The weight of ideology on the attitude of Latin American countriestoward the United States

The article tests the effect of ideology on the attitude of Latin American countries toward the United States, as well as alternative explanations, to respond to the expectation that left-wing governments are critical of the US and right-wing governments are friendly. The findings are that the alternative explanations are less relevant and that ideology has the expected effect.


Introduction
I t is often expected and suggested, due to past US actions against left-wing governments in Latin America, that the ideological position of ruling parties in the region foretells the attitude of their administrations toward the United States. In this sense, left-wing governments in Latin America would be critical of the US, whereas right-wing governments would be friendly.
The purpose of this article is to examine the effects of four variables, including two ideological variables, on the vote coincidence of seventeen Latin American countries (Argentina, Brazil, Bolivia, Colombia, Costa Rica, Chile, Ecuador, Guatemala, El Salvador, Honduras, Mexico, Nicaragua, Panama, Paraguay, Peru, Uruguay and Venezuela) with the US in the United Nations General Assembly (UNGA), which was considered as the indicator of the attitude of these Latin American countries toward the United States. The two ideological variables considered were the location of the ruling party and the president in the left-right scale, according to the evaluation of legislators from each country per period of government. Two other variables were also tested in order to evaluate alternative explanations about the attitude of Latin American nations toward the US.

Copyright:
• This is an open-access article distributed under the terms of a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided that the original author and source are credited.

Revista Brasileira de Política Internacional
Another alternative explanation can be drawn from the neorealist theory. According to Waltz (1979), the military and economic powers are the main determinants of international action of the States, which are unitary actors who, based on their relative capabilities, establish their preferences and make rational decisions. Under the different configurations of the international system, the options of States would be limited to either joining forces to counterbalance the most powerful ones or letting themselves be conducted by the most powerful ones. In the case of the first option, it would be possible the joining of forces among two or more States or the mobilization of internal resources to resist the most powerful ones.
Applying these postulates to the relations between Latin American nations and the US, it could be considered that the Latin American countries which have more national capabilities would be more willing to resist and vote in a divergent manner with the United States in the UNGA.
This article adopts a larger universe of cases than in the literature that addressed to some extent the influence of ideology in the attitude of Latin American countries toward the United States, and it tests, then, the hypothesis that left-wing governments in Latin America are critical of the US and right-wing governments are friendly, as well as the alternative explanations.

Methodology
The attitude of Latin American countries toward the United States, the variable to be explained, is measured by the simple convergence indicator in the UNGA. According to this indicator, there is convergence when the votes given by two countries are equal, in the case of this study, each one of the seventeen Latin American nations chosen and the US. The annual report made by the US Department of State, which provides the data, assesses the coincidence proportion with all countries represented in the cases that the United States voted "Yes" or "No".
The decision to adopt vote coincidence in the UNGA as indicator of the attitude of Latin American countries toward the United States is supported by authors who consider that these votes are a valid indicator of foreign policy orientation (Tomlin 1985;Thacker 1999;Voeten 2000). Among those authors that resort to this procedure, the measures that stand out are the concordance index, from Lijphart (1963), and the simple convergence indicator, utilized here. As for Amorim Neto (2011, 64-65), the justification of this choice is that the US Department of State, diplomats and formulators of foreign policy, including those from Latin America, adopt this measure.
In the case of explanatory variables, the locations of the ruling party, understood as the president's party, and that of the president himself in the left-right scale are the indicators of the ideological position of each government, according to the evaluation of respective national legislators. The data source is the Observatorio de Élites Parlamentarias de América Latina, a project from the University of Salamanca, in Spain, which performs interviews with parliamentarians of each legislature in a group of Latin American countries. Among the questions, legislators are asked to locate others parties than their own and politicians in the left-right scale, considering the left as 1 and the right as 10. The location in the left-right scale is widely accepted and utilized in the literature on parties and party systems in Europe (Sani and Sartori 1983;Mair 1997) and in Latin America (Coppedge 1997(Coppedge , 1998Colomer and Escatel 2005). Coppedge (1997, 10-12), who does an extensive classification of Latin American political parties over time and countries, recognizes that the left-right dimension can be problematic in Latin America, but justifies its use for several reasons, including the fact that the specialists consulted on each country were able to apply it for the vast majority of the cases and, therefore, certified the importance of this criteria.
The classifications applied here come from interviews with legislators, who are asked to classify other parties and politicians in a scale in which 1 is left and 10 is right, with no extra guidance on what each category means. The choice made in this work to use the locations of the ruling party and the president as indicators of the ideological position of the respective government and, consequently, as explanatory variables tested occurs due to the fact that the interviews with Latin American legislators offer, most of the time, answers to both questions.
For measuring the degree of dependence on loans from the international credit organisms, one of the alternative explanations considered for the attitude of Latin American countries toward the United States, the proportion of external debt of each nation is used, chiefly focusing it in relation to the country's Gross Domestic Product (GDP) 1 .
A more difficult task is to measure the national capabilities of a country, the other alternative explanation to be considered. For this, a known tool is used: the Composite Index of National Capability (CINC), originally elaborated by Singer et al. (1972). This index measures the quota of a State, in relation to the world totals, of six resources considered military relevant, iron and steel production, military expenditures, military contingent, primary energy consumption, total population, and urban population 2 .
The selected countries are all of those in which the Observatorio de Élites Parlamentarias de América Latina has accomplished at least one round of interviews, except for the Dominican Republic, once this article has opted not to include the Caribbean countries. The period studied extends from 1995 to 2013, due to the availability of data for the variables, especially the interviews with Latin American legislators, since the research project on parliamentarian elites of Latin America began to be performed in 1994.
For the regression analysis, with panel data, the ideological variables were separated and tested in models which also included the other two explanatory variables, as a way to avoid the multicollinearity, since the locations of the ruling party and the president are highly correlated. Each set of variables was tested in three models, pooled, with fixed effects and with random effects.

Results
Between the years of 1995 and 2013, in the seventeen countries, there were ninety-six governments either complete or incomplete due to term interruptions, temporariness of the president, term beginning before the year of 1995 or continuity until a posterior year to 2013. From these ninety-six governments, seven were ruled out due to their extremely short duration which conveyed no participation in the UNGA 3 .
In the studied period, there was a high variation in vote coincidence with the United States, which reached the peak of 68.8% with Argentina under Carlos Menem in 1995, and dropped to the limit of 6.3% with Venezuela under Hugo Chávez in 2007. These two years, 1995 and 2007, are also the ones that present, respectively, the highest (46.9%) and the lowest (12.1%) convergence means for the group of the seventeen countries (see table 1). The trend of this mean was a drop between 1995 and 2007, after which a rise started, reaching 45.4% in 2011. (1) Refers to the coincidence proportion when the United States voted "Yes" or "No" Source: US Department of State.
3 Formally, there are four more governments that have not been accounted, three from Venezuela and one from Ecuador. In the Venezuelan case, two governments had, each, the duration of one day, one de facto, from the Coup d'état against Hugo Chávez in April of 2002, and one temporary, substituting the latter. Since Chávez continued his term and never resigned, these two governments have not been considered in the total number. There is still a first term by Chávez, which began in 1999. With the Constitution of 1999 coming into force, new presidential elections were held in 2000, which gave Chávez a six-year term. For practical purposes, the period between 1999 and 2007 has been counted as one single government. As the same situation occurred in Ecuador, it was also applied to Rafael Correa in the period between 2007 and 2013. Despite its condition of main ally of the United States in South America, Colombia also registered a vigorous drop regarding vote coincidence with the United States between 1995 and 2007. In fact, the third lowest convergence of all period between 1995 and 2013, of 7.4%, was with Álvaro Uribe's Colombia in 2007. The Colombian mean in this entire period was only higher than those of Bolivia, Ecuador and Venezuela.
Both the drop trend in the vote coincidence with the United States until the second half of the 2000's and the one of recovery since then are maintained when the analysis unit becomes the governments. However, the indexes present significant variation between countries in each of these two moments. Therefore, if, on the one hand, these trend suggest that the foreign policy of George W. Bush (2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009) and its reaction in Latin America may be the explanation for the vigorous drop in the vote coincidence with the United States, on the other, the significant variations in the indexes of countries at the same moment, of drop or rise, suggest that there are other explanations, such as ideology.
These classifications are the means obtained from the answers of legislators interviewed in each round of the Proyecto Élites Parlamentarias Latinoamericanas when questioned about the ideological location of political parties and politicians in the left-right scale. The number of rounds varies according to the country. In many cases, new interviews were performed in each new legislature since the first half of the 1990's. In others, the interviews started later. The question on ideological location of some political parties is present in the application of questionnaires done in every country. In it, the interviewed legislators does not classify their own party. For the countries in which there were no rounds of interviews with parliamentarians from all legislatures between 1995 and 2013, it was utilized the classifications which show in the interviews with parliamentarians from other legislature which do not coincide with that government. As for the question on the location of certain politicians, it is not present in some few questionnaire applications or it was done, however it does not include the president between the politicians mentioned. In these cases, the solution was to repeat for the president the classification received for his political party. There are few cases in which the classifications of the ruling party and the president are very divergent (see table 3). Argentina is an example, with Néstor Kirchner's government (2003)(2004)(2005)(2006)(2007), as the president was classified as left-wing, whereas his party was located on the right-wing. Another example comes from Guatemala, with Alfonso Portillo's term (2000)(2001)(2002)(2003)(2004). In this case, the highest discrepancy is not in the orientations, once the president was classified in the center and his party on the right, however, it is in the punctuations received by each one. Portillo received 5.30, whereas his party received 8.94.  (1992 -1996) Bucaram (1996 -1997) Mahuad (1998 -2000) Noboa (2000 -2003) (  Calderón (1994Calderón ( -1999 (2) (1999 -2004) (  (1993 -1996) Arzú (1996 -2000) Portillo (2000 -2004) Berger (2004 -2008) (1994 -2000) (2) Fox (2000 -2006) ( In the regression analysis, the ruling party's ideology and the president's ideology, tested separately with the other two explanatory variables to avoid the multicollinearity, were statistically significant in the expected direction (see tables 4 and 5). After doing the comparative tests between the models, pooled, with fixed effects and with random effects, those which resulted to be the most appropriate were the models with fixed effects. In these models, apart from the ideological variables already mentioned, the national capabilities of each Latin American country was statistically significant, but in the opposite direction to that expected, since the result observed was that the greater the national capacity, the greater the vote coincidence with the United States in the UNGA.  For an increase of 1.0 in the ruling party's ideology, the vote coincidence with the United States rises 1.09 percentage point. That is, the more the ruling party of a Latin American country is to the right, the greater the convergence of his government with the United States in the UNGA. The model explains 6.26% of the vote coincidence with the US.

Flores
In the case of the president, for an increase of 1.0 in his ideology, the vote coincidence with the United States rises 1.40 percentage point. This means that, as well as the ruling party, the more the president of a Latin American country is to the right, the greater the convergence of his government with the United States in the UNGA. The model explains 7.51% of the vote coincidence with the US.
If the variables are transformed in logarithms, as a way to standardize them, since the values for the dependent variable are expressed in percentages, the ruling party's ideology and the president's ideology remain statistically significant in the expected direction (see tables 6 and 7). In turn, the variable associated with the national capabilities of each country ceases to present significance. With the variables transformed in logarithms, the models that resulted to be the most appropriate, after the comparative tests, were the pooled ones.

Deviant cases
The analysis of deviant cases enables the identification and explanation of exceptions to the trend found, that is, Latin American governments which had a level of convergence with the US in the United Nations General Assembly greater than one would expect based on the ideology of who governs.
Verifying the average coincidence per administration and the ideological locations of the respective presidents and ruling parties, three deviant cases were selected for a qualitative analysis: a left-wing government which had a convergence higher than expected, the one of José Mujica (2010)(2011)(2012)(2013)(2014)(2015) in Uruguay; and two right-wing administrations whose coincidence with the United States in the UN was below than expected, both of them from Alvaro Uribe (2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010) in Colombia.
In the case of Mujica, from Frente Amplio (FA), the average convergence was 42.1%, despite the president and his party having received scores of, respectively, 3.26 and 3.71, which mean locations on the left. This proportion of coincidence takes into account four of the five editions of the UNGA in which the participation of Uruguay was under the responsibility of the Mujica's government, which started in March 2010. The last Uruguayan participation under the administration of the former president was not computed because the data on the convergence of positions in the UN covers until 2013.
The FA attempted an approximation with the United States since its first government in Uruguay, with Tabaré Vázquez (2005-2010, during which the possibility of signing a free trade agreement with the US was discussed in the governmental and partisan levels. In the face of resistance, including the one offered by the other members of the Southern Common Market (MERCOSUR), since it would affect its status as a customs union, which establishes the joint negotiation of trade agreements with other countries or blocs, the FTA with the United States was not materialized in the administrations of Vázquez nor Mujica. Yet a Bilateral Investment Treaty (BIT), which was signed at the end of the government leading up to Vázquez, came into force in 2006 and was complemented by the signing of a Framework Agreement of Trade and Investment in 2007.
The BIT with Uruguay was part of George W. Bush's strategy of offering bilateral trade agreements to Latin American countries after the failure of the negotiations of the Free Trade Area of the Americas (FTAA). At the same time, the foreign policy of his government generated a negative reaction in the region, which reflected in the drop of vote coincidence with the US in the United Nations General Assembly. This factor and ideology would help to explain why, despite the attempted approximation, the Vázquez administration had a convergence of 23.4% with the United States in the UN. As Bush had already finished his second term when Mujica took office in 2010, the new Uruguayan president could face less resistance to increase the coincidence of positions with the US, achieving a rate of 42.1%.
In the case of Uribe, who ruled Colombia for two consecutive periods without being formally affiliated to a party, the average convergence for the first and second terms were, respectively, 17% and 19%, despite having received scores of 8.17 and 7.96, which means a locations on the right. As the ruling party, the ones considered were, for the first term, Uribe's group of the Partido Liberal (PL), of which Uribe was part before becoming president, and, for the second term, the Partido de la U, which was created to represent the Uribe's political base. These forces have received scores of, respectively, 8.47 and 7.34, which also mean locations on the right.
The average coincidence of 17% with the US in the United Nations General Assembly for the first administration of Uribe was the second lowest level in the group of 17 Latin American countries throughout the studied period, only being ahead of the first government of Evo Morales (2006)(2007)(2008)(2009)(2010) in Bolivia, with 16.7%.
The Colombian case seems more contradictory and difficult to explain. One possible explanation is that the negative reaction in Latin America to Bush's foreign policy has included Colombia, at least with regard to the positions in the UN. At the same time, Uribe administration was an American ally in high politics issues, thus ensuring logistical and financial support of the United States for its policies to combat drug trafficking and insurgent groups. It was also under Uribe's government, more precisely at the start of his second term, that Colombia signed an FTA with the US. The agreement had to wait until 2012 to come into force, because of demands made by the American Congress to approve it.

Conclusion
The expectation that left-wing governments in Latin America are critical of the United States and right-wing governments are friendly has been confirmed. The ruling party's ideology and the president's ideology are statistically significant to explain the attitude of Latin American countries toward the US, measured by the vote coincidence between each of the seventeen Latin American nations which were studied and the United States in the UNGA. The ruling party's ideology is significant at the 10 percent level, whereas the president's ideology at the 5 percent level.
The relationship found among each of these ideological variables and the vote coincidence with the US had the expected direction, that is, the more the ruling party or the president of a Latin American country is to the right, the greater the convergence of his government with the United States in the UNGA. The model which included the ruling party's ideology and the other two variable tested explains 6.26% of the vote coincidence with the United States, whereas the model with the president's ideology explains 7.51% of the convergence.
The variable associated with the national capabilities of each Latin American country is also significant in both models, but in the opposite direction to that expected on the basis of neorealism, since the result observed was that the greater the national capacity, the greater the vote coincidence with the United States in the UNGA. The national dependence on external loans resulted not to be significant in both models.
If the variables are transformed in logarithms, the ruling party's ideology and the president's ideology remain statistically significant in the expected direction, respectively at the 5 percent and the 1 percent levels. In turn, the variable associated with the national capabilities is no longer significant in the most appropriate models with each of the ideological variables. The national dependence on external loans continues resulting not to be significant in both models.
Results corroborate the evaluation about the importance of the president in the foreign policy of presidentialist systems, mainly in Latin American countries, where, in most cases, the Head of State relies on relatively strong constitutional powers (Shugart and Mainwaring 2002). Even in the case of presidents with fewer constitutional prerogatives on policy and law formulation, there is the possibility of exerting influence via partisan powers, which depend on that the president is the leader of a party or coalition with legislative majority.
At the same time, as can be seen by the proportion of the dependent variable that the models tested explain, there are more factors that influence the attitude of Latin America countries toward the United States than ideology. Even about the trend found that left-wing governments are critical of the US and right-wing governments are friendly, there can be variations, as the analysis of the deviant cases has showed.
Finally, the results obtained serve as incentive for further studies on the determinants of the attitude of Latin American countries toward the United States: to seek other factors besides ideology which could be relevant explanations, and to continue testing the relationship found here, that is, the more the president or the ruling party of a Latin American nation is to the right, the greater the convergence of his government with the US.