemative cases. In the first one, deep integration takes place every period. In the second one, deep integration takes place once and for all. Assume unit tading costs C on imports from the partner are (9) C = C[SKM)], 8C/aM = (BC/8SK).SK' < 0, where M is the volume of imports from the parter (and M = BI = A2 under symmetty). In this case, there are two sources of positive extemalities - rather thaone - associated with imports from the partner. Individual Ii who imports from the partner generates an extenality in 20 terms of additional secunty. And additional secunty raises utility directly (equation (1)) as well as indirectly by lowering the unit cost of intra-bloc trade. The total trading cost TC on imports M from the parter and the marginal social cost of tading MSC are (10) TC = M.C[SK(M)], MSC = C.(I + £), E - (OC/8).(MC) < 0. Individual I; takes the private cost C of importing from the parter into account but not his negative effect on unit costs s.C. The optimum subsidy SE is (I 1) SE = Z .(oSK/&v)/ - £;.C > S, with S defined in equation (8). Equivalently, the optimum is obtained with a common extenal tariff CETE = SE/(I - SE) > CET = S/(1 - S) and a tax on home consumptionTE = CETE> T = CET. Altmeatively, assume that if deep integration takes place, it does so at a given point in time (e.g., in 1992 for EC-92) and is pemmanent. For instance, assume that trust and security SK1 do not depend on the flow of current intra-bloc trade but on past flows as well (see Section 6 on that) and that deep integration will take place when SK; is sufficiently large and the cost of deep integration is sufficiently low. Tlen, since increased intra-bloc trade generates extenalities in terms of lowering the cost of future deep integration, the optimum subsidy on intra-bloc trade (or optim CET) is higher before deep integration. However, once deep integration has taken place and intra- bloc trading costs have fallen, the optimum subsidy (or optimum CET) is lower. 21 Thus, if the cost of deep integration is unrelated to the degree of trust between Countdes 1 and 2, not only are trade preferences not required in order to implement deep integration measures but deep integration actually implies a reduction in optimum trade preferences. On the other hand, if deep integration vanishes every period and its cost declines with the degree of security and trust, optimum trade preferences are larger. And if deep integration takes place once and for all and is permanent, optimum trade preferences are higher before deep integration takes place and lower thereafter.'5 6. Dynamics The EC has reduced its extenal trade barriers on manu ing products with respect to the ROW over time. Average tariffs on manufacturing products have fallen from about 13% in 1958 to about 3% after the Uruguay Round." Similarly, developing countries forming RIAs in the 1960s imposed high extenal trade barrers while recent RIAs and new incamations of old RIAs have tended to impose lower external trade barriers. Is this gradual reduction of extemal trade barriers over time predicted by our model of regional integration? So far, security capital SK has been assumed to depend on the current flow of imports from the partner. In other words, SK has been assumed to depreciate fully at the end of each period. In l' Of course, there are different degrees of deep integration, and though the 'first' degree (say, EC-92) occurs once and for all, it may be followed by a 'second' degree of deep integration (say, monety union). The inplications for the evolution of external trade barriers is ambiguous in this case. It depends on the 'production functions' of the later degrees of deep integration. If, as seems plausible, the marginal product of the sum of the early degrees of deep integration in the production of later degrees of deep integration rises with the degree of deep integration, and the marginal product of intra- bloc trade falls with the degree of deep integration, then optimum external trade barriers fall as the degree of deep integration increases. 16 Note that the reduction in the EC's external trade barriers over time has taken place despite the increase in the number of member countries from six to fifteen, i.e., despite the increase in the optimum CET level based on market power considerations. 22 fact, though, SK (trust and goodwill) depends not only on current behavior but on past behavior as well. In other words, it does not depreciate entirely at the end of every period. Assume that the level of SKt at time t depends on the curet flow of imports from the partner and on SKt.1 at t-1, with a rate of depreciation of SK equal to 8 < 1. Expressing variables in continuous time, equation (3) becomes, under symmetry, the law of motion (3'): (3') SK =FM) - 8.SK; 8 <1, where SK m aSK/at is the time derivative of SK, F(M) is the curent gross addition to the stock of security capital (with F' > 0), and the subscript 't' on SK and M has been deleted to simplify notation. Since Y2/Y, = YIN2 = 1 under symmetry, the terms have been omitted from equation (3'). Start at SK = 0. As SK increases, its marginal utility Z' falls (equation (1)). If governments are myopic and choose S to maximize current welfare W = L.U in each period, then from equation (6), S falls over time. However, optimizing govemments will also take into account the impact of current subsidies on the level of SK and thus on future welfare. Assume the governments of Countries 1 and 2 select an optimum time path for the subsidy S on imports from the partner in order to maxmize V, the present value of W= L.U. The Hamiltonian is (12) H = W(SK, S, t) + q. SK (SK, S, t), 23 where q is the marginal value of SK, WVmA/oSK, where VmAx is the maximum value of V obtained with the optimum path for S. The solution (or maximum principle) is (l3a) 8Haq = SK = F(M) - 8.SK (equation (3')), (13b) 8W8S = 8W/8S + q.(aSKI8S) = O, and (13c) 8W8SK = 8W/ASK + q.(aSK/8SK) = - q. From equation (13a), aSKIaS = (aSKlaM).(M/aS) > 0. Thus, from equation (13b), MW/MS < 0. Why is 8W/oS < O? If a govemment myopically maxdmized current welfare without regard for future welfare, it would choose a subsidy level Sw so that aW/oSw = 0. However, an increase in the subsidy also genertes future welfare gains because it raises M and thus raises SK Thus, the subsidy Sv which maximizes V is larger than the myopic Sw which maximizes W. Consequently, 8W/8S < 0 at the optimum. Assume SK is low (say following a series of wars, e.g., in the 1950s for France and Germany) so tbat SK rises with imports M > 0 between the two partns (equation (13a)) and the value of SK falls, i.e., q < 0. From equation (13b), a reduction in q over time implies that 8W/OS fills over time in absolute value or becomes less negative since aSKloS falls with SK as well in absolute value.17 In other words, along its optimum path, Sv falls oVer time and approaches SW. 17 asKlas = '(asK1asK).(asKI).(MIas) = -8.(aSKI8M).(&M/aS). The depreciation rate 8 is constant, 8MM86 is ind epeneof SK since U is additive, and SK/M * SK' fails with M, so tbat 8SK/aS falls widt SK in absolute value. 24 Note that S also fills as SK nses because the marginal im of SK on W fiall with SK (Z" < 0). Equivalently, the opdmum common external tariff CEr (recal we a assuming symmetry) fills over ime. Note that the decline in the CET will stop once a steady state is attained, i.e., once SK= (equation (3')). On the ote hand, there may be times of crnsis or rversal where SK suddenly falls. TIhe, the optimum CET would suddenly iaeb e grdually sarting to fail againL If the cisis is temporary, with no impact on behavior, CET eventuay ream to its previous steady-stae value. On the otber hd, a deep crsis may affct security prefenes or the Z-funm in equation (1) as well as the secunty production fumtion in equation (3), with a pemaent chge m optimm exoteal barriers. 7. EnlarSement and Domino Eflht 7. i. Symmetry We now examine the impact of bloc enklareent on welae and on the opdmum CET. Assume a fourth country (Coutry 3) which produces good C, eaot it to Countis 1, 2 and ROW, and imports goods A, B and R. Country 3 is symmeic with Coutie 1 and 2. Eah one of Countries 1, 2 and 3 enjoys the same degee of trt (or sufs the same degee of mistrust) with spect to the other two. Utility for individual I4 in Country i is now (I') U, X(A1,Bi, Cl,RO+Zi (SK1; i - 1, 2, Xj,Z '> O, XZE, A"< 0, Xi U0) m X * aj(j - 4 b Ci, Rj. 25 The level of secrity in the home comtry is positively related to imports from both puter counties (and positively related to its relative income with respect to both), i.e., (3") SKI = SK (B1, CI); SK2 = SK (A2, C2); SK3 = SK (A3, B3), with SK symmetic in both imports (and with the relative icome tems not included in the equation for simplicity). Assume the ree countries (Counties 1, 2 and 3) impose opimum subsidies on trade with each other. Under symmetry, each country imposes the same subidy rate S as the other two and imposes the same subsidy on its imports from both partn The some esult is obuined with a common extemal tariff rate CET = S/(1-S) and a tax rate T - CET on the consumption of the home First, we note the each member country's welfame nses when the RIA expands from two to three countries. In the two-country RIA case, th optimum subsidy is appihed on the imprts of one country only wbile security benefits are obtained through imports from two countries. This case is equivalent to a consrained optimizaion where one of the two subsidies is set equal to zero. The optimum in this case rsults in a level of welfare which is necesarily lower than under unconstraie optimizaion in the case of the thre-country RIA, since in the ter case all xtrnalities associated with the impact of imports on security are intalized. Consequently, startng from a two-member country RIA (say, CountLries I and 2), there is both a demand (by Country 3) and a supply (by the RIA members) for enlargement In this sense, there is a domino 26 effect: if a two-country RIA does exist, the third country will join. However, if no two-country RIA exsts, the tre countries have a incentive to form a RIA simulta sly. We now turn to the impact of enlagment on the optimum CET. Define the opimum subsidy S and optmum CET for a bloc with k membes (k - 2, 3) by ek and CET&, reectively. What is the relationsbip between S2 and S3 (or CET and CET3)? We show in the Appendix that Sk need not increasme with ge, though e2 < S3 is likely to hold on aveage. This result gnizes to S> < Sm for a bloc xansion from m - I to m member countries in a woid of m symmetric countries (m 3), but not to Sl < Sk for a customs union ofsi k (k < m) in a wod of m symmetic countries. In a word of symmetic blocs with the optimum CET detemined by maiket power, the CET increases with enlargement In our modeL the optmum CET need not icrease with enagee though it is likely to do so on avae. 7.2. y We now eamine enagement ndr a specific case of asymmetry. Even thugh the main security concem in the 1950s in Western Europe was with Fne and Germany, four countries (Belgium, Nethnds, Lu g and Italy) joined France and Germany to form the EC-6. These four counties were smaller economically and weaer miitrily. Simily, Argetina and Brazil decided in the 1980s to integrate their economies, and Uruguay and Pguay - two small On dmino effects, se Baldwin (1995). 27 neighbors - decided to join them and form Mercosur. And Chile and Bolivia recty signed FrAs with Mercosur.19 Can the present model say anything about dommo effects in such asymmetric situations? Imagine that Countries 1 and 2 are symmetric and are large compared to Country 3 (though they ae stdll small economically compared to the ROW). Assume that, given the small size of Country 3, security in Countries 1 and 2 depends on the level of trade between them and not on the level of trade witi Country 3. Counties I and 2 have an incentive to form a RIA. The latter reduces the level of security in Country 3 becu Counties 1 and 2 import less from it. Thus, Country 3 has a stnger incentive to join in a tree-country RIA tan it did to join with dther Country 1 or 2 before the latter formed a two-country RLA. On the other hand, Counties 1 and 2 have no incentive to allow Country 3 to join the RIA and Country 3 may have to offer some compensation to Counties 1 and 2 in order to be allowed to join. If the gains to Country 3 from joining are larger than the losses (if any) to Counties 1 and 2 (say, because they now trade more with Country 3 and less among themselves), Country 3 will compensate them and join the RIA.?0 For a North-South model where the developing country makes a side payment to the developed country in order to generate a RIA, see Ethier (1996). 19 On a comparson of the impact for Chile of a FrA with Mercosur versus a FTA with NAFTA which does not include security externalities, see Harison, Rutherford and Tarr (1996b), Schiff and Sapelli 1996) and Schiff and Ingco (1996). In some cases, the gains for Country 3 may be very large. For instance, Paraguay suffered a devastating defeat in the 1860s in its war (known in Spanish as 'Guerra de la Triple Alianza') with a coalition made up of Brazil, Argentina and Uruguay, and where most of its male population over 15 years of age perished. Such a collective memory may provide a strong incentive to avoid future potential conflicts. 28 S. Conclusion This paper has examined a world in which regional trade agrements offer scope to reduce security tensions between neighbors Assumning that security with a neighbor incases as imports from tbat neigbbor incease, and making no assumptions about the relative sizes of trade diversion and trade creation, we have shown tha: 1) The formation of a customs union (CU) - ad by apppiate domestic xes - proides an optuum economic arrangement under symmtry, and the sme holds for a fiee trade agreemet (FTA) under asymmety; 2) Deep integration (such as EC-92) implies lower optimm extenal tariff if it is excogenous, higher opmum exmal tff if it is endogenous and vanishes annually, and higher optimum extenal tarff before deep intgation and lower ones thereafter if deep integmtion is edogenous, take place once and for all and is pmanent; 3) If the level of secuity depends on current as well as past trade flows and is in stady state in he absence oftrade barriers, the optimum extnl decline ovrtime; and 4) En t of bloc siz (in tems of the number of simmetrc countries) implies higher welfare, vwth an ambiguous impact on the optimum CET though it is likely to be higher, and some form of domino effect easts. Though exteralities associated with security matters imply that a RIA may maxim welfae, the model suggests that the RIA is a transitory arangement in the sense that optimum trade prefernces are highest at the time the RIA is formed (wben secuity is low) and tend to decline over time. In other words, the RIA's extenal trade policy becomes inreasingy open over time. 29 9. Extensions First, our model abstacts from defense enditues as an altenative way to gerate secuity. This was done for the sake of simplicity and to avoid game-theortic issues at tis stage. We recognize the importnce of such issues, which are on our research agenda. In the case of defense expnditures, the cooperative solution is clearly optimal as the altertve may be a prisoner's dilemma situation with large defense penditures (and not necessarily a higher level of security). In that case, the "peace dividend? foilowing formation of a RIA may be substntial (see Srinivasan 1994, p. 7, for a discussion of dtese issues in the case of India and Pakistan). Second, the model assumes that imports fiom - rather an trade with - the partner country geneates secuity benefits in the home country. Altermatively, one could assumne that total trade (imports plus exports) with the parter generates secuity benefits. In that case, RIAs are likely to be superior to optimum intra-bloc trade subsidies. Thue are both game-thoretic and pracdcal reasons for this. First, each country will use a subsidy to maxmize national welfre. However, individual decisions in the home countly generate externalities not only for other individuals of the home country but also for those of the partner country becae home country's imprts are the paruer's exports. Thus, intra-bloc trade subsidies which maximize national welfare generate a Nash equilibrium, with welfare and Nash subsidies lower than in a cooperative equilibrium which could be obtained bv forming a RIA. Second. the Nash eauilibrium implies subsidies on both imports and exports. Export subsidies are GAT-illegal so that even the iufeior Nash solution cannot be implemented. The superior cooperative solution implicit in the fonnation of a RIA (with appropriate domestic taxes) gets around this problemL 30 We examine here dhe impact of enargement on the optimum subsidy or CET. Equation (1') in Section 7.1 can be rewritten as (14) U = X(A, B, C, D) + Z[SK(B, C)], where the subscript 'i' has been deleted for simplcty. Without loss of generality, we examine the issue from the viewpoint of Country 1. Define Sk as the optimum subsidy in a bloc of k counties (k = 2,3). From equation (14), we have (15) dU/dS = [(MXIOA).(dA/dS) + (OX/OB).(dB/dS) + (MX1UM.(dC/dS) + (8X/aR).(dR/dS)] + Z'[(SK/aB).(dBtdS) + (aSK/MC).(dC/dS)]. Note dtat sine inome mesured at world prices, Y, - A = aL, doe not chae with the subsdy, and since all wod pri are unity, balaced trade implies dA + dB + dC + dR = 0. Assme we start with S = S3 and apply it to imports of both B and C. Since e3 is an optimum, dU/dS = 0. Given symmety, B = C > A - R, OX/BA - 8XIOR>8MM - XBuc, dB/tdS -dC/dS -dA/dS - -dR/dS> O, and .SKM/B = 2fK/aC. Substituting these in equat (15), ue have (16) dU/dS - 2(dB/dS). I(MXUaB - MX/BA) + Z'.(aSKaB)] - 0. 31 Altbraively, assumthat S = S 3aplies only to imports of B and the subsidy on import of C is set equal to zro. Such a configration of subidies is not an optimum, so dU/dS o 0. In this case, A - C R < B, dA/dS dC/dS dR/dS < O, dB/dS = -dAdS - dC/dS - dR/dS - -3(dA/dS), and dX/8A = 8X/8C - 8XU8L Substibing in equation (15), we have (17) dU/dS - (dB/dS).4(Xa/B - MX/MA) + Z'.(8SK/8B)] - (Z'/3).(8SKI/8C). The objective is to evalte dU/tS in equation (17). If the sign wer negative (positive), then S would have to be rdced (increased) to reach the optimum subsidy s2 in the two-countty RIA case, i.e., then 2 < S3 (S2 > S3). Note the similarities betwe equatios (16) and (17). The term r(A, B, C, R) - [(8XU8B - 8X/OA) + Z'.(aSKIoB)] = 0 in equation (16). If both equations were evaluated at the same values for A, B, C and R, then dU/dS in equation (17) would be d/dS - r(A, B, C, R) - (Z/3).(SK/8C) - - (Z'/3).(aSK/aC) < 0, and tus e2 < S3. Howeve, equations (16) and (17) are not evaluaed at the same values for A, B, C and R, so that r(A, B, C, R) is not necessaly equal to zero in equation (17). We now show that the value of r(A, B, C, R) (8W8B - 8X8IA) + Z'.(8SKI8B) can be ether positive or negative when evaluated at the quantities pevailing in equation (17). The diffence between equations (16) and (17) is that the subidy at the re S3 in equation (16)-is iet equal to zero on good C in equation (17). hub, B and A are both larger in equation (17), 8XWB and 8X/8A are both smaller, and the inpact on (8Wo8B - 8X/8A) is ambiguous. lhe impact on Z'.(MSK/OB) is ambiguous as well. At the larger value of B in equation fl 7, aSK/aB is likely to be 32 smaller, but since SK is also lIkely to be smaller, Z' is larger, and the impact on Z'.(aSK/aB) is ambiguous.2' In other words, the sign of r(A, B, C, R) in equation (17) is ambiguous. For some fimctional forms of X, Zand SK, r(A, B, C, R) will be negative, while for othrs, it will be positive or zero. One would need to know the exact fmnctional fonns of X Z and SK in order to be able to deterine the sign of r(A, B, C, R) in equation (17). However, under symmetry of the postive and negaive values for r(A, B, C, R) across all functional forms for X, Z and SK, one may expect its average value to be zeron?2 And since -(Z'/3).(aSK/aC) < 0, one may expect the average value of dU/dS to be negive. This implies that on average, S2 < e. It is easy to show that the reswlt gnlizes to Sw' < Sm for a bloc expansion from m - I to m member counies in a world of m symmetric countries (m 2 3), but not, unforuately, to S-1 < S' for a customs union of size k (k < m) in a worid of m symmetiic countries 21Comparing equiLibria in equations (16) and (17), a second-degre Taylor expansion of the difference in SK is: ASK - [AB.(aSK/8B) + AC.(MSKQ8C)] + [AB2.(82sK1) + AC2.(82SK/aCt)J2 + (AB.AC.(O2So'IB8C)J. What is the sign of ASK? Ihe first term in square brce is negative. To show this, notetbat B- C in equaon (16) so hat MSK/B - aSK8C, but AB < -AC (the ihse in B is smaLer than the decline in C because removal of the subsidy on C also leads to a rise in consumption of A and R, and from the budget conain, -AC - AA +AB + AR > AB). The second tenn in square bracks is negative sine the scond derivatives are negative. (ntutvely, as C fills, 8SK8C rises so dtat tie loss in SK from addidonal reducto in C increases; similarly, the gams in SK from increases in B declie with B). On the otier hand, 8oSK/aBC may be negative (if Coutry 1 is lss conee about securitywih respect to Country 2 (3) if security with respect to Country 3 (2) is higher), and the third term mi squae bmckes would then be positive. Thus, toug it may seem likely ht removiag the subsidy on imports of good C results in ASK < 0, this need not be the case. However, no matter what the sign of AM, the fact remains tht te difference in Z'.(8SK/8B) is ambiguous. The reason is that with an ambiguous change in SK, the changes in both Z' and in aSK/aB are ambiguous. 22Stricdy speing, whether the average value of r(A, B, C, D) in equation (17) is zero or not depends on the definition of the domain of the functions X. Z and SK. 33 Rerence. Richard E. Baldwin.1989. "The Growth Effects of 1992". Economic Policy 9: 247-81. _. 1995. "A Domino Theory of Regionalism", in R. E. Baldwin, P. Hayand J Kiander (eds.) Expanding Memebrship in the Europan Union. Cambridge Univ. Press: Cambridge. Jens Bastia. 1996. "Trade Diplomacy and Regional Integration", Report on World Bank Workshop, London School of Economics, July 3. Gary S. Becker. 1996. "Prefrnc and Values", Chapter 1 in his The Making of Prefces and Values. Harvard Univ. Press: Cambridge, Mass. Jagdish Bhagwati. 1987. "The Generalized Theory of Distortions and Welfare, in J.N. Bhagwad (id.) Internadonal Trade: Selected Readinws (Second edidon). M.I.T. Press: Cambridge, Mass. Bial Christopher Bliss. 1994. "Trade Blocks and Migration", chapter 5 in his Economic Theory and Policy for Tading Blocks. Manchester and New Yorc Manhester Univ. Press. Olivier Cadot, Jaime de Melo, and Marcelo Olarr. 1996. '¶Regional Integation and Lobbying for Tariffs Against Non-Members". CEPR Discussion Paper # 1448. Steve Chan. 1984. "Mirror, Mirror on the Wall ... Are the Freer Counthies More Pacific?' Jomnal of Conflict Resolution 28: 617-48. Wilfied J. Ethier. 1996. "Multilaeral Roads to Regionali". Seminar Paper 96-15, Centre for Internaonal Economic Studies, Univ. of Adelaide (October). Raquel Fenandez 1997. "Returns to Regionalism: An Evaluation," mimeo. Intemational Trade Division, World Bank. 34 Marc Gasoriowski and Solomon Polachek. 1982. "Conflict and Interdependence: East-West Trade Linkages in the Era of Detente". Joumal of Conflict Resolution 26: 709-29. Gene Grossman and Elbanm Helpman. 1994. . 1995. Ernst Haas. 1958. The Uniting of Europe. London: Stevens & Sons. Glenn W. Harrison, Thomas F. Rutherford and David Tarr. 1996a. "Increased Competition and Completion of the Market in the European Union: Static and Steady State Effects". Joumal of Economic Integation 11 (3): 332-65. _ 1996b. 'NAFTA, MERCOSUR or Multilateralism: A Quantitative Evaluation of Chile's Trade Policy options", mimeo, IECIT, World Bank. Albert 0. Hirschman. 1981. Essays in Trespassing. Economics to Politics and Beyond. Cambridge: Cambridge Univ. Press. 1982. "Rival interelations of miarket society: civilizing, destructive, or feeble?' Journal of Economic Lit vol. XX, pp. 1463-84. Cordell Hull. 1948. The Memoirs of Cordell Hull, MacMillan, New York Ronald Jones. 1993. "Discussion" (ofKrugan, Ch. 3), in J. de Melo and A. Panagariya (eds.) New Dimensions in Regional Integrtion, Cambridge Univ. Press: Cambridge. Immanuel Kant. 1992. Perpehtal Peace. A Philosophical Essay. Thoemmes Press, Bristol. Paul R. Krgan. 1991. -s Bilatealism Bad?', in E. Helpman and A. Razin (eds.) Internional Trade and Trade Policy. Cambridge, Ma.: MIT Press. Robert Z. Lawrence. 1996. Regionalim, Multilateraism, and Deeper Integraon. Washington, D.C.: The Brookings Institution. 35 F. Machlup. 1977. A History of Thought on Economic Integon, MacMillian, Londo Marcelo Olarega and Isidro Soloaga 1996. "Explaining Mercosur's Tariff Stuture: a Political Economy Approach". Mimeo, IECIT, World Bank (October). Sheila Page. 1996. -Intensity Measures for Regional Groups." Paper prepare for EADI 9th General Confence, 11-14 September, Vina Arvind Panagaiya. 1995. "Rethinking the New Rgionalis", paper peented at the Jan. 1995 UNDP-World Bank Trade Expansion Project confence, World Bank, Washington, D.C. Solomon W. Polachek. 1992. "Conflict and Trade: An Economics Approach to Political Inctions", in W. hard and C.H. Anderton (eds.) Economics of Anns Reduction and the Peace Process. Amsterdam: North Holland, 89-120. .forthcoming. "Why Democracies Cooperate More and Fight Less: Te Relionship Between Intational Trade and Cooperation", Review of Imational Economics. H. Rieben, J. Monnet and G.A. Cbevallaz. 1971. Jean Monnat Centr de Recherches EurKop6ees. Laumsame. Maurice Schiff. 1992. "Social Capital, Labor Mobility, and Welfare: The Impact of Uniting States". Rationality and Society, 4(2), April 1992 :157-75. . 1996a. "Small is Beautiful: Prefentl Trade Agreements and the Impact of Country Size, Market Share, Trade Policy and Smuggling", World Bank Policy Research Workdng Paper Series No. 1668. . 1996b. "Social Capital, Trade and Optimal Migration Policy". Mimeo, EECIT, World Bank. . 1996c. "Socio-Cultumal Capital, Ethnic Diversity and Economic Refom in Sub-Saharan Africa", Revue d'economie du developpement, 3. 36 and Merlinda Ingco. 1996. "Will Chile Benefit from NAFTA? Not Much". Mimeo, IECIT, World Bank. and Claudio Sapelli (eds). 1996. Chile in NAFTA. Free Tade Agreementsversus Unilateal Liberalization. Santiago: Intrnational Center for Economic Growdh (in Spash). Alasdair SmidL 1992. "Measuring the Effects of '1992'", in D. Dyker (ed.) The European Economy. London and New York: Longman. and Anthony Venables. 1988. "Completing the Intenal Mket in the Eumpean Community: Some Industry Simulaions". Europe Economic Review 32 (7). T.N. Srinivasan 1994. "Regional Trading Arrangs and Beyond. Exploring Some Options for South Asia Theory, Empirics and Policy", Report no. IDP-142, South Asia Office of the Chief Economist, World Bank (July). D. Swann 1992. The Economics of the Common Macrt 7th edition, Penguin, Hamondsworzh T. N. Srinivasan 1994. "Regional Trading and Beyond. Explonng Some Options for South Asia. Theory, Empirics and Policy'. ReportNo. IDP-142, Office ofthe Chief Economist, South Asia Region, World Bank (July). L. Alan Wmters. 1993. 'Expanding EC membership and association accords: recent exrience and future prospec", in K. Anderson and R Blaclhurst (eds). Regional Integration and tie Global Trading Systr" Harvester Wbeatsheaf, St. Marns Press. . 1995. 'What can European experience teach Latin America about integration?" Paper prepaed for the Seminar on The Hemispheric and Regional ntgmation in Perpective, Bogota, 1-3 November. Wallace. 1996. 37 Policy Research Working Paper Series Contact Title Author Datd for paper WP81773 The Costs and Benefits of J. Luis Guasch June 1997 J. Troncoso Regulation: Implications for Robert W. Hahn 38606 Developing Countries WP81774 The Demand for Base Money Valeriano F. Garcia June 1997 J. Forguas and the Sustainability of Public 39774 Debt WPS1775 Can High-inflation Developing Martin Ravallion June 1997 P. Sader Countries Escape Absolute Poverty? 33902 WPS1776 From Prices to Incomes: Agricultural John Baffes June 1997 P. Kokila Subsidization Without Protction? Jacob Meorman 33718 WPS1777 Aid, Policies, and Growth Craig Bumslde June 1997 K Labrie David Dollar 31001 WPS1778 How Government Policies Affect Szczepan Figiel June 1997 J. Jacobson the Relationship between Polish Tom Scott 33710 and World Wheat Prices Panos Varangis W11PS1779 Water Allocation Mechanisms: Ariel Dinar June 1997 M. Rigaud Principles and Examples Mark W. Rosegrant 30344 Ruth Meinzen-Dick WPS1780 High-Level Rent-Seeking and Jacqueline Coolidge June 1997 N. Busjest Corruption in Afrcan Regimes: Susan Rose-Ackerman 33997 Theory and Cases WPS1781 Technology Accumulation and Pier Carlo Padoan June 1997 J. Ngaine Diffusion: Is There a Regional 37947 Dimension? WPS1782 Regional Integration and the Prices L Alan Winters June 1997 J. Ngaine of Imports: An Empirical Won Chang 37947 Investigation WPS1783 Trade Policy Options for the Glenn W. Harrison June 1997 J. Ngaine Chilean Govemment A Quantitative Thomas F. Rutherford 37947 Evaluation David G. Tarr WP81784 Analyzing the Sustainability of Fiscal John T. Cuddington June 1997 S. King-Watson Deficits in Developing Countries 31047 WPS1785 The Causes of Govemment and the Simon Commander June 1997 E. Wtte Consequences for Growth and Hamid R. Davoodi 85637 Well-Being Une J. Lee WP81786 The Economics of Customs Unions Constantine Michalopoulos June 1997 M. Patela in the Commonwealth of David Tarr 39515 Independent States Policy Research Working Paper Seribe Contact Title Author Date for paper WPS1787 Trading Arrangements and Diego Pug. June 1907 J. Ngalne Industrial Deopment Anthony J. Venable 37947 WPS1788 An Economic Analysis of Woodfuel Kenneth M. Chomltz June 1997 A Maranon Management In the Sahel: The Case Charle Griffiths 39074 of Chad WP81789 Competition Law In Bulgaria After Bemard Hoekman June 1997 J. Ngalne Central PlannIng Simeon Djankov 37947 WP81790 Interpreting the Coffident of Barry R. Chlawick June 1997 P. Singh Schooling In the Human Capltl 85631 Eamings Function WPSI791 Toward Better Regulation of Private Hemant Shah June 1997 N. Johl Pension Funds 38613 WP81792 Tradeofs ftrom Hedging: Oil Price Sudhakar Satyanarayan June 1997 E. Somensatto Risk in Ecuador 30128 WPS1793 Wage and Pension Pressure on the AJain de Crombrugghe June 1907 M. Jandu Polish Budget 33103 WPS1794 Ownership Structure, Corporate Xiaonian Xu June 1997 J. Chinsen Govemance, and Corporat Yan Wang 34022 Performance: The Case of Chinese Stock Companles WPS1795 What Educational Production Lant Pritchett July 1907 Sheia Fallon Functions Really Show A Positive Deon Fllmer 38009 Theory of Education Spending WP81796 Cents and Sodability Household Deepa Narayan July 1997 Sheila Fallon Income and Social Capital In Rural Lant PrKchett 38000 Tanzania WPS1797 Formal and Informal Regulation Sheoll Pargal July 1997 E. de Castro of Industrial Pollution: Hemnamala Hetflge 89121 Comparative Evidence from Manjula Slngh Indonesia and the United States David Wheeler WPS1798 Poor Areas, Or Only Poor People? Martin Ravallion July 1997 P. Sader Quentin Wodon 33902 WPS1799 More fbr the Poor Is Less for the Jonath B. Gelbach July 1997 S, Fallon Poor The Politics of Targeting Lant H. PrKchett 38009 WPS1800 Single-Equation Estimation of the John Baffel August 1997 P. Koldbs Equilibrium Real Exchange Rate Ibrahim A Elbadawi 33716 Stephen A O'Connell