eP S 17 3 POLICY RESEARCH WORKING PAPER 1573 North-South Customs North-South trade accords can serve as 'credibiity- Unions and International enhancing' mechanisms for Capital M obiliy *the treatment of foreign investment, inducing additional inflows of foreign Eduardo Fernandez-Arias capital. Mark M. Spiegel The presence of sovereign risk changes the tradeoffs between trade: creation and diversion, enhancing the potential for welfare- increasing, trade-diverting North-South regional trade accords. North-South integration involves different issues than did previous trade accords, but a nation still does best by the integration that yields the greatest gains from trade. The World Bank International Economics Department International Finance Division February 1996 | POLIcY RESEARCH WORKING PAPER 1573 Summary findings The primary distinction in a North-South trade accord is impact on investment flows. Fernandez-Arias and Spiegel likely to be that the Southern nation experiences more try to understand the channels through which trade capital scarcity than its Northern trade partner. So the accords can affect North-South investment flows. trade accord's impact on the Southern trading partner's A potential link between trade accords and investment ability to attract capital may have welfare implications flows may be how the accords affect the ability of the for both nations. Southern partner government to make commitments Fernandez-Arias and Spiegel extend the traditional about the treatment of foreign investment. They show analysis of customs unions to allow for international that these accords can affect both the magnitude and capital movements. Their results indicate that trade pattern of inward foreign investment and production, accords may affect the ability of Southern nations to implying the possibility that both trade and financial attract capital and may divert capital between Southern diversion can stem from a bilateral regional trade accord. nations. Novel effects that emerge under sovereign risk must be Moreover, the welfare implications of North-South addressed when assessing the welfare implications of trade accords may differ from those that predict the trade accords. The greatest gains from integration are North American Free Trade Agreement's (NAFTA) still achieved when integration takes place between the minor third-country effects, holding factor endowments countries with the greatest potential gains from trade. constant. But Fernandez-Arias and Spiegel make a distinction: The key implications of North-South trade accords these gains now include both current trade and inter- such as NAFTA are generally perceived to involve their temporal trade through foreign investment. This paper -a product of the International Finance Division, International Economics Department - is part of a larger effort in the department to analyze foreign investment in emerging markets. Copies of the paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Sheilah King-Watson, room N3-040, telephone 202-473-1047, fax 202-522-3277, Internet address skingwatson@'worldbank.org. February 1996. (26 pages) The Policy Research Working Paper Series disseminates the findings 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 fully polished. The papers carry the names of the authors and should be used and cited accordingly. The findings, interpretations, and conclusions are the authors' own and should not be attributed to the World Bank, its Executive Board of Directors, or any of its member countries. Produced by the Policy Research Dissemination Center North-South Customs Unions and International Capital Mobility Eduardo Fernandez-Arias International Economics Departnent The World Bank Mark M Spiegel Department of Economic Research Federal Reserve Bank of San Francisco Helpful comments were received from Jonathan Eaton, Jim Markusen, Ronald McKinnon, Andres Velasco, and seminar participants at the Fed Systems Committee Conference, the FRBSF and CEPR conference, and the University of Colorado, Boulder. The views presented in this paper are those of the authors alone and do not necessarily reflect those of the Federal Reserve Bank of San Francisco, the Board of Governors of the Federal Reserve System, or the World Bank. I. Introduction The recent debate concerning the North American Free Trade Agreement (NAFTA) in both policy circles and the popular press suggested that the primary impact of the regional trade accord would not be a small reduction in already low tariff levels between the NAFTA partners, but a fundamental change in the attractiveness of Mexico as a location for investment, resulting in a large amount of capital movement between the Northern and Southern trading partners. This claim was popularized by such opponents of NAFTA as Ross Perot who claimed that NAFTA would create a "large sucking sound" as capital flowed from North to South in response to the trade accord. Indeed, there is some evidence that "financial diversion" has taken place in response to North-South integration, both in the case of Portugal and Spain in the EEC and Mexico under NAFTA [Primo Braga (1993)]. Despite the consensus that the impact of trade accords on capital movements is potentially important, the issue seems to have received less attention in the standard literature on the welfare implications of customs unions,1 which center on the trade-off between trade creation and trade diversion.2 We should note that this is even true of more recent analyses of trade accords which stress both strategic issues [McLaren (1993), Bagwell and Staiger (1993)] and the political-economic implications of trade accords [Grossman and Helpman (1993)]. The reason for this gap in the literature is probably historical. Prior to NAFTA, successful regional trade 1See Lipsey (1960) and de Melo et al (1993) for extensive early and more recent surveys. 2Notable exceptions include Miyagiwa and Young (1986) who explicitly introduce factor mobility into their analysis of customs unions in a different context, and Manchester and McKibbin (1994) who analyze the implications of an ad hoc decrease in Mexico's risk premium as a result of the NAFTA accord. 1 areas tended to be between Northern countries, such as the EEC and the US-Canada free trade area. As the first major North-South free trade area, the NAFTA raises issues which are not central to North-North trade accords. The primary distinction in a North-South trade accord is likely to be that the Southern nation is physical capital scarce relative to its Northern trade partner. Consequently, the impact of the trade accord on the ability of the Southern trading partner to attract capital may have welfare implications for both Northern and Southern nations. In this paper, we extend the traditional analysis of customs unions to allow for international capital movements. Our results indicate that trade accords may affect the ability of Southern nations to attract capital, and may divert capital between Southern nations. Moreover, the welfare implications of North-South trade accords may differ from those which predict minor third-country impacts of NAFTA holding factor endowments constant [Safadi and Yeats (1993)]. We begin by introducing a model of a bi-lateral free trade accord between a Northern and Southern nation which relies on capital from the North for production in its foreign sector. The model is one where the Southern nation faces a standard timing-inconsistency problem concerning levels of taxation on foreign investment. The Southern nation then commits to national treatment, or equal levels of taxation, towards foreign investments under the trade accord. We then show that the trade accord can be an effective mechanism for achieving enlarged levels of commitment towards foreign investments and therefore can generate additional capital inflows from the North. There is some anecdotal evidence that Southern nations already understand the potential of bi-lateral and multilateral treaties as a mechanism for achieving greater international credibility. Mexico has explicitly committed to national treatment and codified numerous investment regime liberalizations towards its NAFTA 2 partners under the NAFTA accord, even though it had previously underwent a large reform program in 1989 [Hufbauer and Schott (1992)]. Apparently, the reason for repeating these liberalizations under NAFTA was that tying their commitments to the trade accord provided an explicit mechanism for penalizing their violation towards investors from NAFTA partner countries. For example, prior to the NAFTA accord, Hufbauer and Schott (1992) predicted that "Since regulations are more easily changed than laws, the United States and Canada are likely to seek commitments from Mexico in the NAFTA to make regulatory reform more permanent." Along similar lines, Chile offered to unilaterally commit to some liberalizing policies under the Uruguay round of the GATT.3 The ability to use trade accords as credibility-enhancing mechanisms may therefore be an aspect of regional trade accords which may be particularly important in North-South agreements.4 We next examine the implications of a trade accord between a Northern nation and a Southern nation in a 3-country model of foreign direct investment in which Northern investors choose the magnitude of investment in the Southern partner nation and an alternative potential Southern target nation. We demonstrate that the standard analysis of trade-diverting customs unions must be adjusted to account for changes from the union in the risk characteristics of the trading partners. The union enhances the safety of investments in the Southern partner nation, increasing the 3We thank Andres Velasco for providing this example. 4Perroni and Whalley (1993) interpret the same liberalizations as concessions to larger entities in return for insurance against trade wars. These two interpretations do not necessarily conflict; if a mutually beneficial agreement requires these liberalization "concessions," a Southern trade partner may be unable to achieve the agreement in the absence of the greater commitment capacity the accord brings in the context of our analysis below. 3 gains experienced by the Northern partner nation from trading with its Southern partner. This enhances the potential for the regional trade accord to be welfare increasing, even when trade is diverted from a more productive alternative Southern location for production. The remainder of this paper is divided into three sections. Section 2 introduces the basic model and demonstrates how the trade accord can affect the credibility of the Southern partner government and the magnitude of its capital inflows. Section 3 examines the 3-country case with both trade creation and diversion. Section 4 concludes. II. A Simple Two-period Model of Foreign Direct Investment Under a Trade Accord 2.1 Setup In this section, we introduce a simple two-period model of foreign direct investment from a Northern nation to a Southern nation enforced by a bi-lateral trade accord. We assume that there are 2 countries in the accord, which we call the Northern partner and the Southern partner. The Southern partner nation produces a single good, s, and exports that good to the Northern partner. For simplicity, we assume that all output is exported, so that the Southern partner government is only interested in maximizing domestic income. In the absence of a trade accord, the Northern partner charges its "most-favored-nation" tariff of T on all imports of s, including those from the Southern partner. We take the exports of the Northern partner as numeraire, and assume that the Northern partner nation is large relative to the Southern partner nation. The Southern partner enjoys a tariff advantage under the trade accord. Exports originating in the Southern partner nation earn price P(1 +T) (the consumer price inclusive of the tariff) under the accord, and P in its absence. 4 The Southern partner nation is endowed with a domestic sector which produces quantity ss of good s, where s is taken as exogenous. The foreign sector produces quantity sn of good s using its imported foreign capital K , which is determined below, according to the constant returns production function: (1) s = AK where A is a constant. The Southern partner charges a tax on the output of domestic and foreign owners of capital, and is assumed to have made a "pledge" of national treatment towards foreign investment. Formally, we model the implications of this pledge as a commitment to tax domestic and foreign output at the same rate, which we designate as t. In the absence of a trade accord, the only mechanism which binds this commitment is the value that the Southern partner government places on its "goodwill" in the international community.5 We model the penalty for loss of goodwill as a proportion of the value of the economic resources of the economy. Since the importance of goodwill is uncertain to investors, and likely to be affected by external parameters, we specify this proportion as a stochastic function, 02, where 0 is an exogenous positive constant fraction and o is distributed according to the density function f(n) along the interval [0,11,6 f'>O. Note that since f(Q) is invariant with respect 5National treatment is assumed only for concreteness; the results hold true for any given tax pledge t. Here we assume that the nationality of foreign investors cannot be observed and, therefore, the loss of "goodwill" applies equally to all foreign investment and all Northern investors equally benefit from the accord. In practice, some degree of nationality discrimination favoring firms associated with the Northern partner nation is likely to be feasible. 6The restriction to the unit interval is done for notational simplicity and without 5 to t, a Southern partner nation choosing to abrogate national treatment will choose to fully tax the proceeds of the Northern subsidiary.7 The strategic interaction between the North and the South has the following extensive form: In the first stage, Northern investors choose the magnitude of capital, Kn, to extend to the Southern partner. In the second stage, Q is realized and the penalty faced by the Southern partner for violating her pledge of national treatment to foreign investors is determined. In the third stage, the Southern partner decides whether to comply with its pledge of national treatment or to tax foreign investment at a higher rate. We assume that all government revenues are distributed lump-sum to domestic agents. 2.2 Southern Partner Government Taxation Decision. To insure sub-game perfection, we begin with the stage three decision of the Southern partner concerning compliance with national treatment. The Southern partner government is assumed to levy a domestic tax of rate t on output (t 0. (7) ae/aKn = -f(n )(1+i)A(I-t)s /O(s +s)2 < O. 2.3 Northern Agents' Investment Decision. The Northern investors are assumed to be risk neutral and have access to international capital markets at rate r. We assume that the investment of the representative owner of foreign capital in the Southern partner nation is sufficiently small that its liability associated with this investment is not limited. Since individual foreign investors are small relative to the Southern partner nation, they also take o as given when making their investment decision. The zero-profit condition then satisfies: 8 (8) e(1-t)P(1+T)A = r The above expression implicitly defines the equilibrium solution Kn(i). As a notational convenience, we can express the price earned by the Southern partner nation as P(1 +T), where in the absence of the accord T=O and under the accord T=T. Substituting Tr for T in (6), (7), and (8) and totally differentiating yields: dKn 0/(1+T) + (ao/8l) > O dT aol a 8Kn There are two independent channels, which are captured by the numerator of (9), through which an increase in T serves to attract foreign capital. The first term reflects the terms of trade improvement experienced by the Southern partner nation. The second term reflects the decrease in the probability of expropriation due to the enhanced potential penalty associated with the price effect of losing the trade accord. The greater are the price benefits of the accord to the Southern partner nation as measured by T, therefore, the greater is the increase in the penalty for expropriation and hence the increase in the magnitude of capital inflows. 2.4 Impact on Southern Partner Income While the small country assumption maintained in this section precludes any welfare implications for the North, the current setting provides a good overview of the income implications of the trade accord for the Southern Partner nation. Define E(U) as the expected income earned by the Southern partner nation. E(U) satisfies: Q (10) E(U) = oW + IWf()dQ. Invoking the envelope theorem, differentiating E(U) with respect to -r 9 yields: * aE(U) nK (11) 8 = Pe(ts +sSj + [OP(J+T) + f (1-O2)f(Q)d2]d a-r n1 fo dr Equation (11) identifies two channels through which the welfare of the Southern partner nation is affected by the trade accord: The first term reflects the direct positive increase in earnings due to the tariff advantage. The second term reflects the increase in expected revenue from the foreign sector due to the increase in foreign investment, defined by equation (9). Both of these terms are positive. It is interesting to note the characteristics of the Southern partner nation which would cause it to experience greater income gains from the trade accord. First, the increase in income will be greater the greater is the magnitude of trade with the Northern partner nation. This stems both from the direct price effect on Southern exports and from the impact of the trade accord on the magnitude of capital inflows the Southern partner nation can attract, as shown in (9). Second, the Southern partner will experience greater benefits from the trade accord the greater is the initial safety of the Southern partner nation. Since increasing T enhances the payoff under compliance with national treatment and fails to affect the payoff under violation, the impact of a change in the initial safety of the Southern partner nation will be a greater increase in the magnitude of capital inflows. III. Extension to Two Potential Target nations 3.1 Setup In this section, we extend the analysis to allow the trade accord to affect the pattern of investment across the South. This requires us to treat the Southern partner country as "large" in terms of having market power with respect to the price of its export good. To accomplish this in as simple a manner as possible, we assume 10 that there are two potential locations of production: Country A, the Southern partner nation, and country B, an alternative location, that can be thought as the rest of the South. We assume that the output of the two nations are perfect substitutes and maintain the assumption that individual northern firns behave competitively, despite the fact that their country has market power in aggregate. We also maintain the above assumption that all output is exported to the Northern partner nation, so that exports from country A earn price P(1+T) while exports from country B earn price P. Let reflect the output of the foreign sector of country Z (Z=A,B). We assume that the production function for SZ satisfies: (12) Sz = ZKe. (Z=A,B; z=a,b). where a and b are constants representing the productivity of the Southern target countries A and B respectively, while Kn is the magnitude of foreign capital n attracted by country Z. Let reflect the total output of exportables in country Z, Sz=Sz+ where s4 represents the output of country Z's domestic sector and let s represent aggregate industry output, s=sA+sB. The demand for the output of the Southern nations is assumed to be a downward-sloping function of output. Market demand satisfies P(1+T)=D(s), where D'S50. As in equation (4), the zero-profit conditions for foreign investments in countries A and B with country A in the trade accord satisfy: (13a) eA (1-tA)P(l+)a = r. (13b) OB (1-tB)Pb = r In addition to the producer terms of trade, P, there are four determinants of the magnitude of foreign investment which will be attracted by the target countries. The first is country risk, determined by eA and 0B, the probabilities that 11 countries A and B will comply with national treatment. The second is tA and tB, the tax rates faced in the two countries given compliance. The third determinant is the relative productivity in countries A and B, determined by parameters a and b. The last determinant is the magnitude of the terms of trade advantage enjoyed by the Southern partner nation, T. 3.2 Target nation taxation decision. We again assume that both Southern nations have committed to national treatment towards foreign investors. We also assume that each nation faces a penalty of losing a OZQ share of its output for violation of national treatment where z is an exogenous constant fraction and 0 is defined as above.9 Let WB represent the payoff to country B from choosing compliance with national treatment. WB satisfies: (14) p (tB + =s). Alternatively, let WB represent the payoff to country B from choosing complete taxation, which satisfies: (15) 4B = p(s! + sB( B ). Let QB represent the realization of 2 which leaves country B indifferent between compliance and complete taxation (W B=W4). OB satisfies: (16) Q =- 9Allowing the countries to have different distributions of Q would produce qualitatively the same results and would only be relevant in terms of diversification potential, which we do not focus on in this paper. 12 When making its taxation decision, country A must consider the additional loss from losing the benefits of the trade accord. Let WA represent the payoff from complete compliance with national treatment. W4 satisfies: (17) W4 = T+ s Alternatively, let ;V4 represent the payoff to country A from choosing complete taxation, which satisfies: (18) Vl = p(sA + sA )(J _ 0A0) Let OA* represent the realization of n which leaves country A indifferent between compliance and complete taxation (W=A4). QA satisfies: (19) D = A [I - (1+T)(tS +S)/(s +<) ] As noted in the previous section, a positive probability of full taxation requires a parameter restriction.10 To motivate this parameter restriction, we make the assumption that under national treatment investment in countries A and B would earn positive profits in the absence of country risk. This is ensured by the constraint: (20) (1 tZ)P 2 r, (Z=A,B; z=a,b). oA and 8B can then be found by substituting QA *and QB respectively for Q in equation (5) above. Totally differentiating with respect to KAand KByields: 10Altematively, we could have generalized the analysis to include zero foreign investment in the form of corner solutions. 13 (21) 8 = - f(4*)(1+i)a(J-tA)sA/,1A(s4;s+4)2 < o aK4 and: (22) Be = f (B*)b(l-tB)4/1B(s 5s)2 < O. n As above, we express the price earned by the Southern partner nation as P(I +-r), where in the absence of the accord T= 0 and under the accord T=T. Substituting T for T and totally differentiating yields: (23a) eT f(/)(J/p A ) > 0. a-c aeB (23b) = O. 3.3 Implications for International Capital Flows We can again examine the impact of the trade accord by conducting comparative static exercises concerning the implications of an increase in T. We show in the appendix that the comparative static results satisfy: dKA [ B] (24a) [- jl ± T ]c |89P2eB (24b) dn A aoA A 8[ (l+raBBbPlPA < 0. where A is the determinant of the matrix of the system shown to be positive in the appendix and P'=8P/8s50d The comparative static exercises demonstrate that the impact of the trade accord is an increase in investment in the Southern partner nation and a decrease in investment in the other Southern nation. The magnitude of this impact for both 14 nations is increasing in OA, the probability of compliance by country A with national treatment, and aeA/aT, the increase of this probability due to increases its terms of trade advantage under the trade accord. Consequently, the trade accord, by enhancing Southern partner credibility through granting it a tariff advantage, generally increases investment in the Southern partner nation at the expense of the Southern non-partner nation. It is easy to show by contradiction that the expansion in output in country A must be larger than the contraction in output in country B, resulting in a price decline. Assume that P does not decline. Equations (lOa) and (lOb) imply that in equilibrium both countries would be riskier (oZ would be smaller). This is possible only if investment increases in both countries, which implies higher output and a decline in price. More formally, recall that the change in output with respect to -r in country Z (Z=A,B) satisfies zaKZ/la. The relative output change in the two countries satisfies: b(dK4/dT) abeBPI a (dKO/dt) abeBP, + aP(a B/aK1B) which demonstrates that aggregate output increases. However, the impact on aggregate foreign investment is ambiguous depending on the relative productivity of the Southern partner. By equations (21a) and (21b): dKB/dT B (26) - -d-/d - =boBP, + P(aeB/aKB). Examining equation (26), if asb (the Southern partner is less efficient than the rest of the South), the increased output must be supported by increased investment. If a>b, however, the superior productivity of the Southern partner may lead to smaller aggregate foreign investment from the North. In this case, while Southern aggregate output expands, investment from the rest of the South is diverted more than fully. 15 The potential for investment diversion will also depend on the relative size of the Southern partner nation to other potential Southern locations for investment. To see this, let A represent the relative size of the Southern partner nation within the South, x =sB1(sA +sB), and £ represent the elasticity of Northern demand, £=-D'sID. Substituting into (24) then yields: dKBI/dT aeB (27) = dKA/dT beBeA + m. where m=- (aEB/aKe)(P2/sA) > 0 and bounded. If the Southern partner nation represented the entire South, A converges to 0 and the offsetting effects vanish. For more moderate values of A, the results depend on the value eX. Smaller aggregate investment would result if the Southern partner is more productive (a>b) and eA is large enough (i.e. demand is highly elastic and/or the Southern partner is relatively small within the South). An interesting extreme case is obtained when the Southern partner is extremely small within the South (i.e. A is close to 1). In that case it is easy to obtain that in the limit equation (21a) converges to equation (9). In general, the expressions obtained in the previous section correspond to the limiting cases A = 1, or the Southern partner being "small" within the South. 3.4 Welfare Analysis 3.4.1 Impact on Northern Partner Welfare Finally, we turn to the question of the welfare implications of the trade accord. Since citizens of countries A and B do not consume the output of the industry and production is competitive with the world rate of interest r taken as given, there are two channels for the trade accord to affect welfare in this simple model: By affecting consumer surplus in the Northern partner nation, and by affecting the 16 expected magnitude and distribution of Southern income. We begin by analyzing the welfare impact of the trade accord on the Northern partner nation. As we showed above, the existence of the trade accord will result in a diversion of investment from country B to country A. See Figure 1. Let PO and Pl represent the equilibrium consumer price in the absence and presence of the trade accord respectively. The trade accord achieves a reduction in the price paid by consumers in the Northern trade partner from P0 to P1. However, only imports of the B non-partner nation, represented in the graph by sI, are subject to tariff. Assuming that all tariff revenues are repatriated to consumers, then, the net impact of the trade accord on Northern partner consumer surplus is equal to the area of region III (reflecting the mitigation of the trade distortion), plus area V (reflecting the B price reduction in the portion of trade subject to tariffs, s1) minus area IV (reflecting the reduction in tariff revenue in the portion not subject to tariffs, in excess of the price decline). We analyze the impact of the trade accord formally by positing that the accord yields an advantage of terms of trade advantage of T, where OT5-Tr, and then integrating the implications of changes in T from 0 to T. Let AWN represent the resulting welfare change in the Northern partner nation. We show in the appendix that AWN satisfies: (28) AWN [ + a(I ] psA]dT where 1P is defined in the appendix. T>0 is a necessary, but not sufficient, condition for AWN>O. AP is of ambiguous sign due to the loss in tariff revenue from country B resulting from trade diversion under the trade accord. This loss is increasing in b and oB, two terms which reflect the attractiveness of country B as a location for 17 production. Given that ' is positive, which necessarily holds when the Northern partner nation benefits from the accord, (26) demonstrates that AWN is increasing in both A and (aoA/a-). Integration with country A is more likely to be welfare enhancing for the Northern partner country the greater the security of investments in the Southern partner nation, and the greater the positive impact of the trade accord on the riskiness of investments in that country. The intuition behind this result follows from the traditional customs union analysis of the trade-off between trade creation and trade diversion. Lowering the sovereign risk of your trading partner enhances your gains from trade with that trading partner. In the traditional literature, the welfare analysis of trade- diverting customs unions centers on choosing the proper customs union partner, i.e. that with which you will enjoy the greatest gains from trade. The trade accord, by enhancing the environment for foreign investment in country A, actually enhances the gains from trade with that country. Moreover, the degree to which the trade accord enhances the penalty faced by the Southern partner nation for non-compliance with national treatment is directly increasing in the gains from trade between the Northern and Southern partner. Consequently, the impact of the accord on sovereign risk provides an independent reason that the potential for welfare gains are increasing in the gains from trade between the two partners. 3.7.2 Impact on Southern income. The trade accord will affect the income of the two Southern nations through its impact on the value of the output the domestic sector of those nations and through its impact on tax revenues generated by foreign direct investment. Note that expected revenues must include payoffs in states where expropriation takes place. 18 Define E(LF) as the expected income earned by country Z (Z=A,B). E(UZ) satisfies: (29) E(Z)= oZvZ + f WZf(2)do; (Z=A,B). Invoking the envelope theorem, differentiating E(U4) with respect to T yields: aE(UO) A (assA+sA) + + dKny + A1 (30) = P[(- p) + a d[oAl+) + wA where: *Z (31) Z = f (l1-,OZ)f(Q)dQ, (Z=A,B). Equation (30) identifies three channels through which the welfare of the Southern partner nation is affected by the trade accord: The first term reflects the direct positive increase in earnings due to the tariff advantage. The second term reflects the decline in terms of trade suffered by the Southern partner nation due to the increase in supply. The third term reflects the increase in expected revenue from the foreign sector due to the increase in foreign investment. While the first and third terms are positive, the net impact on the Northern partner nation is ambiguous because of the negative terms of trade effect captured by the second term. Similarly, differentiating E(W3) with respect to r yields: (32) ~~~~aE(LJB) asEO dI4 (32) = p (a + b d [BtB + B] < o. The components of (32) are analogous to the final two components of (30). The first term reflects the decline in terms of trade suffered by country B due to an increase in supply from the South, while the second term reflects the decrease in country B's expected revenue from its foreign sector, due to the decrease in its foreign 19 investment. These two effects are both negative, leading to the unambiguous result that the trade accord reduces the welfare of the alternative Southern nation. With the possible exception of a large drop in market price due to the erosion of monopoly power, then, the accord favors the Southern partner nation and harms the other Southern nation. The impact on the South as a whole is therefore ambiguous. Combining equations (30) and (32) and simplifying yields: jj ° Solving: (S 3)d = p0A + (I+z)][SP +eBbP ]/E (A.4) n = aoBPPA + _(1+T)l/T