Presidential Diplomacy and the Institutional Underpinnings of MERCOSUR: An Empirical Examination

The relative success of the Southern Common Market (MERCOSUR) is a puzzle for most theories of regional integration. This is due to its having achieved remarkable progress in spite of lacking features such as significant levels of previous interdependence (demand factor) or major regional institutions (supply factor). To account for this puzzle, it has been claimed that the operation of MERCOSUR rests on presidential diplomacy. Such a mechanism is understood as the resort to direct negotiations between the national presidents whenever a crucial decision has to be made or a critical conflict solved. This article argues that presidential diplomacy—understood as political, summit diplomacy as opposed to institutionalized, professional diplomacy—is insufficient to account for the performance of MERCOSUR. Through the empirical analysis of three critical episodes, the article shows how institutional structures, shaped by the system of government of the member countries, have sustained presidential intervention and, hence, the process of regional integration.


INTRODUCTION
The experience of the Southern Common Market (MERCOSUR) does not fit mainstream theories of regional integration. The two major contemporary currents, namely liberal intergovernmentalism (Moravcsik 1998) and supranational governance , regard society as the point of departure for integration, as transnational transactors increase their exchanges and subsequently call on national chief executives of non-presidential regimes-such as prime ministers in parliamentary democracies (Danese 2001). Supposedly, in MERCOSUR, the successive Argentine and Brazilian presidents since 1985 have carried out presidential diplomacy with a free hand to accommodate the issues at stake in the way they saw fit. This article reconsiders the "presidential diplomacy" argument on the grounds that, while not incorrect, it does not provide a truly accurate explanation of proceedings and results of MERCOSUR. Through an empirical analysis of three critical episodes of MERCOSUR history, the article shows how institutional incentives and constraints, shaped by the presidential format of the member countries, have supported the process of regional integration.
Most presidential regimes in Latin America have been described as asymmetric, unbalanced, hyperpresidential or concentrationist when compared to the U.S. case (Mainwaring and Shugart 1997;Nino 1992;Shugart and Carey 1992;von Mettenheim 1997). This article advances the hypothesis that, in MERCOSUR, the member states' presidents concentrate enough power as to either prevail over or circumvent other veto players such as the cabinet 2 and congress in a way that is impossible for either parliamentary executives or balanced presidential systems. It examines the role of presidents as both policy crafters and dispute settlers in order to assess the real magnitude of the role played by chief executives in key policy areas. The institutional features that shape presidential intervention are identified and evaluated, with a view to setting forth an account of Southern Cone regional integration that explains its unexpected singularities.
After a brief introduction that discusses the genesis and operation of MERCOSUR, the article focuses on three events that have marked its short history. These include two controversies over the special regimes that were established at the beginning of the process and which are still in force for automobiles and sugar. The third is the outburst and management of one of the most serious crises the bloc has faced to date, triggered by the Brazilian devaluation of 1999. The methodology draws partly on the supranational governance approach, as it is especially sensitive to different policy areas and to day-to-day politics. At the same time, it draws on the intergovernmental approach to preference formation, for the hypothesis tested here challenges the liberal assumption that states act as agents of domestic social actors.
2. In a presidential system, the cabinet is not formally a veto player. However, the fact that most presidential regimes are governed by party coalitions means that partisan veto players may shift from the legislature to the cabinet, making it an additional veto point.

A BRIEF OVERVIEW
MERCOSUR is presently an incomplete customs union that aims to become a common market. It was created on March 26, 1991, when previous agreements between Argentina and Brazil were expanded in order to include Paraguay and Uruguay. Although it has delivered less than it originally promised, the bloc has tripled intraregional trade flows and has become an acknowledged actor in world trade and a consistent target of foreign investment. MERCOSUR is one of the most advanced regional groupings in terms of formal integration. 3 All these achievements notwithstanding, it has not developed any kind of autonomous regional authority and is ruled exclusively by intergovernmental bodies made up of representatives of the member governments. The highest formal authority is the Common Market Council (CMC), integrated by the foreign and economy ministers of each country. According to former Argentine Foreign Minister Dante Caputo, there was an early decision not to institutionalize the integration process until it had reached a safe "cruising altitude" (Caputo 1999). The structure of MERCOSUR contrasts sharply with that of the European Union, as the latter has created strong institutions since its inception in 1951. Moreover, three of its highest bodies-the European Parliament, the Court of Justice and the Commission-are of a supranational nature.
Both the embryonic (1985)(1986)(1987)(1988)(1989)(1990)(1991) and the postfoundational stage (from 1991 onwards) of MERCOSUR have been characterized by a remarkable degree of presidential activism. Argentina's Raúl Alfonsín and Brazil's José Sarney took the first, unexpected step toward creating MERCOSUR in 1985, when both countries were inaugurating their democratic regimes and there was no social demand for a rapprochement between the two, historically aloof, neighbors. In 1991, Argentina's Carlos Menem and Brazil's Fernando Collor de Mello went a step further, inviting Paraguay and Uruguay to join the venture. The four countries signed the Treaty of Asunción, which was the birth certificate of MERCOSUR. Shortly thereafter, an additional agreement was signed: the Protocol of Brasília, as it was called, established an ad hoc-as opposed to standing-dispute-settlement mechanism. Some years later, in 1994, Menem and Brazil's new president, Itamar Franco, signed the Protocol of Ouro Preto, consolidating the formal structure of MERCOSUR, which has been kept virtually untouched since then. After 1995, it was Fernando Henrique Cardoso, the new Brazilian president, who took over the command of MERCOSUR, first with Menem and later with Fernando de la Rúa. The presidents of the smaller states would play a lesser role, although their intervention has always been greater than any other actors' except that of the presidents of the bigger states.
The presidents have exhibited significant autonomy in decision making regarding MERCOSUR affairs. Nonetheless, their decision to embark on an integration process and their performance in this area have been shaped by institutional incentives and some (albeit few) constraints. As striking as it may seem, the presidents adopted decisions that appeared to compromise national sovereignty without facing serious vetoes from other relevant actors. This was due mainly to two factors: first, a high degree of concentration of power in the hands of the chief executive enabled them to overcome or circumvent potential veto players (Cheibub and Limongi 2002;Figueiredo and Limongi 2000;Malamud 2001); second, foreign policy typically offers chief executives greater room to maneuver than domestic politics do (Rogowski 1999;Schlesinger 1974;Silva 1989). Some recent works analyze thoroughly the impact of the type of democracy on policy outcomes, thus paying greater attention to the link between presidential systems and public policy (Eaton 2000;Haggard and McCubbins 2001); however, they do not focus on foreign policy but rather mainly on budgetary and regulatory policies. Academics such as Milner (1993Milner ( , 1997 have studied the differential influence of presidential and parliamentary systems on foreign policy, but rarely have any presidential democracies outside the United States been included in the analyses. This article contributes to filling this lacuna by analyzing how a specific brand of presidentialism has dealt with a specific foreign policy matter, namely regional integration. To do so, it makes the case that the presidents of MERCOSUR member states have concentrated power and enjoyed high levels of autonomy, mainly because of institutional factors, which includes decisional autonomy from the assembly and the cabinet, legislative initiative, decree power, and veto power. The following case studies provide empirical support for this hypothesis.

THE SPECIAL REGIME: AUTOMOBILES
There are many reasons why cars have become a central issue for MERCOSUR. In the first place, global competition everywhere is undermining the national bases of the automobile industry. However, it would be improper to speak of an integrated world automobile market, as the sector is organized into regional clusters rather than on a global level (Bastos Tigre et al. 1999). Second, cars are believed to be MERCOSUR's equivalent to coal and steel for the early European Community. In other words, it is the sector in which intra-industrial complementation has developed most because it is a highly dynamic sector where expanding investment and labor markets are concerned. In the third place, the automobile regime was established at the beginning of the integration process and acquired a special status as the only sector that persisted despite all the subsequent changes. As key Argentine negotiator Roberto Lavagna pointed out, this is "the only sectoral agreement that survived the methodological change of 1990, thus becoming a successful example of productive expansion and technological modernization with a reasonable balance between its parts" ( Lavagna 1999, 7-8). According to the same analyst, the reason for this outcome was that "the auto sector has had enough political strength to prevent the dismantling of the sectoral project" (Lavagna 2000). Whether or not these sectoral interests had the political strength that Lavagna claims, the decision to continue with the sectoral agreements in this area was not politically contested. Instead, there was a joint decision between Argentina and Brazil to go on with this strategy, which was heralded, defended, and rescued-when in jeopardy-by the national executives.
The foundation that fostered integration between the national auto industries was established by Protocol 21 of the Economic Complementation Agreement no. 14, which Argentina and Brazil signed at the end of 1990. It was explicitly defined as provisional, and its terms were supposed to be renegotiated a few years later. Indeed, Protocol 21 was eliminated four years later, when the Ouro Preto Protocol consolidated the customs union when it was signed in December 1994. In its place, CMC Decision 29/94 created an ad hoc technical committee that was requested to elaborate, before June 1, 1995, a proposal for a common auto regime that was to enter into force on January 1, 2000. Meanwhile, in addition to Decision 29/94, a bilateral agreement between Argentina and Brazil reciprocally acknowledged the national regimes still in force in both countries. 4 Both partners also agreed not to introduce unilateral restrictions on trade within the Free Trade Zone. From June 1995 onwards, bilateral agreements could be modified only to increase intra-regional trade and never to bestow advantages on the respective industries with the purpose of attracting investment. This was to change dramatically shortly afterwards, though.

The Remaking of the Automobile Regime and the 1995 Crisis
Between 1990 and 1996, there was a sharp increase in intraregional trade in cars as a consequence of tariff reductions and agreements to balance imports and exports. 5 However, by mid-1995 Brazil was suffering from economic distress as a result of the Mexican crisis. Argentina was facing economic turmoil as well, although its public and financial sectors were more solid than Brazil's. Automobile production remained stable in Argentina essentially because of its exports to Brazil, which generated a growing imbalance and raised bitter complaints among Brazilian producers. The government was expected to take protective measures, but these were not supposed to apply within MERCOSUR but to be addressed toward extra-regional trade.
The crisis finally came to a head on June 13, when Brazil announced that it would limit automobile imports during the second half of the year to 50 percent of the total of imports of the first half of the year. Surprisingly, Provisional Measure 2410 considered no exception regarding MERCOSUR partners. 6 This fact angered the Argentine authorities. The Brazilian presidential spokesman, Sergio Amaral, communicated the decision to two Argentine negotiators, Jorge Campbell and Carlos Magariños, who had flown to Brasília. However, the Brazilian cabinet was divided about the decision, as Finance Minister Pedro Malan opposed trade restrictions, whereas Planning Minister José Serra supported them. Both were close aides to the president, so he usually mediated between them, but this time the hard-liners carried the day.
The following day, Argentine papers covered their front pages with the alarming news: "Brazil Limits Car Imports from Argentina." 7 Actually, although the quotas were not limited to Argentina, the issue became highly sensitive and threatened MERCOSUR. Argentine officials and producers considered Cardoso's decree a violation of the Protocol of Ouro Preto, which guaranteed free trade among all MERCOSUR partners. However, this protocol also set up the requisite of balanced trade, so the Brazilian government justified the measure by pointing to the lack of equilibrium between imports and exports.
The conflict was one of the most serious that MERCOSUR had faced thus far. Furthermore, it was exacerbated by the proximity of the MERCOSUR Economic Forum, a meeting organized by the World Economic Forum, which was scheduled to take place later that week, from June 18 to 20, in São Paulo. On June 14, Menem threatened to boycott the most expectantly awaited event of the forum, the presidential 5. It should be noted that, apart from free trade, there were other reasons contributing to this surprising performance, including parallel macroeconomic developments and economic recovery (Bastos Tigre et al. 1999).
6. Provisional measure 2410 established a restriction to Brazilian imports regardless of the country of origin, thus making no difference between Mercosur members and non-members.
7. "Brasil limitó la importación de autos argentinos," Clarín, June 14, 1995. summit. Menem's threat engendered feverish negotiations aimed at preventing an incident that would have such a negative effect on the reputation of MERCOSUR. Faced with this strong stance, Cardoso ultimately decided to postpone the enforcement of the decree for thirty days. It is somewhat puzzling that either the presidents themselves or second-level officials carried out the negotiations, whereas top cabinet ministers were left out. Argentine Foreign Minister Guido Di Tella was stuck in Rome due to an airlines strike, and his deputy, Fernando Petrella, did not take part in the process either. Only the number three at the foreign ministry, Jorge Campbell, got involved. The presidents decided to hold direct conversations in order to work out a solution. As put by an Argentine journalist, "notwithstanding Di Tella's transportation troubles, the concrete thing is that the crisis was personally handled by Menem, who consulted Economy Minister Domingo Cavallo rather than the foreign ministry." 8 According to public sources, Cardoso himself called Menem at 5 p.m. on June 15 and sent him a personal letter at 8 p.m. This time, direct presidential intervention was not limited to the main countries: Paraguay's Juan Carlos Wasmosy and Uruguay's Julio Sanguinetti also offered themselves as personal mediators in the dispute between their larger partners (Sanguinetti 2001).
Over the weekend, Menem responded to Cardoso with a letter in which he accepted the invitation to attend the regional meeting. It was formally agreed that all four presidents would meet and that Menem and Cardoso would hold a bilateral summit thereafter. In spite of the provisional accord, Brazil ratified the validity of the restrictions but accepted that it should negotiate them with Argentina. Strangely enough, neither partner ever mentioned the possibility of appealing to the Protocol of Brasília, the institutional instrument by which MERCOSUR partners are supposed to resolve their disputes.
Shortly afterwards, the Brazilian administration itself was caught up in the controversy. The official line was that the establishment of a national automobile regime had to be implemented prior to any regional agreement in the area. Therefore, cabinet hawks criticized the negotiators who had consented to establish a regional regime at Ouro Preto when only one of the main countries involved, Argentina, already had a national regime. On Monday 19, Menem and Cardoso finally held the bilateral summit but only agreed to sign a declaration stating that negotiations were to continue. The discussion was leaked to the public, and neither Wasmosy nor Sanguinetti were able to prevent it. Cardoso explicitly asked Menem to act "politically, in order to restore a balance between the parts. The most important thing here is political will." Again, 8. Clarín, June 16, 1995, 5. no mention of established institutional procedures was made. In turn, Menem replied: "we can negotiate, but only within the frame of the signed agreements that cannot be changed-as in Argentina they have been incorporated into national legislation." 9 The Argentine discourse was more institutional, whereas the Brazilian focused on the issue of political will.
The following day, the situation took a dramatic turn. Menem and Cavallo admitted that the negotiations of 1994 had left Brazil at a disadvantage and toned down their accusations. Meanwhile, a heated discussion arose within the Argentine cabinet. Di Tella and Cavallo blamed each other for the situation, although Cavallo later grudgingly acknowledged the reasonable nature of Brazilian claims when, during the bilateral negotiation, he was informed that MERCOSUR technical committees had never met to adjust and enforce the automobile agreement. President Cardoso finally promised there would be a negotiated settlement of the dispute and suspended the application of the quotas. Subsequently, a comprehensive agreement was reached: Brazil would not apply quotas to Argentina for the remainder of 1995, and the two countries would begin negotiations to establish a definitive common regime that would last until 2000 (Cason 2000).

Circumventing Regional Institutions: Interpresidential Dynamics
Both Argentina and Brazil considered the automobile issue highly sensitive because it was about much more than just trade. As the Argentine industry secretary Carlos Magariños noted, "the real dispute is over investments" (Cason 2000, 31). Therefore, it was not only trade balance but also industrial planning and even macroeconomic stability that were at stake. This fact explains why national governments were so sensitive to the emerging conflict and so eager to solve it promptly. What it does not explain, however, is why the presidents got involved when this was an issue that lesser diplomatic or technical public officials and domestic bureaucracies could have managed in a technical way.
The automobile sector in MERCOSUR was built around a few large transnational companies that did not coordinate with one another. They designed their strategies with an eye on the integration process but got used to dealing directly with national governments from the outset. For their part, governments were receptive to the demands of companies and followed company preferences, as manifested by the maintenance of the automobile regime. As a result, the companies were used to getting what they wanted: direct links with decision makers and active policies in their favor. The governments were also satisfied, as they 9. Clarín, July 20, 1995, 2-3. wanted to attract investment and create jobs. Brazilian Provisional Measure 2410 disrupted the cross-border alignment between governments and companies, splitting up interests along national lines.
The automobile market was anything but free. Government intervention guaranteed not only positive incentives but also protective measures against extra-and intraregional competition. This intervention was not only accepted but also actively demanded by auto firms: as a Brazilian minister put it, "the auto firms love quotas" (as cited, unnamed, in Cason 2000, 30). On the Argentine side Oscar Salvi, director of Toyota Argentina, defended the protective measures, stating that "the traditional policies of Brazil . . . have been, if not imperialistic, at least overwhelming" (Cason 2000, 31).
The strategies of the car makers were therefore not market-oriented but rather government-oriented. The Brazilian government was well aware of this, and it feared that existing trade regulations for the auto sector would influence transnational investments in MERCOSUR to the benefit of Argentina. On the other hand, many in Argentina feared that the country would have become a "province of Brazil" had the situation not been reverted, as bluntly stated by well-known political journalist Mariano Grondona. 10 Paradoxically, the only surviving sector of the original integration project was deeply embedded in domestic, nationalist sentiment.
The dynamic of conflict within the automobile sector was the result of the interests and strategies of the actors involved. In almost every episode, the preferred alternative was to delay fulfilling treaties and postpone negotiations to achieve a permanent settlement. Likewise, the favored course of action was always based on informal, bilateral (i.e., Argentina-Brazil) negotiations and never on the request for formal procedures. Whenever the possibility of resorting to the mechanisms established by the Protocol of Brasília or the Protocol of Ouro Preto arose, businessmen insisted that it was impossible to respect them "for the time being, at least" (Sajem 1999, 79). The outcome of the 1995 crisis in the auto sector was the same as that of all subsequent conflicts over the sector: the maintenance of the status quo. Neither buildup nor spillover turned out, 11 but for most observers the preservation of MERCOSUR was a good enough result, given the region's disappointing experience with integration and the magnitude of the conflict.
With MERCOSUR in place for only four years, few people were ready to consider its institutional performance satisfactory. Fernando Henrique Cardoso himself strongly criticized some of its features: We, Brazilians, invented the theory of a non-institutional MERCOSUR, based on the direct relation between the governments and, fundamentally, the presidents. Whenever there is a crisis, Menem, Wasmosy, Sanguinetti and myself have a talk. We even decide on minor problems . . . Nonsense. Presidents should not get involved in such things . . . I even had to solve a problem regarding automobile quotas with Menem in São Paulo! (Cardoso and Toledo 1998, 126).
What Cardoso calls a Brazilian-crafted theory and then calls mistaken, is nothing less than the way MERCOSUR has worked since its inception. "Inter-presidentialism," a term coined to identify this governing style (Malamud 2003), kept MERCOSUR ticking. Because of their institutional capabilities and the historical preeminence of the presidency, the presidents were targeted as the only possible suppliers of decisions, enforcement, and dispute resolution. Hence, every demand-from the most significant to the most trivial, arising from national producers or foreign governments-was channeled directly through the presidents. When applied to the 1995 auto crisis, this procedure proved to be predictable and effective.

THE SPECIAL REGIME: SUGAR
The sugar regime is a prime case for gauging presidential intervention in MERCOSUR, for the hypothesis advanced here faces a direct challenge. The reason is that a number of authors have argued that national parliaments (Vigevani et al. 2001), the joint parliamentary commission (Mustapic and Geneyro 2000), or domestic social actors (Lucca 2001;Romero 2001;Vigevani et al. 2001) in MERCOSUR have played an important role in shaping this regime. Consequently, the analysis of the relative weight of each actor-executives, legislatures, and social actorsin shaping the final outcome is likely to feed controversy and, in so doing, stimulate further research on the working of MERCOSUR.

The Sugar Regime: Nature and Conflicts
On June 30, 2000, a presidential communiqué issued in the wake of a CMC meeting held in Buenos Aires advocated the integration of sugar into the free trade zone and the common tariff. Concrete proposals were notably absent, however. Between 1985 and 2000, no improvement in cooperation or integration had been achieved in this area. Stagnation was evident; its causes were not.
Sugar has been a hot issue since the beginning of MERCOSUR. Argentina and Brazil had, and still have, opposing interests resulting from a high productivity differential. The differential is not due to different uses of technology, which are similar in both countries, but to the indirect subsidies that Brazil allocates to the production of sugar cane. 12 This policy 12. It is also true, however, that economies of scale, transportation infrastructure, and the weather favor Brazil more than Argentina in sugar production. dates back to 1975, when the first oil crisis led the Brazilian government to encourage the use of cane alcohol in gasoline-the so-called Programa Proálcool, which was seen as a strategic choice of paramount importance by the military administration (Fávero 2000). From then on, the production of sugar was artificially increased, and it became, not the main product of cane, but a byproduct tradable at marginal prices (Lucca 2001).
The effects of the productivity differential on trade are sizeable. The subsidies to cane have turned Brazil into the largest producer of sugar in the world: its share in global trade ranks close to 30 percent. To properly assess the magnitude of this figure, it should be taken into account that sugar is one of the most protected commodities worldwide (Fávero 2000;Romero 2001). Argentina, on the other hand, faces a difficult situation, which is due to tough foreign competition and aggravated by the domestic conditions in which the production of sugar takes place. The Argentine sugar crop is spread over five provinces in particular, the most important of which are Tucumán, Salta, and Jujuy, all located in the poor northwestern region. 13 Sugar constitutes the basis of the regional economies in all three provinces, and it is the most significant source of employmentboth direct and indirect-after the public administration. Given the federal arrangements established by the Argentine constitutional regime, and given the social and economic reasons just mentioned, the senators and deputies representing the northwestern provinces constitute a solid bloc in Congress that seek to defend their "sugar constituents." The Brazilian government has always insisted on the prompt inclusion of sugar into the free trade zone. By contrast, Argentina has sought to delay this outcome through a variety of strategies. The most conciliatory sectors in Argentina have accepted only the gradual incorporation of sugar into the free trade area. Their aim was to establish a ten-year transition period during which the sector would establish protective mechanisms to guarantee supply for the domestic market. Brazil, on the other hand, proposed a transition period of no more than twelve months. The two stances were impossible to reconcile. Time was not the only source of disagreement: some of Argentina's toughest sectors had even asked Brazil to abolish the Proálcool system, a demand that was rejected outright. Brazil claims that there are strategic, political, and ecological reasons-not just economic ones-to maintain the plan, so it could not be abandoned.
Given the complexity of the issue and the contradictory interests that had prevented any agreement, in 1994 the CMC resolved to create an ad hoc committee and to establish the special sugar regime. Three decisions were adopted between August 1994 and December 1996 14 whereby the 13. By contrast, Brazilian sugar is grown mostly in the wealthy states of the South. 14. CMC Decision 7, August 5, 1994;CMC Decision 19, December 17, 1994;and CMC Decision 16, December 16, 1996. committee was requested to define a regime that would adapt the sugar sector to the customs union. There were two parameters: the proposal had to aim at gradual liberalization of intra-MERCOSUR sugar trade, and it had to neutralize distortions stemming from the asymmetries between different national policies. The initial deadline was November 1, 1995 but was later extended to May 31, 1997. The latter date brought with it the first of two serious crises that MERCOSUR faced over sugar, with the second taking place in 2000. They are analyzed below.

Circumventing the Parliaments: Dispute Settlement through Presidential Competences
Between 1995 and 1997, Argentina acrimoniously questioned the "MERCOSURness" of Brazilian sugar subsidies. According to Senator Alberto Tell (Partido Justicialista [PJ], Jujuy), "against subsidized sugar Argentina cannot compete, as it has a free and deregulated regime." 15 Cane croppers and their representatives demanded that there should be no additional protection, but rather that sugar imported from Brazil should cost as much as that coming from any other country. In the same vein, Brazilian deputy Paulo Bornhausen (Partido da Frente Liberal [PFL], Santa Catarina) countered that: "Argentina's wheat exports are based on strongly subsidized crops. This fact makes Brazilian producers unable to compete in fair conditions with Argentine-imported wheat." 16 The menace of retaliation was clear. These positions delineated the scenario in which both congresses would confront one another, embracing defensive strategies on behalf of their respective interest groups.
In May 1997, the Argentine Congress passed a law (Ley 24.822) imposing the same tariff on imports of Brazilian sugar as that applied to sugar from non-MERCOSUR countries. It was intended to be in effect as long as the asymmetries caused by the subsidies to the Proálcool Program remained in place. Although the Brazilian authorities denied having exerted any pressure on the Argentine government, 17 Menem vetoed the law through Decree 471/97. The alleged reason was that Ley 24.822 contradicted MERCOSUR treaties, a stance that coincided with the official Brazilian position (Vigevani et al. 2001).
Some months later the dispute over the status of sugar trade was reinitiated, and in September 1997 the crisis peaked. On September 3, the Argentine Senate rejected the executive veto, sticking to the position adopted earlier by Congress. Against the position manifested by 15. "El Congreso rechazó el veto a la ley del azúcar," Clarín, September 4, 1997. 16. "Câmara ameaça retaliação contra Argentina," O Estado de São Paulo, September 6, 1997.
Economy Minister Roque Fernández and Foreign Minister Guido Di Tella, the congressional vote was unanimous. The electoral campaign for the congressional mid-term elections was under way, and no legislator wanted to be exposed to criticism by paving an open road for Brazilian sugar imports. The argument escalated bitterly after the overriding of the veto. Reaction on the Brazilian side was not long in coming. In the Brazilian congress "there were retaliation threats, aimed at limiting imports of Argentine wheat. Apart from its socio-economic significance, the episode had intense repercussions because it took place within a delicate political context, marked by public divergences between Brazil and Argentina regarding the reform of the UN Security Council" (Seixas Corrêa 1999, 252). By public divergences, Seixas Corrêa meant the obstacles that Argentine diplomats were placing to Brazilian attempts to secure a permanent seat on the UN Security Council should a reform of the UN charter take place. These divergences were no more than background noise; however, the real issue was conflicting sectoral interests and their parliamentary echoes.
The Brazilian executive adopted a consistently prudent stance throughout the conflict. Cardoso and his aides were aware that Argentine legislators were acting under pressure of the electoral campaign, so they opted to trust the efforts that the Menem administration was making to limit the damage. The Argentine executive publicly condemned the congressional position and eschewed legislative proposals that ranged from proposing a new bill revoking Law 24.822 to contesting its constitutionality before the Supreme Court. Simultaneously, the government requested more time to fix the problem from its Brazilian counterpart. There were some officials on the Brazilian side-notably Foreign Minister Luiz Felipe Lampreia 18 -who were not convinced of the sincerity of the Argentine position, but Cardoso chose to give Menem the benefit of the doubt.
The situation appeared to have culminated in a dangerous deadlock. According to Danese (2001), it was then "clear that the situation was getting to an unbearable point and that a firm presidential intervention would be necessary to move things forward. Therefore, a presidential meeting was agreed upon." On November 10, Menem traveled to Brasília. The trip was not an isolated gesture; as the Brazilian deputy foreign minister later testified, there is an observable pattern according to which MERCOSUR trade crises are overcome through top-level political bargains, which invariably reaffirm the commitment to regional integration . . . The official visit of Menem to Brazil, on November 10, 1997, reversed the negative atmosphere created by the sugar crisis of September 1997 (Seixas Corrêa 1999, 252-3).
The controversial law was still in force, however, and a formal solution had yet to be found. Finally, the Argentine president adopted a formula that dated back to the Viceroyalty of La Plata: "acátese, pero no se cumpla" ("acknowledge but do not obey"). In ancient times, this practice worked well for rulings coming from Spain; now, they worked against the laws originating in the Argentine Congress. The tariff was never applied.
Presidential dominance within the national Congresses and the Joint Parliamentary Commission is best underscored by the second sugar episode. On August 24, 2000, the Argentine Congress passed another law that was based on the usual arguments and aimed to protect Argentine cane from Brazilian competition. The law extended the validity of Decree 797/1992, which had established a protection regime for the sugar industry that was set to expire in December 2000 and established no further deadlines. Brazilian officials issued condemnatory responses, whereas the Brazilian Congress threatened to limit imports of food containing sugar from Argentina. President de la Rúa was upset about the Argentine legislative decision. He was expected in Brasília on August 31, where a South American summit was set to take place, and was not pleased to be received in an unfriendly climate. After a bitter cabinet meeting, he sent his foreign minister 19 to negotiate with the deputies from the northwestern provinces in order to avoid "a war between the Brazilian and Argentine congresses." 20 While President Menem's visit of November 1997 had allowed the softening of tensions between the two assemblies, providing a temporary resolution for the first sugar episode, this time resolution appeared to be more difficult. The risk of an outbidding conflict that could jeopardize the whole process looked greater. MERCOSUR had not yet overcome the effects of the 1999 crisis, and legislators from both countries were threatening retaliation even when no electoral campaign was underway. Members of the Joint Parliamentary Commission attempted to mediate, but they could not reach an agreement (Mustapic and Geneyro 2000). On September 1, de la Rúa promised Cardoso that his government would not endorse the controversial law, thus risking a serious domestic conflict. 19. The fact that it was the Argentine foreign ministry-and not another executive office-that handled the situation was acknowledged and welcomed by the Brazilian administration (as confirmed at the Itamaraty's Divisão do Mercosul by Mauricio Fávero, 2000), for this highlighted the political relevance of the conflict. 20. "Ley de protección a la industria azucarera. El azúcar amargó al gabinete," Clarín, August 30, 2000. Just when it seemed that things could get no worse, the Gordian knot was cut by President de la Rúa: on September 4 he vetoed the law, only to sign a resolution with the same contents (though it also established a five-year limit). The goal was twofold. On the one hand, the decision prevented an open-ended escalation between congresses, expediently removing the issue from the legislative arena. On the other hand, a decree in place of a law improved the executive's negotiation power, as it enabled the Argentine president to ensure his Brazilian counterpart that the decree would be lifted as soon as an agreement was reached, without having to go through an uncertain legislative process. 21 Subsequently, both presidents agreed to initiate a round of negotiations with the object of incorporating sugar into the free trade zone, along the lines manifested in the communiqué of Buenos Aires of June 30.
This event shows how the presidents have chosen to take all responsibility whenever MERCOSUR undergoes a serious conflict. They have done so even at the expense of relinquishing the likely benefits that could be extracted from a two-level game (see Putnam 1988). When negotiating with foreign partners, the presidents can no longer argue that domestic institutions keep their hands tied, as they have made it clear that no actor stands above the presidential will. Vigevani et al. (2001) argue that the sugar issue exposes the inadequacy of MERCOSUR structures for overcoming threats posed by negatively affected interest groups. They describe how Argentine and Brazilian social and economic groups organized themselves at the national level to defend their interests. Pressure was aimed at national lawmakers in both countries, bringing about a confrontation between the two congresses and engendering a diplomatic crisis. The evolution of the crisis illustrates the limitations of MERCOSUR's institutional structure, and its resolution confirms that the final dispute-settlement mechanism was inter-presidential bargaining. However, as this case shows, presidential interventions were successful because they were backed by institutional capabilities such as veto and decree powers and because the presidents were not accountable to the legislative branch. By contrast, a prime minister who vetoed or ignored a decision of the parliament might well have faced a motion of no confidence, given the institutional characteristics of parliamentary regimes.

THE GRAND CRISIS OF 1999: MANAGEMENT AND OUTCOMES
This section analyzes the crisis triggered by the 1999 devaluation of the Brazilian currency and the ensuing reaction of the Argentine 21. "De la Rúa firmó una resolución y vetó una ley. El azúcar, protegido hasta el 2005," Clarín, September 5, 2000. government. These developments drove the members of MERCOSUR to the conclusion that tighter macroeconomic coordination and more effective supervision and implementation of the common agreements were necessary, leading to further efforts to relaunch the process in accordance with the summits of Buenos Aires (June 2000) and Florianópolis (December 2000).

The Brazilian Devaluation and its Aftermath
Throughout most of the 1990s, both Argentina and Brazil experienced periods of domestic turbulence due to political (uncertain electoral outcomes) and economic (inflation or recession) reasons. However, such difficulties did not occur in both countries at the same time, which allowed integration to continue since the largest MERCOSUR partners compensated each other in times of trouble. In 1999 the situation changed radically, as both economies went simultaneously into recession. To make things worse, Brazil was compelled to devalue its currency, a measure that threatened to alter intraregional trade and investment flows dramatically.
When the Brazilian decision to devalue was adopted on January 13, Menem was in the United States on an official visit. In Argentina, there was criticism of the fact that the Cardoso administration had neither consulted nor informed its partner government about a decision that had such a deep impact on Argentina. Menem and his economy minister, Roque Fernández, refuted the criticism with two different arguments: for Menem, the measure was technically not a devaluation; for Fernández, it was logical that the devaluation had to be undertaken without previous warning. In any event, the contact between both presidents was so close that Menem received a phone call from Cardoso on January 14, while he was giving a press conference. 22 The Argentine reaction to the Brazilian decision was varied. Whereas the government remained calm and sensitive to its counterpart, many representatives of the industrial sector and the media feared a possible "avalanche" of cheap goods from the giant neighbor. 23 Fears of invasion by Brazilian products were first called "efecto Brasil" and later "Brasildependencia." The former label, somewhat neutral, denoted a likely danger by contagion; the latter reflected a sentiment of subordination that called for a protectionist reaction. The Menem administration took some rapid measures to relieve domestic pressures, such as reducing taxes, diminishing import tariffs for sensitive goods coming from non-MERCOSUR 22. "Menem prometió a los inversores que la Argentina no devaluará," Clarín, January 1, 1999. 23. "Efecto Brasil: Temen avalancha de productos brasileños por la devaluación. Los industriales pedirán medidas de protección," Clarín, January 17, 1999. countries, and asking the Brazilian government to eliminate distorting subsidies. All this was insufficient to dissipate the public concern, and the Argentine Industrial Union (UIA) insisted on a rise in tariffs. 24 The months following the devaluation witnessed increasing tension between the largest MERCOSUR members. All South American countries except Bolivia, the smallest economy in the region, and Argentina had adjusted their exchange rate after Brazil. In Argentina, the Convertibility Plan-a law that tied the national currency to the dollar-prevented such a response. In February, the presidents and ministers of Argentina and Brazil met twice: first in Mexico, in the framework of a Rio Group meeting, and later on February 12 in São José dos Campos, in Brazil. These summits were not enough to ease the apprehensions due to the ongoing Brazilian crisis, so in April the Argentine government had to act to calm the domestic front.
Between April and July, the Argentine executive issued three rulings to promote balanced trade with Brazil. On April 19, Resolution 458 put in motion an antidumping procedure against steel imports; on July 14, Resolution 861 fixed import quotas for textiles; finally, on July 26, Resolution 911 incorporated a Latin American Integration Association (ALADI) norm into Argentine legislation to allow the country to apply safeguard measures. It made no mention of exceptions for MERCOSUR members, so it could be interpreted as an attempt to establish trade barriers via import quotas, in open conflict with MERCOSUR rules. The Brazilian government reacted strongly: for the first time since the creation of the bloc, it announced through Undersecretary of Foreign Trade José Alfredo Graça Lima that all negotiations were suspended. 25 A year later Graça Lima (2000) diplomatically downplayed these events, preferring to speak of "trade confrontation" rather than political crisis. However, the magnitude of the controversy can hardly be overlooked; indeed, most key protagonists viewed it as the most critical moment since 1985.
During the hard times between the Brazilian devaluation in January and the Argentine Resolution 911 in July, some progress had been made. On April 28, the first ruling was issued by the ad hoc tribunal set up after the Protocol of Brasília, generating expectations of a legal institutionalization of MERCOSUR procedures. Shortly after that, on June 7, Menem and Cardoso held a meeting in Olivos, first alone, and then with their foreign ministers, at which time they seemed to reach agreements on fiscal responsibility and macroeconomic coordination. A week later, on June 14, they met again in Asunción in the friendly context of the 24. "Los industriales insisten con subir los aranceles," Clarín, January 22, 1999. 25. "Presión por las medidas 'proteccionistas' de la Argentina. El gobierno de Brasil pidió una reunión urgente del Mercosur," Clarín, August 27, 1999. MERCOSUR biannual summit, together with the presidents of Paraguay and Uruguay. However, by the end of July the dialogue between Argentina and Brazil was over. Domestic pressure for tougher measures was mounting on both sides, and the prospects of MERCOSUR seemed bleak. Paulo Skaf, chief of the Brazilian Garment Industries Association, went as far as to state, "this is the first step towards the end of MERCOSUR." 26 How could such a dismal ending be prevented?

PRESIDENTIAL CAPABILITIES
On July 26, the Brazilian government accused Argentina of "an explicit violation of the Treaty of Asunción," and denounced "an alarming regression in the development of the integration process" in a letter sent to the temporary president of MERCOSUR in Montevideo. 27 After declaring a halt to all negotiations, the government requested an urgent bilateral meeting to protest the quotas and attempt to prevent their extension. Both administrations were internally split over the issue. The Argentine foreign ministry blamed Secretary of Industry Alieto Guadagni for promoting the controversial resolutions, whereas the latter accused Foreign Minister Guido Di Tella of nurturing Brazilian distrust by unilaterally applying for NATO membership. In Brazil, the position of the hard-liners and soft-liners was reversed: while ministers of finance, Pedro Malan, and of industry and trade, Clovis Carvalho, seemed to sympathize with the Argentine position, Foreign Minister Lampreia was extremely harsh. Knowing that Menem was planning to travel to Brazil in August, Lampreia discouraged the visit, publicly declaring on July 27 that the political climate was not propitious for a presidential summit. On July 28, Cardoso's spokesman added that the presidential visit "should be postponed until times are better." 28 However, on July 29 the Argentine presidential plane landed in Brasília unexpectedly, and Menem was on board. After dinner at the presidential residence, the two presidents gave a press conference where they denied the existence of any differences between the two countries. What can explain such a sudden shift from bitter anger to intimate friendship?
Menem had arrived in Brasília from New Orleans, where he had been taking part in a hemispheric conference of energy ministers. A technical stop late on Thursday in Manaus, in the north of Brazil, had been programmed for the trip home. On Wednesday morning, however, when 26. Idem. 27. "Maratón de reuniones en Buenos Aires. Argentina discutirá con Brasil la crisis comercial," Clarín, July 28, 1999. 28. "De Manaos a Brasília," Clarín, July 30, 1999. Menem was boarding the plane to go to New Orleans, Di Tella suggested that he arrange a summit with Cardoso. Menem agreed. Di Tella wanted an open-ended meeting to allow the two countries to harmonize positions without giving the impression of an Argentine defeat. The secretary general of the Argentine presidency, Alberto Kohan, had other plans, though. Kohan was one of the men closest to the president. He was known for his conflictive nature within the cabinet: he had confronted many ministers with the aim of surrounding Menem with people that he trusted. Menem's style of command was based on a double-checking system: he would instruct two subordinates to carry out the same mission separately, enabling him to play one against the other. Kohan was his preferred emissary when it came to leading parallel political operations of this kind, although such operations were usually only used in the domestic arena. Parallel international diplomacy was a new, more serious practice.
Kohan bypassed Foreign Minister Di Tella and telephoned Lampreia. On his way to the United States on Wednesday evening, Kohan told the Brazilian minister that Menem had decided to exclude Brazil from the scope of Resolution 911. They then agreed that the stopover on the way back from New Orleans to Buenos Aires would be in Brasília rather than Manaus so that the two presidents could meet. Notwithstanding his previous warnings against Menem visiting Brazil, Lampreia diplomatically accepted the new arrangement. Di Tella, who had stayed in Buenos Aires, had to come to terms with this fait accompli, which confirmed that he had been sidelined and that his substantive plans had been ignored. In his view, Kohan's conditions were closer to capitulation than to negotiation. However, he called in his associates and decided to adapt his strategy to the new situation, preparing to travel to Brasília the following day. As usual, government officials and business associations alike made no mention of the possibility of resorting to the Protocol of Brasília or the Joint Parliamentary Commission of MERCOSUR.
On Wednesday, before Menem's arrival, Di Tella, Guadagni and the undersecretary of Foreign Trade, Félix Peña, arrived from Buenos Aires and met with Lampreia. The climate was tense. The Argentines informed Lampreia that safeguards against Brazil were to be lifted and requested special treatment in light of the exceptional situation brought about by the devaluation. Lampreia showed little concern, however. He claimed that Argentina had "crossed the red line" and reaffirmed that MERCOSUR would be destroyed if the safeguards were applied. Later on, the foreign ministers participated in the presidential meeting, and Lampreia repeated his strongly worded complaints to Menem, also asking for written confirmation of the new Argentine decision. To this Menem answered that his word was sufficient guarantee. At this point, President Cardoso intervened, siding with his Argentine counterpart to ease the tension. 29 At the end of the day, both presidents appeared together for the "photo op," while their foreign ministers gave the press details of the agreement.
Argentine diplomacy had calmed the Brazilian government, but it was unable to prevent negative public comment. On Friday, the newspapers in Buenos Aires considered that Argentina had given in: "Argentina Gives Up and MERCOSUR Tension Declines." 30 Argentine officials did their best to counter this view, but their efforts were not rewarded. By the time that the following extraordinary meeting of the CMC took place a few days later in Montevideo, Menem had already issued Resolution 955 exempting MERCOSUR countries from the effects of Resolution 911.
In the episode just related, Menem affirmed his supremacy vis-à-vis the cabinet and the diplomatic corps. Under his personal supervision, his right-hand man had gone above the heads of politicians and bureaucrats to arrange a meeting with the Brazilian president, one that incidentally also contravened the public position of the Brazilian foreign minister. Thereafter, the presidents settled the dispute alone. When asked why Di Tella had accepted Kohan's parallel diplomatic efforts without complaint, a chief Argentine official replied off the record that the foreign minister was ready to accept any conditions as long as he could stay in office (author's interview). Thus, on this occasion, as before, he had accepted a presidential fait accompli.
Norberto Moretti, secretary of the Brazilian Embassy in Buenos Aires, argues that Menem was able to stop over in Brasília because of the previous decision to withdraw Resolution 911 (Moretti 1999). He paid a domestic price for this, as the industrial sector and most congressional groups demanded more, not less, protection against the Brazilian "threat." However, at the advice by Jorge Campbell, Menem had decided to prioritize MERCOSUR over transitory domestic issues, and he possessed the institutional capabilities to do so. Clearly, the system of collective responsibility that characterizes parliamentary cabinets would never have given a chief executive such autonomous room to maneuver.

CONCLUSION
The three cases analyzed above show the extent to which presidential intervention boosted the process of integration and shaped its outcome, with presidents acting not only as decision makers but also as dispute settlers and guarantors of commitments. The presidents were 29. "Entretelones de un día muy agitado. El palo y la zanahoria," Clarín, July 31, 1999. 30. "Argentina cedió y bajó la tensión en el Mercosur," Clarín, July 30, 1999. perceived to be efficient problem solvers because they had popular legitimacy and the determination to intervene. However, the tasks they performed were not merely based on charismatic leadership but also on institutional capabilities. In order to negotiate and secure agreements with their counterparts, Menem, Cardoso and de la Rúa made extensive use of the prerogatives granted to them by national constitutions and legislation. These prerogatives allowed the presidents to overwhelm other potential veto players such as the cabinet-including the economy and foreign ministers, who did not always belong to the president's party-and the congress in ways not available either to parliamentary or balanced presidential regimes. The presidents were also reluctant to build up regional institutions or to relinquish their competences to the regional institutions that did exist.
The degree of concentration of power exerts a differential impact on governmental capacities. Concentration of power is regarded as preferable when it comes to steering capacities-active policy making-whereas diffusion of power is more closely related to maintenance and political capacities (Rockman 1997;Weaver and Rockman 1993). This article argues that the concentrationist brand of presidentialism of the Argentine and Brazilian systems of government has helped promote regional integration in two ways: first, the constitutional capabilities of the presidents vis-à-vis other actors have made it credible that their policies would not be blocked. Second, the preeminence of the presidents has provided social actors with a single, non-bureaucratic target through which to channel their interests, allowing for a faster decision-making process. As a result, presidential intervention in the management of MERCOSUR has become a structural element of the integration process.
Conceivably, there are other reasons that help to explain the large room to maneuver enjoyed by the presidents. Among them, the tradition of executive supremacy in domestic politics is not without importance. A second is the feeble institutionalization of MERCOSUR, which features neither veto points nor veto players that can block presidential interventions. Finally, the absence or weakness of relevant social actors calling for enhanced participation facilitates the predominance of presidential decisions. These factors only add to the institutional resources already in the hands of the chief executives.
State-led, presidential-driven integration has been a persistent feature of MERCOSUR, as has been the low level of involvement of business organizations, particularly in Argentina and Brazil (Schneider 2001). As noted above, this represents a deviation from expected theoretical outcomes, since both liberal intergovernmentalism and supranational governance suggest that governmental action is a result of sectoral or functional demands. In MERCOSUR, the evolution of business involvement in the integration process also diverged from that of other Latin American regional schemes.
Schneider (2001) shows that both Chilean and Mexican business organizations have had a much more active role in fostering and shaping open regionalism than their equivalents in Argentina and Brazil. This phenomenon is not only a product of the greater strength and organization of key Chilean and Mexican social actors, as Schneider correctly points out, but also of the lesser power of their presidential offices-be it lesser partisan powers as in Chile, or lesser constitutional powers as in Mexico (Mainwaring and Shugart 1997, 432;Siavelis 2002). 31 In MERCOSUR, the limited intervention of entrepreneurs and business peak associationseven when they were invited to participate-did little to curtail the power held by the presidents. On the contrary, "the scarce participation of the private sector in the institutional structure implied that its interests were channeled through lobbying national negotiators rather than through formal means at the regional level" (Redrado 1999, 143). Thus, informal lobbying through direct executive agents is a relevant factor that also accounts for executive supremacy.
Some public officials believe that the troubles of MERCOSUR do not stem from the integration process itself but from the recurring crises of its member countries (Pereira 2000). In this context, stronger institutionalization would not solve its troubles; rather, it may accentuate them by adding institutional discredit to enduring ineffectiveness. Accomplishing a common market may require stronger institutions, but until 1999, MERCOSUR performed reasonably well without them. However, there was by then a growing awareness of the overwhelming position of the presidents, which ignited public debate on the issue. In 1999, former Uruguayan president Luis A. Lacalle stated that the MERCOSUR secretariat should be reinforced so as not to erode the image of the presidents. 32 In 2000, Chilean president Ricardo Lagos was even more explicit: "Whenever there are difficulties between our countries, the politics of the 'presidential phone' do not work. It is not possible for presidents to work things by getting involved in routine fights on the phone." 33 Contrary to Lagos's statement, and regardless of how undesirable it may appear, this kind of politics was not only possible but also effective. Above all, it was accepted and stimulated by the presidents of the largest member countries. 31. Although Mainwaring and Shugart classify the Brazilian presidency similarly to the Chilean, recent studies have shown that the partisan (and coalitional) powers of the Brazilian president are actually higher than they expected (Cheibub and Limongi 2002;Figueiredo and Limongi 2000).
32. "Mercosur: Lacalle pide no desgastar la imagen de los presidentes," El Observador, August 3, 1999. Lacalle sought to preserve presidential power for crucial issues. Coherently, he advocated the direct effect of regional normative in order to increase presidential power on regional matters, vis-à-vis domestic actors such as the Congress (Lacalle 2001).
33. "El Mercosur tiene que integrarse en serio," Clarín, May 14, 2000. In January 2000, the Brazilian government created the office of the Special Ambassador to MERCOSUR and appointed ambassador José Botafogo Gonçalves to the post. The intention was to demonstrate that MERCOSUR was a priority for the Cardoso administration. Up until then, it had been the Undersecretary of Trade and Economic Integration of the Foreign Ministry, Alfredo Graça Lima, who had conducted coordination tasks for MERCOSUR. As he later underlined, the new office depended on the presidency and was not part of the structure of the foreign ministry (Graça Lima 2000). This meant that, even in the country with the most professional diplomatic corps, the president decided to keep direct command of MERCOSUR affairs.
To be sure, summit diplomacy is not a feature exclusive to MERCOSUR. The originality of MERCOSUR resides in its being the most significant regional bloc whose chief executives are all presidents and whose regional institutions are negligible. Provided that domestic regimes have an impact upon international relations, it is conceivable that not only the democracy-autocracy dichotomy but also the presidentialparliamentary distinction is likely to make a difference. This paper shows how the presidents of MERCOSUR countries have contributed to shape the integration process and to make it work by making use of their institutional and political capacities.