Political parties, foreign policy and ideology: Argentina and Chile in comparative perspective^*

Janina Onuki^I; Pedro Feliú Ribeiro^I; Amâncio Jorge de Oliveira^I

^IUniversity of São Paulo, Brazil

ABSTRACT

The aim of this article is to discuss the distribution of preferences of members of the Chilean and Argentinian Congress on foreign policy issues through the analysis of roll call votes. This goal is guided by the debate in Latin American literature concerning the decision-making process in foreign policy. The predominant argument focuses on the Executive as the principal decision-maker, disregarding the Legislative as relevant in this field. Thus, legislators would tend to abdicate from their preferences in determining foreign policy. Confronting this argument, we have many studies emphasising the importance of domestic actors in the foreign policy decision-making process. This article proposes to analyse two case studies in comparative perspective: the lower houses of the national parliaments of Argentina and Chile. The result is that the party ideology is a relevant explanatory factor of deputies' votes. Although the argument is more evident for the Chilean case, it is possible to argue that there is a similar pattern to the structuring of deputies' votes in the two countries, both on the domestic and on the international arena. The methodology used makes it possible to infer legislators' preferences by means of roll call votes and of the construction of maps of deputies' ideal points in foreign policy terms, as well as the correlation between Chilean and Argentinian parties' ideological classifications. Votes on foreign policy questions during the 2002-2006/2007 legislatures are considered.

Keywords: Legislative; Political parties; Foreign policy; Chile; Argentina.

Introduction

The main aim of this study is to delineate the distribution of preferences among Chilean and Argentinian deputies on foreign policy themes by means of an analysis of roll call votes. This aim is oriented by the debate being had in Latin American International Relations literature regarding the process of foreign policy-making. The predominant argument positions the Executive as the main formulator, disregarding the Legislative branch as a relevant arena. Thus, Latin American legislators would tend to abdicate from their preferences in determining foreign policy, affording presidents full conditions to establish foreign policy without legislative interference.

As a counter-position, several studies emerge in the 1990s stressing the relevance of domestic actors in foreign policy formulation, national congresses among them. Although much of this literature makes reference to the US case, recent studies focusing on Latin American countries maintain that the Legislative is relevant in decision-making on foreign affairs, even in the face of a preponderant Executive. It is precisely within this perspective that the central argument of this article fits, taking the lower house of the Argentinian and Chilean National Congress as its objects of analysis.

The Chilean political system, like many others in Latin America, functions under the aegis of multiparty presidentialism, which makes it extremely unlikely for a president to be elected just with the backing of his/her own party, which, in turn, is unlikely to obtain a majority of congressional seats (Nolte 2003). Hence, the scant possibility of a single party winning the presidency and a majority in Congress makes it necessary to form a governing coalition to ensure the political system's stability and governability. The Chilean binomial electoral system offers strong incentives to the formation of two party coalitions (see for example Carey 2002; Aninat et al. 2004; Vásquez 2006). Beyond institutional factors, the makeup of Chilean political coalitions is strongly oriented by an ideological polarization on the left-right continuum. From the country's re-democratization in the early 1990s until the legislature analysed here (2002-2006), the Concertación centre-left coalition formed the Chilean government, while the Alianza por Chile rightwing coalition played the role of the opposition in the National Congress.

For its part, Argentina has a bicameral presidentialist system and adopts federalism as its political model. This structure seems to directly influence parties' behaviour, not just on domestic policy matters, but also in terms of foreign policy. Although this aspect is not dealt with here, it is worth pointing out that the provinces have taken on more prominence and capacity of influence since the 1994 constitutional reform, and the parties seem to reproduce this structure through their national, provincial and local leaderships. While Partido Justicialista (PJ) and Unión Cívica Radical (UCR) have historically become consolidated as the majority parties, the fragmentation of the Argentinian party-political system is coming into sharper focus. One of the explanations is the fact that Argentina holds legislative elections with closed lists, thus obliging voters to choose parties. Intra-party divergences have led to the formation of new parties, rather than coalitions as is the case in Chile.

There are many discussions going on about party discipline in Argentina that do not fit in this article. Some authors state that the institutional rules of the electoral process, in conjunction with legislative organization, lead to a high level of party discipline (Jones 2002). Others, like Mustapic (2000), argue that high levels of party discipline are not to be found in Argentina. The mismatch in interpretations, these authors allege, is related to measurement problems.

The main thesis espoused here is that the distribution of Chilean and Argentinian deputies' preferences regarding foreign policy is extremely similar when compared with the same distribution for domestic policy. Ideology of the political party appears as a relevant explanatory factor of deputies' votes on domestic and foreign policy matters. Though this argument is more evident for the Chilean case, it is possible to argue for the similarity in the pattern of structuring of deputies' votes in the two countries on the domestic and international arenas. In other words, the factors that determine legislators' votes in the domestic ambit seem to be relevant also in the case of foreign policy.

In section 2, we review the literature pertaining to the theme, stressing two diametrically opposed views on the participation of national congresses in foreign policy. In section 3, we delineate the methodology used to infer legislators' preferences by means of roll call votes, notably the WNOMINATE program. In section 4, we describe the results of this study, chiefly the maps of deputies' ideal points in foreign policy terms, as well as the correlation between the ideological classifications of Chilean and Argentinian parties in the domestic and international arenas. Lastly, in section 5 we present the conclusions.

Political Parties, the Legislative and Foreign Policy

In general, it is possible to identify two opposed views as to the role of the national congresses in foreign policy formulation: the first argues for the preponderance of the Executive in the face of a Congress that is little or not at all assertive on foreign policy questions. The second position sees Congress as active in foreign policy as in the domestic ambit.

One of the central arguments of the first perspective is that the president will always obtain greater space for action in international affairs than in domestic matters. In US literature, this postulate became known as the "two presidents thesis". In other words, there is an imperial-like president in the conduct of US foreign policy, and another one, strongly constrained by a powerful Congress, for domestic matters (Lindsay and Ripley 1992).

According to the thesis of the preponderance of the Executive in international themes, ideology and political parties influence in determinant fashion only those policies circumscribed to the domestic ambit (Edwards 1989; Bond and Fleisher 1990;¹ Ragsdale 1995). Constant congressional support for the president's foreign policy has been termed "bipartisanship", characterized by the joint action of Congress and the Executive in pursuit of common goals, even if a conflict of interest were to emerge. Hence, bipartisanship essentially has two elements: unity on external matters, i.e., political support by the majority of both US political parties, and practices and procedures intended to achieve the desired unity (McCormick and Wittkopf 1990).

The possession of formal powers in the conduct of foreign policy that are denied him/her with respect to domestic affairs is one of the main explanations coined by this literature to justify the preponderance of the president (Lindsay, 1994). Other explanations often provided include the difficulty in obtaining re-election by parliamentarians active in the foreign policy field and the greater technical and operational capacity of the Executive and its agencies to conduct the complex foreign relations of the United States (Kegley and Wittkopf 1995).

Still within this perspective but now focusing on the Latin American literature, Lima and Santos (2001) produced a study of the Brazilian case whose central argument is the abdication of authority by Congress to the Executive in the foreign policy decision-making process. Through a one-dimensional spatial model, the authors argue that the position of the president as the policy initiator, and of Congress as the ex post facto ratifier, generate an equilibrium whereby the median legislator is obliged to accept the policies negotiated by the Executive at international forums due to the calculation of the political cost of rejecting them. They claim this was the case of the trade liberalization policy pursued by Brazil in the early 1990s. In general, in spite of the lack of empirical studies in the field, the specialized Latin American literature tends to evaluate congresses' participation in international questions as weak (Stuhldreher 2003).

The main reasons for Latin American congresses' low level of assertiveness in foreign policy-making pointed out by the literature are the major concentration of power in the presidency, the lack of institutional articulation and instruments of expertise, and low electoral returns (Lima and Santos 2001; Santos 2006; Oliveira 2003; 2005).

In the United States, the extremely negative repercussion of the Vietnam War in the eyes of public opinion was a milestone in the strengthening of the channels of direct participation by Congress in the conduct of foreign policy² (Meernik 1993; Ripley and Lindsay 1993). This is the starting point of much of the literature that counterposes the view of an imperial-like Executive with reference to international themes.

As Warburg (1989) argues, institutional reforms in the US Congress during the 1970s (post-Vietnam War) transformed Congress' institutional environment, less prone as it was to protect the president's foreign policy priorities from congressional opposition. Hence, institutional changes in Congress also make up the set of explanations offered by the literature for the end of bipartisanship in US foreign policy. Prins and Marshall (2001) argue that foreign policy issues exert great influence on the level of Congress' support of the Executive. So depending on the issue in question (high politics or low politics, for example), there is significant variation in congressional support for the president. Lastly, Fleisher et al. (2000) demonstrate that the structural conditions of international politics also influence the level of consensus between the Legislative and the Executive.³

Pursuing this line of argument on the assertiveness of the Legislative in the conduct of foreign policy, Lisa Martin (2000) develops a thesis whose central point is legislators' capacity to influence foreign policy results even when there is a situation of delegation of powers by the Legislative to the Executive. The author argues that in an anarchic international environment, the institutionalization of legislative participation in international cooperation enhances the credibility of the commitments made by States, making international cooperation deeper and more stable. This factor, it is claimed, attributes great importance to legislative activity on the issue,⁴ since an attempt by the Executive to reduce the Legislative's capacity to influence international negotiations would lead to a loss of credibility in the face of the other negotiating party. By ignoring legislators' preferences in the process of negotiation of an agreement, the Executive becomes incapable of offering guarantees of its implementation. For this reason, Martin (2000) rejects the thesis of the Legislative's abdication of formal and informal powers in the conduct of foreign policy.

Moreover, according to Martin (2000), lawmakers' position on public policies in general, including foreign policy, is crucial to their re-election, thus diminishing the chance of abdication. Therefore, legislators' disinterest in the issue is a hypothesis the author discards. In order to understand the role of the Legislative in determining foreign policy, it would be necessary to look beyond the actions carried out by it (for example, the non-approval of an agreement signed by the Executive), privileging an analysis of the relationship between lawmakers' preferences and actual foreign policy results (Neves 2003).

When it was ascertained in the early 1990s that Congress had become more active, this produced a significant impact on the expert literature. Several empirical studies have established the influence of party affiliation, ideology and special economic interests over congressional decisions on foreign policy issues in general, with the predominance of those whose focus is trade policy in particular⁵ (Fordham 1998; Conley 1999; Baldwin and Magee 2000; Bardwell 2000; Fordham and McKeown 2003; Xie 2004; Delaet and Scott 2006). Thus, contrary to the first view presented at the start of this section, the ideology of congressmen and women, constituency, party influence and organized economic interests become important explanatory variables of the results of US foreign and trade policy.

McCormick, Wittkopf and Danna (1997), for instance, show that partisanship exerted great influence over roll call votes on foreign policy matters in the US Congress during the Bush administration and Clinton's first term. Wittkopf and McCormick (1998) observe a much lower level of support for the presidential foreign policy during the post-Cold War period, citing as the main motive congresspersons' distinct ideological preferences. As for trade policy, Baldwin and Magee (2000) also find the strong influence of legislators' ideology on the votes for the approval of North American Free Trade Agreement (NAFTA) and General Agreement on Tariffs and Trade (GATT), and the rejection of Fast Track authority in 1998. Constituency is another influential variable in legislators' decisions on foreign policy questions according to US literature. In the face of events that gain repercussion in the media, it is argued that public opinion exerts major influence on the formation of legislators' preferences (Lindsay 1994).

From the point of view of comparative politics — going beyond the case of the United States — one finds an interesting strand of academic production whose empirical investigation positively correlates party ideology and legislators' foreign policy preference, making up what we term second perspective (see Thérien and Noel 2000; Marks et al. 2006). For example, Milner and Judkins (2004) examine the positioning of political parties on trade policy questions in 25 developed countries (most of which members of the Organization for Economic Cooperation and Development (OECD) between 1945 and 1998. The central argument of this study is the existence of a strong impact of political parties' positions, in a one-dimensional left-right ideological scale, on the positions taken by politicians and legislators on matters of trade policy. Moreover, in the sample used in this study, parties of the left tended to favour free trade, while those of the right tended to oppose it (Milner and Judkins 2004).

With reference to the influence of Latin American Legislative branches and political parties, although incipient, the academic production on the theme contains case studies that stress the influence of national congresses on foreign policy formulation. However, there is a near absence of comparative studies on this subject. As an exception, we may cite the work of Randall Parish (2004). In it, the author produces a model with several systemic, institutional and economic variables to test their significance in the formulation of the foreign cooperation policy of various Latin American countries. Among them, the party system variable stands out. Party systems with strong political parties characterized by high levels of party discipline and electoral stability represent national-level constituencies, like Executives.⁶ This leads the Legislative to work with the Executive on foreign cooperation policy. The opposite happens in fragmented and undisciplined party systems, where lawmakers have few incentives to support national over parochial interests in the electoral equation (Parish 2004).

As for the case studies mentioned above, the following are worthy of note: Neves (2003); Mena (2004); Alexandre (2006); Feliú, Galdino and Oliveira (2007); and Leão (2008). Analysing Executive-Legislative relations in the formulation of Brazilian foreign policy, Neves (2003) finds the strong predominance of the Executive, stressing, though, that this predominance is not the preserve of foreign policy, but common to much public policy-making in general. In the case of Brazil, there prevails the delegation rather than abdication of authority by the Legislative to the Executive in foreign policy-making. Also in Brazil, when analysing the institutional mechanisms of legislative influence in foreign policy, Alexandre (2006) argues that despite clear delegation, there is no abdication on the part of Brazilian parliamentarians on the issue. On the contrary, the use of mechanisms such as provisos, amendments and interpretive clauses of international treaties reveals an attempt at exerting control, even if ex-post facto, over the Executive on foreign policy matters.

According to Mena (2004), the process of trade liberalization that occurred in Mexico during the 1980s was accompanied by a significant increase in the participation of Congress in foreign policy, chiefly from the ratification of NAFTA in 1994 onwards. It is worth highlighting the fact that this enhanced participation was not accompanied by an increase in its constitutional prerogatives with respect to foreign policy, but rather by the use of pre-existing — and previously under-used — constitutional mechanisms (Mena 2004).

Lastly, Leão (2008) has investigated the influence of the Chilean Legislative on the formulation of foreign trade policy during the 1990s. The author argues that the Chilean Legislative, even though restricted to ex-post facto action, has the capacity to influence the trade policy decision-making process, since the Executive incorporates the preferences of the institution's median legislator into its decision-making on the theme. Additionally, Feliú, Galdino and Oliveira (2007), after quantitative analysis of roll call votes relating to free trade treaties in the Chilean Chamber of Deputies since the country's re-democratization, conclude that the location of a deputy's political party in the left-right ideological spectrum is an excellent predictor of his/her votes regarding trade policy. Furthermore, it was found that deputies located more to the left of the spectrum tended to vote in favour of free trade (Feliú, Galdino and Oliveira 2007).

An analysis of the empirical studies alluded to above points to the predominance of three central hypotheses in explaining congresspersons' votes: 1) the influence of the ideology of the congressperson's political party; 2) the influence of the congressperson's constituency; 3) the influence of special economic interests linked to the congressperson. In this article we intend to test the first hypothesis with reference to the chambers of deputies of Argentina and Chile for purposes of foreign policy. Therefore, the question to be answered is: Is there a correlation between the position on the political-ideological spectrum of the parties to which Argentinian and Chilean deputies belong, and their respective votes on foreign policy issues? We thus aim to compare the distributions of preferences of Argentinian and Chilean deputies on foreign policy matters. In the next section, we discuss the methodology employed to answer the question put, as well as the sample of the study.

Methodology

Spatial models are important tools, used more and more often in analyses of roll call votes in Legislative branches. Intuitively, in these models the ideal point for each legislator is represented by a point in the Euclidean space, and each vote is represented by two points, one for yes and one for no. In each vote, the legislator votes as per the result closest to his/her point, at least probabilistically speaking. Taken together, this set of points forms a spatial map that summarizes the roll call votes (Poole 2005).

The empirical operationalization of the spatial theory of the vote depends on the statistical techniques employed. There are various statistical techniques developed in the literature to estimate the ideal point on the basis of roll call voting records (Heckman and Snyder 1997; Jackman 2001). The seminal work in question is the procedure developed by Keith Poole and Howard Rosenthal (Kalandrakis 2006), Nominal Three-Step Estimation (NOMINATE), whose aim is to estimate legislators' ideal point based on a certain sample of policies.

This method of analysis of roll call votes, because it is metric,⁷ permits the estimation of multidimensional positions for legislators and policies (Leoni, 2000). One must also remember that votes are subject to error. In other words, considering the hundreds of votes that a legislator casts over the course of a parliamentary term, it is not impossible for some to be inexplicable bearing in mind the legislator's pattern of behaviour. For example, on a vote about abortion, a liberal legislator might be against because his/her daughter happens to be considering having one, and at that specific moment he/she does not want to encourage the procedure. Each and every legislator has his/her error function, so to speak. Obviously, these "errors" are not mistakes and have an explanation, but they are not explicable by the usual utility function. Hence, the NOMINATE procedure employs a probabilistic model that makes it feasible to use error patterns to recover the coordinates of political results, assuming that some errors are more probable than others, independent and equally distributed among legislators and policies (Rosenthal and Voeten 2004).⁸ Thus, NOMINATE includes a signal-to-noise ratio that measures how strong the spatial component is in relation to the factors that caused the errors (Leoni 2000, 24-5).

In short, NOMINATE should be able to estimate the following parameters: i) the coordinates of legislators' ideal points for a given political dimension; ii) the coordinates of the "Yeses" and "Nos" for a given political dimension; iii) the typical size of the errors. To this end, we can say in a simplified manner that it is necessary to estimate, given the votes observed, which values for the parameters maximize the likelihood of producing the data observed. In other words, a method of estimation by maximum verisimilitude must be employed,⁹ which in essence is what NOMINATE does.

Since the number of parameters to be estimated is very large, instead of simply estimating by maximum verisimilitude, one begins from a given initial location of the parameters. Then one carries out the steps that improve the estimates, until achieving the maximum probability that the parameters reached by the end of this interactive process generate the votes observed.

In practice, the NOMINATE procedure manages to solve the problem of how to determine, based just on the records of legislators' "Yes" and "No" votes, what is the ideal point with accuracy, i.e., with a spatial measurement for each legislator. This technique of estimation determines how many relevant dimensions — patterns — exist, or to be more precise, what the explanatory capacity of each dimension is. In other words, we can compare the success of the model in capturing legislators' behaviour with one dimension with the success of the model with two dimensions, three dimensions etc. This is a relevant aspect of the model, for it allows one to evaluate its ability to fit in with the data and how many variables (dimensions) are relevant to explain the votes observed.

It must be noted, however, that statistical technique by itself is not able to supply substantive meaning to the dimensions found. More specifically, the NOMINATE procedure requires the analyst to inform as a reference point one legislator considered extreme in some relevant dimension. Usually, following the prescription of theory, one informs the position of the extreme legislator for whom there is most consensus as to their ideological position, i.e., within the left-right spectrum.

Hence, the first dimension estimated by the methodology takes as its basis the existence of this legislator and, therefore, has the substantive meaning referenced by this information supplied by the analyst, and that must be confirmed by means of the analysis of the resulting spatial disposition or of other statistical techniques (Leoni 2000, 6). For this reason, the substantive interpretation of the first dimension is usually ideology, while other dimensions that turn out to be relevant will demand interpretive analyses by the analyst (conservative v. liberal, nationalist v. internationalist, from the north v. from the south, government v. opposition etc). From the graphic point of view, NOMINATE produces a spatial map containing the estimated ideal point of each legislator and allows for the visualization of whether intuitive voting patterns exist or not.

On account of the interpretive component of the substantive meaning of the dimensions obtained, we have opted in this article to combine a qualitative approach with the method presented. More specifically, we will analyse the eight most polarized roll call votes (four for each country) in the sample. The analysis was made based on the stenographical notes of the plenary discussion on the day the matter in question was voted on. It is thus possible to identify deputies' motivations for voting one way or another from their speeches. With such information to hand, it is possible to substantively interpret the political cleavages in the Chamber of Deputies on foreign policy issues.

The sample

The dependent variable of the research is composed of the roll call votes of Argentinian and Chilean deputies on foreign policy matters. According to our hypothesis, the main independent variable is the political parties' positioning on the left-right political-ideological spectrum. Analysing the predictive capacity of this variable (party ideology), as well as the number of additional explanatory dimensions of parliamentarians' vote, is precisely the objective of the research.

With respect to the dependent variable, all the roll call votes on foreign policy themes by Argentinian and Chilean deputies during the 2002-2006 legislature in the Chilean case, and the 2002-2007 legislature in the Argentinian case,¹⁰ were included, totalling 267 votes.¹¹ The votes where the minority side did not surpass 2% were excluded,¹² as well as deputies who did not take part in at least 10 votes (cut-off criterion). Tables 1 and 2 show the respective values for each country.

Foreign policy related votes were classed as those that referred to international treaties and agreements signed with foreign countries and international organizations, trade policy, direct actions on international matters, measures referent to the functioning of the Ministry of Foreign Affairs (MRE), diplomatic representations and national defence issues.

"Direct actions on international matters" means legislative initiatives originating in the Chamber of Deputies whose theme is foreign policy. In this classification there are also requests made to the Executive, involving the same theme. With regard to trade policy, the following were included: free trade treaties; tariff and non-tariff barriers; bilateral agreements on investment protection; and agreements referent to double taxation. Tables 3 and 4 display the abovementioned data per legislature of each country.

Results

The number of dimensions needed to represent legislators' ideal points generally is small, given that legislators often decide their votes based on basic dimensions (Poole 2005). In the US Congress, for instance, the liberal/conservative ideological dimension is capable of forecasting the near totality of congresspersons' votes; it is the major structuring factor of roll call votes (Poole 2005). Analogously, in the Brazilian Chamber of Deputies the left/centre/right ideological dimension¹³ is also capable of forecasting the vast majority of federal deputies' votes (Leoni 2000).

Likewise, as may be observed on Map 1, the ideological dimension (first dimension) is capable of forecasting the votes of Chilean deputies on foreign policy issues during the legislature analysed here. Legislators from parties of the left are disposed on the left half of the spectrum (-1 to 0) of the first dimension, while legislators from parties of the right are disposed on the opposite half (0 to 1). This polarized distribution makes more evident the remark on the ideological constraint on Chilean deputies' votes on foreign policy questions.

As for Argentina, as may be observed on Map 2, there is a low level of ideological/party polarization on foreign policy issues during the period studied. One observes a major concentration of points (deputies) at the centre of the ideological spectrum (first dimension), indicating a convergence of preferences by Argentinian political parties with respect to foreign policy.

The comparison between maps 1 and 2 quite visibly reveals the fundamental difference between the preferences of Argentinian and Chilean political parties as to the foreign policy of their respective countries: the level of polarization. The low incidence of Chilean legislators (Map 1) at the centre of the political-ideological spectrum corroborates the finding of party-political polarization in their decisions relating to foreign policy. In this sense, it is possible to note the historical realignment of the Partido Demócrata Cristiano (DC) from the centre to the left of the ideological spectrum, coinciding with the argument of Alemán and Saiegh (2007). This realignment was essential in attributing internal cohesion to Chile's governing coalition, both in domestic and in foreign policy, thus forming a bipolar map of legislators' ideal points, which rests on the axis made up of the governing coalition (Concertación) and the opposition alliance (Alianza por Chile).

In the Argentinian case, the three main political parties in the Chamber of Deputies, when compared with their Chilean peers, display a distance of little significance in the ideological dimension. This means that the respective deputies' preferences are very similar and not very conflictive with respect to Argentinian foreign policy. Not even the parties in opposition to the president, Afirmación para una República Igualitária (ARI) and Unión Cívica Radical (UCR) display relevant distance from the governing party, Partido Justicialista (PJ).

In order to develop the argument initiated above more appropriately, we present Tables 5 and 6, as well as Graphs 1 and 2, which set out the location of Argentinian and Chilean parties on the ideological spectrum when it comes to foreign policy.

Tables 5 and 6 demonstrate some measures whose aim is to determine, based on estimated values for each deputy in the first dimension, the location of the main Chilean and Argentinian political parties in the political-ideological spectrum, as well as their level of internal cohesion. This first column of values (N Deputies) shows how many deputies make up each party's sample. It is worth remembering that owing to Nominate's cut-off criterion (see section 2), the number of deputies analysed in the tables above does not always coincide with the number of deputies in the legislature. Next, we present the mean,¹⁴ median¹⁵ and standard deviation¹⁶ of deputies' ideological values for each of the main Chilean and Argentinian political parties.

The next column identifies the interquartile range (IQR), a measure obtained from the difference between the upper and lower quartiles of the dispersion of the data. The upper quartile represents the values where 25% of the observations located more to the right of the party-political spectrum are found (closer to 1). The lower quartile represents the ideological location of the 25% of deputies more to the left of the party-political spectrum (closer to -1). As in the case of standard deviation, one can use the interquartile range as a measure of the level of political parties' ideological cohesion. The higher the IQR level is, the less cohesive the party, and vice-versa.

The values exhibited on the tables above can be graphically represented by boxplots, undertaken for the two countries analysed here (Graphs 1 and 2). This graphic representation makes it easier to interpret the values on the tables, as well as offering more information regarding the dispersion evaluated. In the boxplots, each "box" (red in the case of Chile and blue in the case of Argentina) represents 50% of the observations, in this case of the ideal points, aggregated by party, for the Chilean and Argentinian deputies in the ideological dimension (first dimension). The horizontal line inside the boxes represents the median of the observations. Additionally, the boxplot informs the sample's outliers, i.e., those values that deviate a lot from the sample's median, located more to the extremity than most other observations. In the Graphs 1 and 2, these are represented by points outside the interquartile range. The number that accompanies the points identifies the deputy in the database developed by this research.

With the data presented, we begin describing the ideological positioning of Chilean and Argentinian political parties on foreign policy issues. Table 5 and Graph 1 show the Partido Socialista (PS) more to the left in the ideological spectrum, followed by the Partido por la Democracia (PPD) and the Demócrata Cristiano (DC). All three main parties that make up the Concertación governing coalition occupy the left of the spectrum, revealing a high level of ideological cohesion on international matters within the coalition. On the other side of the spectrum, one finds the parties Unión Democrata Independiente (UDI) and Renovación Nacional (RN), which make up the Alianza por Chile opposition coalition. As happens with Concertación, the parties of Alianza por Chile also display a high level of ideological cohesion. Each of the coalitions takes up an extremity of the spectrum, thus demonstrating the polarization of party preferences in Chilean foreign policy.

Still on the subject of Table 5 and Graph 1, adopting the mean is more adequate when the distribution of the data is symmetrical, while the use of the median is more accurate in asymmetrical distributions. In the case of the Chilean parties, there is a small or inexistent difference between the mean and the median of their ideological positioning, leading to the conclusion that the distribution of Chilean deputies is symmetrical. This means that there is a low incidence of outliers, or that these values are not significantly altering the mean. The latter option seems more adequate, inasmuch as Graph 1 indicates the presence of outliers in these parties' distribution. Lastly, we observe that the Chilean party with the highest level of ideological cohesion is the DC, as measured by the IQR.

As for independent Chilean deputies, the first relevant observation is their tendency to occupy the right of the ideological spectrum. For example, in the 2002-2006 legislature, represented by Graph 1, we can see that the values considered outliers are situated on the left of the ideological spectrum. Moreover, one finds that the median position of the values is very close to the upper quartile, indicating that half the observations tend be situated on the right. Hence, in the case of the independents, one notices a greater difference between the mean and the median of the ideological values (always in comparison with the others), revealing an asymmetrical distribution, displaced to the right.

Lastly, we observe on Table 5 and Graph 1 the low level of ideological cohesion among independent deputies. This is an expected and significant result, inasmuch as independent deputies are not members of political parties and therefore lack the accompanying ideological constraints. This item of information bolsters the remark made here about the strong influence of party ideology on the votes of Chilean deputies on foreign policy issues.

In the distribution of Argentinian political parties, represented by Table 6 and Graph 2, we find on the left of the ideological spectrum the Afirmación para una República Igualitária (ARI), while the Unión Cívica Radical (UCR) takes up the centre ground and the PJ (Partido Justicialista), the right of the spectrum. Unlike what happens in the Chilean case, the distance between the opposition political parties (ARI and UCR) and the governing party (PJ) is not significantly large. In general, the parties' ideological medians are relatively close when compared with the Chilean case, indicating a tendency of convergence of the preferences of Argentinian political parties on foreign policy issues.

In keeping with what happens in the Chilean case, in general Argentinian parties exhibit high levels of internal ideological cohesion. This is borne out by the low standard deviations and IQRs on Table 6. Graph 2 reveals that the UCR and the PJ display high levels of ideological cohesion, as demonstrated by their low standard deviations and IQRs. This fact notwithstanding, these same parties have the most outlier deputies, who tend to be situated to the right of the spectrum. The ARI, though displaying a certain level of ideological cohesion among its deputies, in comparative terms has the lowest level of ideological cohesion of all the political parties in the sample, including both Chilean and Argentinian parties.

In counterposition with the Chilean case, in Argentina the centre of the ideological spectrum is taken up by an important political party, the UCR. In addition, the distance from the centre of the other two major Argentinian parties (ARI and PJ) is smaller than that exhibited for foreign policy issues by the Chilean parties. The ARI, the Argentinian party more to the left of the spectrum, keeps a distance of 0.39 from the centre, while the DC, the party of Chile's Concertación coalition closest to the centre has a distance of 0.45 from the centre. On the other side of the spectrum, while the PJ is located 0.31 from the centre, UDI, the party of the right-wing coalition Alianza por Chile closest to the centre is situated 0.48 away from the centre.

With respect to the set of Argentinian parties with less electoral strength, termed distritales, one finds what is expected: low ideological cohesion. This is because this category encompasses a considerable number of parties, of which many only have a regional (provincial) reach. The median of their positioning is displaced to the right, very close to the ruling party's median (0.29 for the distritales and 0.28 for the PJ).

Foreign policy versus domestic policy

One of the theses present in the Latin American literature maintains that the president will obtain more space for action on the international arena when compared with the domestic arena, i.e., the legislators' preferences would be manifested and relevant only in the domestic arena. In this section we make a comparison between the preferences of the parties, estimated for questions of foreign policy and domestic policy. To this end, we use the ideological classification of Chilean and Argentinian political parties produced by a survey conducted by Proyecto Élites Latinoamericanas (PELA) (2002), of the University of Salamanca.

In the case of the Argentinian political parties, the ideological classification was extracted from data produced by the PELA survey conducted between 2003 and 2007. Basically, we use the question in the survey that asks deputies to classify the main Argentinian political parties in an ideological scale from 1 (left) to 10 (right). The mean of the values chosen by the legislators of a given party represents that same party's position on the ideological spectrum.¹⁷ The means and medians obtained situate the ARI more to the left of the spectrum, displaying the values of 3.4 and 3 respectively. Located at the centre of the ideological spectrum is the UCR, with 5.79 and 6. Slightly to the right, one finds the governing party (PJ), with values of 6.5 and 7. The grouping of parties termed distritales was not considered in the analysis, for they are not included in the sample of the survey utilized here.

For the case of the Chilean parties, we use data from the PELA survey presented by Manuel Alcántara Saéz (2003) in an article titled La ideología de los partidos políticos chilenos, 1994-2002: Rasgos constantes y peculiaridades. The specificity of this measurement in comparison with that for Argentinian political parties is the incorporation of a larger number of questions to lawmakers and not just the ideological classification formulated by deputies about parties. Thus, the values estimated also include questions referent to democracy, the role of the armed forces and of the State in the promotion of social well-being. The ideological scale is the same as the one used for Argentinian parties, where 1 is left and 10 is right.

Graph 4 shows that the parties that make up the Concertación governing coalition occupy the left and centre of the ideological spectrum, where the party that is more to the left is the PS (2.56), followed by the PPD (3.85) and the PDC (5.28), more to the centre. The parties of the opposition coalition Alianza por Chile, for their part, take up the right of the spectrum: the RN has 6.8 and the UDI, 7.08.

Based on the values mentioned above, we are able to compare the distribution of Argentinian and Chilean political parties' preferences in the domestic and foreign arenas, empirically assessing the consistency of the "two presidents thesis". Graphs 5 and 6 represent the comparison between the positioning of Argentinian and Chilean parties on domestic and foreign policy matters, respectively.

Graphs 5 and 6 reveal an initial, evident aspect: the high correlation between the domestic and international arenas as far as Argentinian and Chilean parties are concerned, the figures being 0.95 and 0.90, respectively. Argentinian and Chilean parties basically maintain the same position on the ideological spectrum with reference to foreign and domestic policy, with their ordering on the same spectrum remaining constant.

The ordering of Chilean political parties on the ideological spectrum estimated via Nominate for foreign policy coincides with that produced via the survey for domestic policy. While the PS, PPD and DC follow the same dispersion, there is a slight change with respect to the UDI and RN. The UDI is the party most to the right of the spectrum as estimated by Nominate, while the RN takes up this position with regard to the domestic arena. Even then, the fact that both are situated on the right half of the spectrum coincides.

As far as Argentinian parties are concerned, the ARI is estimated to lie on the left of the ideological spectrum in the foreign and domestic dimensions, while the UCR's position is estimated in the centre of the spectrum, also for both dimensions. The PJ, in turn, occupies the centre-right of the spectrum in the foreign and domestic dimensions.

Beyond the similarity mentioned above, there is another similarity between the ideological classifications of Argentinian and Chilean political parties in the domestic and international dimensions: the low level of polarization in the Argentinian case and the high level of polarization in the Chilean case. As Mustapic (2000) argues, effective Argentinian political parties have a small ideological distance between each other, i.e., from a general public policy perspective, one observes a low ideological polarization between the main Argentinian parties, and the same occurs with reference to foreign policy (as previously argued).

As for the Chilean case, Carey (2002) created an index to measure the level of unity among Chile's political parties and coalitions by analysing 215 roll call votes in the Chamber of Deputies over two years, 1997 and 1998. The results of the index indicate major proximity among parties that belong to the same coalition. On the other hand, between coalitions there is a high level of polarization, indicating positions that diverge from those of the median voter (Carey 2002). There is significant coincidence between our findings as to foreign policy (analysed in isolation) and those of the authors just mentioned. Alemán and Saiegh (2007) also analysed roll call votes to ascertain political parties' positioning during the 1997-2000 period. The authors estimated Chilean deputies' ideal points, revealing results similar to those found both by Carey (2002) and by ourselves, presented here with reference to foreign policy; in other words, great intra-coalition cohesion and a clear bipolar situation between the two political coalitions.

In order to illustrate the level of ideological polarization, as previously argued, one may consider the difference between the distances of the more extreme parties in the two dimensions. In the Argentinian case, the difference between the ARI and the PJ in the domestic ambit is 0.62, and in the foreign ambit, 0.70. In the case of Chile, this difference rises to 1 and 1.43, respectively.

In sum, both in the Argentinian case and in the Chilean, it is possible to argue that there is a strong similarity between the political parties' ideological classifications in the domestic and foreign dimensions, suggesting that the factors structuring their preferences in the domestic dimension are the same that structure their preferences in the foreign domestic, with party ideology standing out.

Conclusion

This article has sought to analyse the distributions of Argentinian and Chilean legislators' preferences in foreign policy themes, each in two legislatures starting in 2002. The fundamental motivation of the research was understanding to what extent the argument that foreign policy — being less crucial to legislators' interests than domestic matters — would be prone to negligence and abdicative postures on congresspersons' part. The comparison was undertaken on two main dimensions. The first related precisely to the distribution of party-political preference on foreign policy matters, especially along the left-right ideological spectrum, the most relevant differentiating factor of legislators' positioning. The second dimension related to the linearity between the party-political ideological classifications when confronting the countries' foreign and domestic policies.

With respect to the first aspect, we have found a clear distinction between the Argentinian and Chilean dynamics. The former has a distribution of preferences with very little polarization, with legislators tending to concentrate their preference in the centre of the political spectrum when it comes to foreign policy matters. Furthermore, the Argentinian Legislative records a larger number of discrepant preferences. Therefore, Argentina combines preferences that converge to the centre with marginally discrepant positioning.

Chile, on the other hand, has a clearly polarized distribution of preferences. Such polarization coincides in grouping together in opposite sides of the spectrum the opposition-right and the governing-left. In the Chilean case, there exist high levels both of party discipline and of coalitional discipline as far as foreign policy issues are concerned. Also differently from Argentina, Chile features a low index of discrepant behaviours. Put differently, when examining Chilean legislators' distribution of preferences one detects few outliers.If Argentina and Chile differ substantially with regard to the pattern of legislators' distribution of preferences on foreign policy matters, the same cannot be said of the correlation between preferences on domestic and foreign policies. In both countries there is a high correlation between parties' positioning with respect to domestic and foreign policies. In other words, when we use foreign policy issues to classify the countries' positioning, the results in both cases differ little from what occurs when the classification is arrived at by means of domestic questions.

Therefore, the conclusion is that the two countries' party-political structuring is not distinctive, when comparing the foreign and domestic ambits. On the contrary: the positioning on foreign policy matters does nothing more than reflect the structuring of preferences in the domestic ambit. In Argentina, positions on matters of foreign policy are centralized, as are positions on matters of domestic policy. By the same token, in Chile, legislators' positions regarding foreign affairs are polarized, as are their positions on domestic matters.

Clearly, the two cases do not permit any definitive inferences that can be taken at face value in an effort to theorize on the specificity (or lack thereof) of South American or Latin American countries' foreign policies. Only a more exhaustive comparative study — one that, beyond legislators' preferences, takes into account other factors that come together in the formation of political parties' orientation — can make a more consistent contribution to theorization efforts. Nevertheless, it is hoped that this article's findings provide some momentum in this direction.

Notes

1 It is worth highlighting the fact that the findings of Bond and Fleisher (1990) indicate that only Republican presidents possess significant congressional support in the fields of foreign policy and national defence.

2 It is necessary to stress that some scholars point to a relevant participation by the US Congress and political parties already from the start of the Cold War; see McCormick and Wittkopf (1990) and Fordham (1998), for example.

3 About thematic and media influence in the formulation of the president's foreign policy agenda in the US political system, see Wood and Peake (1998).

4 Oliveira (2003) points out three other advantages of the Legislative's foreign policy action: decision-making decentralization diminishes the possibility of mistaken policies being formulated; the increase in institutional constraints, via the Legislative, can act as bargaining instrument on the international plane; and the production of information on the Legislative's part.

5 For a view that minimizes the impact of lawmakers' constituency and ideology on US trade policy issues, see Biglaiser, Jackson and Peake (2004). The authors maintain that depending on the political party of the President, the variables constituency and ideology can lose importance, fundamentally among Republican representatives.

6 Here the authors implicitly refer to the theory that the Executive, by virtue of being accountable to the electorate at the national level, privileges policies that benefit the general well-being. In the US literature, this gets translated into a defence of trade liberalism. On the other hand, legislators benefit localized interests (notably protectionist interests) owing to the fact that they are responsive to a local electorate (Milner and Rosendorff 1997).

7 A metric space is one where legislators' spatial position can be measured quantitatively (and not just qualitatively).

8 Despite the fact that other estimation models differ as to the choice of the distribution of errors (logistical, uniform or normal), all of them assume that the errors are independent and identically distributed (Rosenthal and Voeten 2004). According to Rosenthal and Voeten (2004), in legislatures where there is strong party discipline and great variability in the level of party loyalty among political parties, in parametric models of estimation, such as Nominate, there can be violation in the assumption of independent and equal distribution of errors.

9 For a normally distributed sample database (an assumption of Nominate), the technique of maximum verisimilitude identifies the values of the parameters of the mathematical model that display greater verisimilitude in relation to the data. In other words, it is a technique that selects the values of the parameters of the sample that best "fit into" the population that the sample represents. When one assumes that the data of the sample are uniformly distributed, the estimation of maximum verisimilitude coincides with the estimation of the likeliest values, i.e., those values of the parameters of data whose probability of occurrence is the largest possible, get selected (Aldrich 1997).

10 The Argentinian case displays a peculiarity in relation to the Chilean. In Argentina, half the seats in the Chamber of Deputies are renewed every two years. Furthermore, some parliamentary terms are for two years, while others are for four. The total number of seats is 256, but in the present sample this number jumps to 618, for it includes all the deputies who voted in the plenary during the 2002-2007 period.

11 Votes during the 2002-2006 legislature in Chile were obtained from the website of the Chilean Chamber of Deputies (www.camara.cl). Votes during the 2002-2007 legislatures in Argentina were obtained from the website of the Argentinian Chamber of Deputies (www.diputados.gov.ar).

12 It is important to stress that NOMINATE only considers "Yes" and "No" votes, including abstentions as "Did not vote".

13 The meaning of left, centre and right certainly depends on the historical and cultural context; it is only necessary for a large part of the actors to share this meaning (Leoni 2000).

14 The mean is obtained from the addition of the values, divided by the number of values observed. This measure summarizes the set of data in terms of a central position or typical value; it is very appropriate in situating the party in the ideological dimension.

15 The median evaluates the centre of a set of values, dividing the distribution of the data in half, i.e., leaving the 50% lower values on one side and the 50% higher values on the other.

16 Standard deviation supplies information about the variance or heterogeneity of the values analysed. In this case, the standard deviation acts as an excellent measure of the level of deputies' ideological cohesion or homogeneity within the same party on matters of foreign policy.

17 It is worth stressing the fact that other measures, like the median and the mode, were also taken, with extremely similar outcomes.

Bibliographical References

Aldrich, John. 1997. R. A. Fisher and the making of maximum likelihood 1912-1922. Statistical Science 12 (3): 162-76.

Alemán, Eduardo, and Sebastián Saiegh. 2007. Legislative preferences, political parties, and coalition unity in Chile. Comparative Politics 39 (3): 253-72.

Alexandre, Cristina V. M. 2006. O Congresso brasileiro e a política externa (1985-2005). Master's diss., Postgraduate Programme in International Relations, Catholic University of Rio de Janeiro.

Baldwin, Robert and Christopher Magee. 2000. Congressional trade votes: From NAFTA approval to fast-track defeat. Washington: Institute for International Economics.

Bardwell, Kedron. 2000. The puzzling decline in house support for free trade: Was fast-track a referendum on NAFTA? Legislative Studies Quarterly 25 (4): 591-610.

Biglaiser, Glen, David Jackson, and Jeffrey Peake. 2004. "Back on track": Support for presidential trade authority in the House of Representatives. American Politics Research 32 (6): 679-97.

Bond, Jon, and Richard Fleisher, 1990. The president in the legislative arena. Chicago: University of Chicago Press.

Carey, J. M. 2002. "Parties, coalitions, and the Chilean Congress in the 1990s." Legislative Politics in Latin America. In Legislative Politics in Latin America, ed. Scott Morgenstern and Benito Nacif. Cambridge: Cambridge University Press.

Conley, Richard. 1999. Derailing presidential fast-track authority: The impact of constituency pressures and political ideology on trade policy in Congress. Political Research Quarterly 52 (4) 785-99.

Delaet, James, and James Scott. 2006. Treaty-making and partisan politics: Arms control and the U.S. Senate, 1960-2001. Foreign Policy Analysis 2 (2): 177-200.

Edwards, George. 1989. At the margins: presidential leadership of Congress. New Haven: Yale University Press.

Feliú, Pedro; Manoel Galdino, and Amâncio Oliveira. 2007. Política de comércio exterior, ideologia partidária e interesses locais: Um estudo sobre o caso chileno. Revista Cena Internacional 9 (2): 33-57.

Fleisher, Richard, Jon Bond, Glen Krutz, and Stephen Hanna. 2000. The demise of the two presidencies. American Politics Quarterly 28 (1): 3-25.

Fordham, Benjamin. 1998. Economic interests, party and ideology in early cold war era. International Organization 52 (2): 359-96.

Fordham, Benjamin, and Timothy McKeown. 2003. Selection and influence: Interest groups and congressional voting on trade policy. International Organization 57: 519-49.

Heckman, J. and Snyder, J. 1997. Linear probability models of the demand for attributes with an empirical application to estimating the preferences of legislators. RAND Journal of Economics 28: 142-89.

INTER-AMERICAN DEVELOPMENT BANK. 2006. A política das políticas públicas: Progresso econômico e social na América Latina. Rio de Janeiro: Editora Campus.

Jackman, S. 2001. Multidimensional Analysis of Roll Call Data via Bayesian Simulation: Identification, Estimation, Inference and Model Checking. Political Analysis 9, No. 3, pp. 227-41.

Jones, Mark. 2002. "Explaining the high level of party discipline in the Argentine Congress". In S. Morgenstern, and B. NACIG, (eds.), Legislative politics in Latin America. Cambridge/New York: Cambridge University Press.

Kalandrakis, Tasos. 2006. Roll call data and ideal points. Working Paper 42, University of Rochester.

Kegley, Charles and Eugene Wittkopf. 1995. American foreign policy - Pattern and process. New York: St. Martin's Press.

Leão, Flávio. 2008. Poder legislativo e política externa: Um estudo sobre a influência do Congresso Nacional do Chile na formulação da política comercial durante a década de 1990. Contexto Internacional 30 (1): 89-139.

Leoni, Eduardo. 2000. Ideologia, democracia e comportamento parlamentar: A Câmara dos Deputados no Brasil pós-constituinte. Master's dissertation, Department of International Relations, University of Brasília.

Lima, Maria Regina Soares and Fabiano Santos, 2001. O Congresso e a política exterior. Lua Nova, no. 52: 121-50.

Lindsay, James. 1994. Congress and the Politics of US Foreign Policy. Baltimore: The Johns Hopkins University Press.

Lindsay, James and Randall Ripley. 1992. Foreign and defense policy in Congress: A research agenda for the 1990s. Legislative Studies Quarterly 17 (3): 417-49.

Marks, Gary, Liesbet Hooge, Moira Nelson, and Erica Edwards. 2006. Party competition and European integration in the East and West: Different structure, same causality. Comparative Political Studies 39 (2): 155-75.

Martin, Lisa L. 2000. Democratic commitments: Legislatures and international cooperation.Princeton: Princeton University Press.

McCormick, James, and Eugene Wittkopf. 1990. Bipartisanship, partisanship, and ideology in congressional-executive foreign policy relations, 1947-1988. The Journal of Politics 52 (4): 1077-100.

___, and David Danna. 1997. Politics and bipartisanship at the water's edge: A note on Bush and Clinton. Polity 30: 132-49.

Meernik, James. 1993. Presidential support in Congress: Conflict and consensus on foreign and defense policy. The Journal of Politics 55 (3): 569-87.

Mena, Antonio Ortiz. 2004. O processo de formulação política de comércio nas américas: Lições da experiência mexicana. Paper presented at the 45th Annual Convention of the ISA, March, in Montreal, Canada.

Milner, Helen, and Peter Rosendorff. 1997. Democratic politics and international trade negotiations: Elections and divided government as constrains on trade liberalization. Journal of Conflict Resolution 41 (1): 117-46.

Milner, Helen, and Benjamin Judkins. 2004. Partisanship, trade policy, and globalization: Is there a left-right divide on trade policy? International Studies Quarterly 48: 95-119.

Mustapic, Ana María. 2000. Oficialistas y diputados: Las relaciones Ejecutivo-Legislativo en la Argentina. Desarrollo Económico 39 (156).

Neves, J. A. Castro. 2003. O papel do legislativo nas negociações do Mercosul e da ALCA. Contexto Internacional 25 (1).

Nolte, Detlef. 2003. El Congreso chileno y su aporte a la consolidación democrática en perspectiva comparada. Revista de Ciencia Política XXIII (2): 43-67.

Oliveira, Jorge Amâncio. ed. 2005. Entraves e possibilidades institucionais e políticos da integração regional sul-americana. OPSA Report, September.

___. 2003. Legislativo e política externa: Das inconveniências da abdicação. Caeni, No. 3, São Paulo. Manuscript.

Parish, Randall. 2004. Presidents, parties, and foreign policy: Domestic institutions and intestate behavior in Latin America. Paper presented at the 2004 Annual Meeting of the Midwest Political Science Association at Chicago, April 15-17, in Illinois.

Poole, Keith T. 2005. Spatial models of parliamentary voting. Cambridge: Cambridge University Press.

Prins, Brandon C., and Bryan W. Marshall. 2001. Congressional support of the president: A comparison of foreign, defense, and domestic policy decision making during and after the cold war. Presidential Studies Quarterly 31 (4): 660-78.

Proyecto Elites Latino Americanas - PELA. 2002. Base de datos. http://americo.usal.es/oir/Elites/.

Ragsdale, Lyn. 1995. President and Congress. New York: Simon & Schuster Press.

Ripley, Randall, and James Lindsay, ed. 1993. Congress resurgent: Foreign and defense policy on Capitol Hill. Ann Arbor: The University of Michigan Press.

Rosenthal, Howard and Erik Voeten. 2004. Analyzing roll calls with perfect spatial voting: France 1946-1958. American Journal of Political Science 48 (3): 620-32.

Saéz, Manuel Alcántara. 2003. La ideología de los partidos políticos chilenos, 1994-2002: Rasgos constantes y peculiaridades. Revista de Ciencia Política XXIII (2).

Santos, Fabiano. 2006. Democracia, Legislativo e os novos desafios para o capitalismo brasileiro. Análise de Conjuntura OPSA, no. 9, September.

Stuhldreher, Amalia. 2003. Régimen político y politica exterior en Argentina: El caso del primer gobierno de Carlos Menem (1989-1995). Perfiles Latinoamericanos, no. 22 (June).

Thérien, Jean-Philippe, and Alan Noel. 2000. Political parties and foreign aid. The American Political Science Review 94 (1): 151-62.

Vásquez, David. 2006. El sistema electoral binominal y las elecciones parlamentarias 1989-2005. Serie Estúdios, no. 1. Santiago: Biblioteca del Congreso Nacional de Chile.

Warburg, Gerald F. 1989. Conflict and consensus: The struggle between Congress and the president over foreign policymaking. New York: Harper & Row.

Wittkopf, Eugene R., and James M. McCormick. 1998. Congress, the president, and the end of the cold war: Has anything changed? The Journal of Conflict Resolution 42 (4): 440-66.

Wood, Dan B., and Jeffrey S. Peake. 1998. The dynamics of foreign policy agenda setting. American Political Science Review 92 (1): 173-84.

Xie, Tao. 2004. Congressional roll call voting on China trade policy. American Politics Research 32 (6): 679-97.

* A preliminary version of this paper was presented at the International Congress of the Latin American Studies Association (LASA), from 11 to 14 June, 2009, in Rio de Janeiro. We thank Rosana Miranda for her help in feeding the database, and the researchers of the Centre for the Study of International Negotiations of the University of São Paulo (CAENI/USP) for their comments.