Presidential Delegation to Foreign Ministries: A Study of Argentina, Brazil, and Mexico (1946–2015)

Journal of Politics in Latin America Creative Commons CC BY: This article is distributed under the terms of the Creative Commons Attribution 4.0 License (https:// creativecommons. org/ licences/ by/ 4. 0/) which permits any use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https:// us. sagepub. com/ enus/ nam/ openaccessatsage). Presidential Delegation to Foreign Ministries: A Study of Argentina, Brazil, and Mexico (1946–2015) Octavio Amorim Neto1 and Andrés Malamud2 Abstract When do presidents delegate policymaking authority to their foreign ministries? And is foreign policy unique in this respect? We posit that six international, national, and personal factors determine the opportunity and motivation of presidents to delegate, and then analyse the cases of Argentina, Brazil, and Mexico in 1946–2015. By applying fuzzyset qualitative comparative analysis, we find that four combinations of factors are sufficient paths to delegation: (1) international stability and elite consensus on foreign policy; (2) international stability, rightwing president, and low diplomatic professionalisation; (3) international stability, rightwing president, and low presidential expertise on foreign policy; or (4) absence of authoritarianism combined with elite consensus on foreign policy and rightwing president. Our study of foreign ministries reinforces some of the main findings of the scholarly literature on other ministries, thus challenging the view of foreign policymaking as different from domestic policy areas. Manuscript received: 30 December 2019; accepted: 30 June 2020 Keywords Argentina, Brazil, Mexico, Latin American politics, foreign policy, diplomacy, presidents, presidential diplomacy, fuzzyset QCA Journal of Politics in Latin America 2020, Vol. 12(2) 123–154 © The Author(s) 2020 Article reuse guidelines: sagepub. com/ journalspermissions DOI: 10. 1177/ 1866 802X 20944184 journals. sagepub. com/ home/ pla Research Article 1EBAPE – Fundação Getulio Vargas, Rio de Janeiro, Brazil 2Institute of Social Sciences, University of Lisbon, Portugal Corresponding Author: Andrés Malamud, Universidade de Lisboa, Instituto de Ciências Sociais, Av Prof Anibal de Bettencourt, 9, Lisboa, 1600-189, Portugal. Email: [email protected] Journal of Politics in Latin America 12(2)124 Motivation and Puzzle Why do chief executives delegate authority to professional bureaucrats instead of politicians, privatesector technocrats, interestgroup representatives, or cronies? This is a key question to understand policymaking in presidential regimes, given that presidents, as opposed to prime ministers, hold individual rather than collective responsibility for government.1 It is no wonder that there has been a growing comparative literature on the issue, with a strong focus on economic and social ministries in Latin America (Camerlo and PérezLiñán, 2015; Centeno and Silva, 1998; Dargent, 2014; Kaplan, 2017). Yet despite this growing literature, there is a dearth of comparative research on delegation to professional diplomats. This is puzzling, because professional diplomats rank among Latin America’s besttrained bureaucrats and are pivotal for the formulation and implementation of effective foreign policies. International relations scholars, for their part, tend to eschew the development of crosscountry operational indicators and focus on single cases instead (Belém Lopes, 2013; Bosoer, 2005; Cason and Power, 2009; Castro, 2009; Dalbosco, 2014). With few exceptions (such as Figueroa and Schiavon, 2014), comparative research on the relations between political leaders and professional diplomats in Latin America lies in limbo. The most we can say is that presidential ideology is the primary determinant of foreign policy change (Merke et al., 2020). But then, knowing whether presidents conduct foreign policy themselves or delegate it becomes crucial, as presidential ideology can get diluted, distorted, or even challenged in the delegation process. Why was Brazilian president Fernando Henrique Cardoso (1995–2002) mockingly nicknamed Viajando (“Travelling”) Henrique Cardoso? (Pinheiro Machado, 2018: 466). Why has Mexican president Andrés Manuel López Obrador (2019–) been titled “a president who does not travel”? (Carretto, 2019). The comparative literature lacks a systematic explanation of the factors behind such contrasting approaches, respectively presidentialisation and delegation, to foreign policymaking. Recently, Amorim Neto and Malamud (2019) conceptualised and measured the policymaking capacity of Latin American foreign ministries with an empirical focus on Argentina, Brazil, and Mexico in 1946–2015. Foreign ministries with high policymaking capacity are capable of insulating themselves from the policy preferences of other government agencies. In so doing, they are able to advance and implement their preferences while retaining internal and external legitimacy. Three features build up a foreign ministry’s policymaking capacity: the professionalisation of the diplomatic corps, ample institutional attributions, and extensive presidential delegation. The first is a necessary condition, while the second and third features are substitutable: a foreign ministry can have its policymaking capacity heightened either by institutional attributions or by presidential delegation. In this manuscript, we endeavour to explain the latter. The article proceeds as follows. The next section reviews both the literature on delegation to technocrats/bureaucrats and the literature on foreign policymaking to pin down the state of the art and in search of hypotheses. We posit that international, national, Amorim Neto and Malamud 125 and personal factors determine the opportunity and motives of presidents to delegate power to their foreign ministries. The third section describes our methods, indicators, and data. In the fourth section, we use Ragin’s fuzzyset qualitative comparative analysis to identify the combinations of conditions conducive to presidential delegation; in addition, we identify the conditions that lead to nondelegation. We discuss the findings and provide robustness checks in the fifth section. The sixth section concludes. State of the Art and Explanatory Factors Research on technocrats in Latin America tends to focus on contemporary examples. However, the technocratic phenomenon has roots that go back to the late nineteenth and early twentieth centuries. After a prolonged period of postcolonial frustration with the failure to accomplish rapid change, politicians in late nineteenthcentury Latin America turned to science for solutions. The technocrats of this period, of which the Mexican cientificos became the most renowned, were expected to fulfil these new hopes and aspirations (Baud, 1998: 13). This positivist mindset was expressed in the Brazilian national motto, Ordem e Progresso, which has been stamped on its flag since 1889, and in Argentina’s Paz y Administración campaign of President Julio Roca (1880–1886). Traditional Latin American populism was not dismissive of technocrats but put them firmly under the rein of the political leader. This hierarchic or complementary rather than adversarial relationship is consistent with the most recent research. Bickerton and Invernizzi Accetti (2017) assert that both populism and technocracy are similar in their “opposition to two key features of party democracy: political mediation and procedural legitimacy.” Lee and Schuler (2020: 531) further contend that autocrats “do not fear technical competence; they fear political competence.” Therefore, they “promote ministers with technical competence but punish the politically competent.” Although autocracy is not the same as populism, we consider them as functional equivalents in terms of the relationship between the political leadership and the bureaucratic/technocratic echelons of the administration. After multiple decades of bureaucratic authoritarianism (in Argentina and Brazil) or autocratic rule (in Mexico), neoliberalism established a foothold in the 1990s. Neoliberalism, with a technocratic flair not dissimilar to that of bureaucratic authoritarianism, occurred in a more heterogeneous fashion across the region, however. In Argentina and Brazil, the neoliberal turn took place under neopopulist presidents Carlos Menem and Fernando Collor. In Mexico, it evolved under traditional Institutional Revolutionary Party (PRI) rule. In the three cases, strong presidents alternated with weaker ones, and the relationship between them and their national bureaucracies displayed great variation. With regard to foreign policymaking, we want to understand the factors that led some presidents to lean on technocratic government and others to accentuate political (presidential) discretion. In other words, this is an analysis of professional diplomacy versus presidential diplomacy. Complementing recent findings on technocratic and partisan appointments in presidential cabinets (Camerlo and PérezLiñán, 2015), our Journal of Politics in Latin America 12(2)126 research goes beyond minister retention to unveil the logic behind presidential direct intervention in policymaking. Six Explanatory Factors We posit that six factors are connected to a president’s decision to delegate power to the foreign ministry. These factors are spread over three dimensions: international, national, and personal. Cason and Power have already applied this threefold categorisation to the study of Brazil’s foreign ministry, and we intend to supplement their analysis and extend it to assess our other two cases (Cason and Power, 2009). The international factor concerns the presence of systemic stability. In relatively stable times, presidents have incentives to delegate foreign policy management and focus instead on domestic issues. In contrast, during systemic crises such as the end of a world war, a global financial meltdown, or a sharp decline of a superpower, presidents may find both a motive and opportunity to intervene more directly in foreign affairs. The first domestic factor is political and concerns the presence of an elite consensus on foreign policy – or the absence or politicisation thereof. When such a consensus is present, presidents have incentives to delegate policymaking authority to professional diplomats; when the consensus breaks or never develops, more presidential foreign activism is to be expected. The rationale underlying this factor is identical to the logic of central bank autonomy: when political and economic elites converge on the goal of keeping inflation low and stable, they grant autonomy to a central bank staffed with highquality economists (Cukierman, 2008). The second domestic factor is institutional and concerns the nature of the political regime. The presence of democratic institutions generates incentives for chief executives not to delegate to the foreign ministry because democratic presidents are under pressure to consider the views of voters, interest groups, and political parties. That is to say, democracy’s accountability mechanisms compel presidents to consider preferences other than those of professional bureaucrats. In contrast, nondemocratic presidents are freer to delegate authority to the foreign ministry. The third domestic factor is ideological and concerns the programmatic orientation of the government. Rightwing parties and leaders have the motivation to rely on technocratic management, and therefore to delegate authority to professionalised foreign ministries (Kaplan, 2017; Kiely, 2017). Moreover, given the frequent elitist bias of the diplomatic career, there is an elective affinity between diplomats and rightwing politicians because the latter tend to be more closely associated with elites. Conversely, leftist presidents have the motivation to advocate policy change and frown upon such bias and, thus, to display a handson approach to professional diplomacy. The last domestic factor is bureaucratic and regards the degree of professionalisation of the foreign ministry. As mentioned above, the classical problem of public agencies becoming independent from their political masters increases when bureaucrats are professionalised. In the words of Weber (1946: 232), “The ‘political master’ finds himself in the position of the ‘dilettante’ who stands opposite the ‘expert,’ facing the trained Amorim Neto and Malamud 127 official who stands within the management of administration.” Professional bureaucrats claim a monopoly on their technical expertise, form deeply entrenched preferences over policy, establish clearcut boundaries between their bureau and the outside world, and develop an “esprit de corps” that is jealous of any hint of encroachment by outsiders. These are the drawbacks of a strong public organisation – from a chief executive’s point of view, particularly from a democratic one. While strong bureaucracies are desirable because they are an instrument of effective governance, their drawbacks generate a control problem. Conversely, when a public bureaucracy is organisationally weak, it is relatively easy for the chief executive to penetrate or marginalise it. Therefore, the presence of a professional foreign ministry motivates presidents not to delegate authority to it. Finally, the personal factor is psychological and concerns presidential personality. Broadly speaking, a weak personality can be associated with delegation, while a strong personality is connected to presidential activism in foreign policy. Psychological factors are multidimensional and very difficult to measure. Here we rely on part of Preston’s theorisation to develop one hypothesis regarding presidential personality. Preston’s study found that three individual characteristics – need for power, complexity, and prior policy experience – played a critical role in shaping presidential leadership styles in the United States (Preston, 2001). However, here we analyse only prior policy experience because it bears directly on presidents’ motivation to delegate authority to the foreign ministry, is relatively easy to observe, and therefore is relatively straightforward to operationalise on the basis of a president’s career before he or she was sworn in. Prior policy experience of leaders has a significant impact upon presidential style, the nature of advisory group interactions, and how forcefully leaders assert their own position on policy issues (Alcántara, 2012, 2017; Araya, 2015; Preston, 2001). Experience provides leaders with a sense of what actions will be effective or ineffective in specific policy situations, as well as which cues from the environment should be attended to and which are irrelevant. Leaders with a high degree of policy experience insist upon personal involvement or control over policymaking (Preston, 2001: 11–12). Therefore, a less experienced president has incentives to rely on experts for conducting foreign policy (Saunders, 2017). Conversely, a president with considerable experience in international affairs and for whom foreign policy is a salient issue has the motivation to take charge personally (Allison and Zelikow, 1999; Saunders, 2017). Cases, Methods, Measures, and Data Following Amorim Neto and Malamud (2019), our case selection is based on three facts. First, Latin America features the world’s strongest presidentialist tradition. Second, our chosen cases are the largest powers in the region, which confers greater impact to their foreign policymaking process. Third, the three cases exhibit wide variation in presidential delegation to the foreign ministry. This variation allows us to expect that our results may travel well to other presidentialist countries. As our objective is to identify the causes of a consequence, not the consequences of a cause, the appropriate methodology is the comparativehistorical method (Mahoney, Journal of Politics in Latin America 12(2)128 2010). Moreover, as our purpose is to pin down combinations of conditions conducive to presidential delegation to foreign ministries – and given the nature of our evidence, measures, and size of our sample – it is advisable to use Ragin’s fuzzyset qualitative comparative analysis (henceforth, fsQCA [Ragin, 2008]). Deploying fsQCA allows us to illuminate what Ragin calls “explicit connections” rather than the mean effects of a set of variables, as would be the case were we to conduct a multivariate regression analysis. Hocking (1999: 14) claims that “generalisations about foreign ministries, their role and significance are fraught with danger.” Precisely because we agree on this statement, a configurational analysis is a suitable instrument, as it aims at only “contingent generalisations” (George and Bennett, 2005: 31–32). Last but not least, as recently argued by Goertz (2017: 90–122), QCA is particularly useful for testing “constraint theories” such as the one developed here. Calibrating Presidential Delegation Following Amorim Neto and Malamud (2019), we consider three indicators of our dependent variable, presidential delegation to foreign ministries. The first indicator regards direct intervention in foreign affairs, which is measured through a proxy: international travel. Presidents who embark on more foreign trips, and remain longer abroad, are considered to delegate less authority on foreign policy than presidents who tend to stay home (Cason and Power, 2009).2 This delegation may be more or less implicit. Intentionally or not, diplomatically inactive presidents are more likely to empower the foreign ministry than active ones. The second indicator is the maintenance of a stable ambassadorial elite composed mostly of career diplomats. Given that the appointment of ambassadors is a presidential attribution, the decision to fill a country’s key diplomatic posts with professional diplomats is a signal of a presidential delegation that strengthens the foreign ministry policymaking capacity. The third indicator is the appointment of foreign ministers with an autonomous political basis. Amorim Neto and Malamud (2019) highlight that ordinary bills of law are a less necessary tool for foreign policy than for other policy areas. This implies that the conversion of presidents’ preferences into foreign policy will depend more on executive control over the bureaucratic apparatus than on legislative majorities.3 Presidents for whom foreign policy is a salient issue will prefer the selection of experts or bureaucratic operators acquainted with the diplomatic apparatus. The more a president wants to control foreign policy, the less he or she will be willing to delegate authority to the diplomatic corps. Hence, the chief executive will be less inclined to appoint a foreign minister with independent political support, such as, for example, a party leader. This kind of president is even likely to appoint a diplomat to head the foreign ministry. Diplomats offer him or her the double benefit of being politically weak, therefore controllable, and knowing the inner working of the bureaucracy. Amorim Neto and Malamud (2019) hypothesise that the three indicators that capture the level of delegation are substitutable, namely: (1) (low) diplomatic activism by Amorim Neto and Malamud 129 presidents, (2) the stability and the professional makeup of the ambassadorial elite, and (3) the appointment criteria of foreign ministers. For every presidential term, they allocate 0, 0.5, or 1 point for each dimension in which delegation proceeds. When a country meets a given condition in a given year, it is coded as 1; when a country meets that condition partially, it is coded as 0.5, and 0 otherwise. Since the three indicators can be cumulative, each presidential term can receive the following scores: 0, 0.5, 1, 1.5, 2, 2.5, and 3. We calibrated the scores for the purpose of conducting fsQCA. Therefore, when the score of a presidential mandate is 3 (maximum delegation), it was coded as 1; when it is 2.5, it was coded as 0.83; when it is 2, it was coded as 0.66; when it is 1.5, it was coded as 0.5; when it is 1, it was coded as 0.33; when it is 0.5, it was coded as 0.17; when it is 0, it was coded as 0. A note on Brazil’s Juscelino Kubitschek (1956–1961) is illustrative of how Amorim Neto and Malamud’s measures of delegation correspond to presidents’ behaviour in foreign policy. Kubitschek’s score on our calibration of delegation is 0.66. He travelled little and appointed foreign ministers with an autonomous power basis (both features implying delegation), but rotated his ambassadorial elite frequently and selected relatively many nonprofessional ambassadors (implying relegation). This means that he was more or less in the set of delegation. Does this summary evaluation match his behaviour in foreign policy? Kubitschek is known for personally engaging in one important foreign policy initiative, the Pan American Operation (OPA), in 1958. He did not agree with the foreign ministry on how to deal with Washington: while the president wanted a multilateral approach, the foreign ministry preferred a bilateral one. Thus, the president created an interministerial committee and set out to bring the Latin American countries together so as to negotiate US foreign aid. “In June, Foreign Minister Macedo Soares resigned, not because he opposed OPA, but because he had been excluded from its planning” (Weiss, 1993: 116). Here is a clear example of a president withdrawing delegation, although in other regards he let the foreign ministry run diplomatic affairs according to its preferences. We now turn to the calibration of the six explanatory factors or conditions. Stability of the International System Our independent international factor, systemic stability, is the only one that remains unchanged through all national cases. It is calibrated in the simplest manner, that is, as a dummy. We have identified four periods of systemic change, meaning either that the system polarity was altered or that major economic turmoil conferred extraordinary activism to lesser powers. Every one of these periods, with an assigned score of 0, includes a starting year plus the four that followed. We have defined all other years as stable and assigned them a score of 1. The first period of change is 1946–1950, when the Cold War order set in and crystallised. The second period is 1973–1977, when the first oil crisis disrupted the international economy and put an end to the thirty glorious years in the West. The third period is 1989–1993, encompassing the fall of the Berlin Wall and the dissolution of the Soviet Union. Finally, the 2001–2005 period marks the peak and Journal of Politics in Latin America 12(2)130 end of the US unipolar moment, after 9/11 and its aftermath, including the rise of China, which signalled the advent of an increasingly multipolar order. The remaining explanatory factors are located at the domestic and individual levels. Elite Consensus on Foreign Policy The category of elite consensus involves two separate questions: who is included in the elite (the actors), and what is agreed or disagreed upon (the substance). We understand the foreign policy elite as representing parties’ principal leaders, the top brass of the armed forces, and the most prominent members of the diplomatic corps, whose views are commonly intertwined with academic paradigms. As to substance, the crucial Latin American issues concern national development strategies, trade policies, and relations with the United States. Development strategies have generally alternated between a resourcebased export model and an importsubstituting industrialisation model. Accordingly, trade policies have varied between free trade and protectionism. Finally, the political dimension of foreign policies has oscillated between USfriendly and “antiAmerican” approaches. Because this compound variable is multidimensional, it is advisable to adopt a finegrained calibration rather than a dichotomy. Based on an extensive literature review and expert consultation, we operationalised it in a trichotomous manner. If there is a consensus on the political dimension and on the development/trade dimension of a president’s foreign policy, the score is 1. If there is consensus only on the political dimension or only on the development/trade dimension, the score is 0.6. This value is more appropriate than 0.5 because the latter indicates maximum indifference, which is not the case as far as consensus is concerned (Schneider and Wagemann, 2012).4 If there is no consensus on either dimension, the score is 0. Political Regime We adopt a minimalist definition of democracy as a regime in which rulers are elected through relatively free and fair (though not necessarily direct) elections, and calibrate this factor as a dichotomous variable: 1 for periods of authoritarianism or nondemocracy; 0 for democratic presidents. Our coding of Latin American political regimes coincides with that of Mainwaring and PérezLiñán (2001). Professionalisation of the Foreign Ministry If under a given president, diplomatic recruitment and promotion has not been professionalised, this president’s term receives the score of 0. Note that professionalisation began in 1963 in Argentina, 1946 in Brazil, and 1967 in Mexico (Amorim Neto and Malamud, 2019), dates to which we assign a membership score of 0.6.5 That is, for the year when professionalisation started for each country, they received a membership score greater than 0.5, as they are now more in than out of the professionalisation condition. From then on, the degree of professionalisation evolves 0.1 every five years, as successive cohorts of diplomats are recruited and promoted according to meritocratic Amorim Neto and Malamud 131 criteria. Twenty years after the professionalisation of the diplomatic service – that is, once professionalisation is mature – presidential terms are assigned a score of 1. Presidential Ideology The ideology of a president is that of their party, and party ideology is initially based on the categories and operationalisation proposed by Coppedge (1997), further revised and updated with resort to case experts as reported in Amorim Neto and Malamud (2015). As for calibration, this factor is a fuzzyset: 1 for rightist presidents; 0.75 for centrerightists; 0.5 for centrists; 0.25 centreleftists; 0 for leftists. Presidential Expertise On the basis of presidential biographies, we have first established a scale of expertise ranging from 1 to 5. Presidents who had been ministers of foreign affairs or were acknowledged specialists in international affairs were coded as 5; presidents who had been ambassadors, heads of a ministry related to foreign affairs (such as defence or foreign trade), attachés, representatives of an international organisation, or had some type of substantial international experience were coded as 4; presidents whose careers had had some kind of systematic contact with foreign affairs (e.g. a legislator who had been a member of the foreign affairs committee) were coded as 3; presidents whose careers had not provided them with any systematic contact with foreign affairs, but who had travelled abroad for political or business affairs, or had participated in international political missions were coded as 2; all others were coded as 1. Reelected presidents were automatically upgraded to the maximum score. As for calibration, we proceeded as follows: 5 was converted to 1; 4 to 0.75; 3 to 0.5; 2 to 0.25; and 1 to 0. Unlike the elite consensus variable, here it is relevant to retain the score of 0.5, as the presidents who are at the socalled crossover point can go either way in terms of delegation. The scores for systemic stability and professionalisation of the foreign ministry have been identified above, so we now discuss the four other factors for each country. Argentina Full elite consensus on foreign policy has not existed in Argentina since 1946, when Perón came to power and redrew the political system in two camps that have lasted until today. The only exception was the coalition government that, after the 2001 economic and political meltdown, led Duhalde to head a caretaker administration tantamount to a national salvation government (2002–2003). A partial consensus was reached in the 1963–1974 period, when a tacit agreement sunk in on nonalignment, import substitution, and soft contestation of existing international organisations (Russell and Tokatlian, 2006: 266). Dissent was prevalent during the rest of the time due to two stable dichotomies: one facing the proUS military administrations versus the more proautonomy civilian opposition, the other reflecting the allout adversarial relationship between Peronists and Radicals (Malamud, 2011; Saraiva and Tedesco, 2001). Journal of Politics in Latin America 12(2)138 the outcome. This prevents or allows the presence of counterfactuals in the minimisation process (Ragin, 2008: 174). Thus, we set that four conditions (elite consensus, international stability, authoritarianism, and right-wing president) contribute to the outcome when present and two when absent (diplomatic professionalization, and presidential expertise), as hypothesised. Let us thus focus on the intermediate solution. As seen in Table 2, the intermediate solution yields four substitutable sufficient conjunctions for delegation: (1) international stability and elite consensus; (2) international stability, rightwing president, and low diplomatic professionalisation; (3) international stability, rightwing president, and low presidential expertise; (4) absence of authoritarianism combined with elite consensus and rightwing president. This solution’s consistency is 0.86, with a coverage of 0.64. While this indicates a good fit in terms of consistency, in terms of coverage, that is, empirical relevance, it falls short, according to the 0.75 minimum suggested by Legewie (2013). Paths 1, 2, and 3 corroborate five of our six factors, while path 4 goes against one of our hypotheses because democracy (or absence of authoritarianism) appears as one of the factors conducive to delegation. In addition, international stability and rightwing president appear in three of the four conjunctions generated by the intermediate solution. Consensus appears in two. Low presidential expertise and low diplomatic professionalisation are present in one. Authoritarianism is not found in any of the four conjunctions. Therefore, the first five factors (international stability, elite consensus, rightwing president, low expertise, and low professionalisation) seem to be the conditions connected to presidential delegation to the foreign ministry, and in varying permutations. Authoritarianism is our weakest factor. Table 2. Results From the Necessity and Sufficiency Assessments for Delegation. Path Conjunction Relation Consistency Coverage Raw Unique Necessary condition S + R ←D0.96 0.63 – S + P ←D0.95 0.60 – ~A + R ←D0.95 0.60 – Intermediate solution 0.86 0.64 – 1S*C + 0.86 0.37 0.12 2S*R*~P + 0.91 0.21 0.03 3S*R*~E + 0.92 0.41 0.12 4~A*C*R →D0.94 0.19 0.07 Note: S = stability, A = authoritarianism, C = consensus, R = right wing, E = expertise, P = professionalisation, D = delegation. The symbol “+” denotes logical OR, and “*” denotes logical AND. “~” indicates the absence of a condition. Necessity is represented as “←” and sufficiency as “→”. Amorim Neto and Malamud 139 In order to provide empirical support for the solution, it is useful to identify the cases that hold a membership of greater than 0.5 in the conjunctions yielded in the truth table analysis. Thus, Table 3 shows the presidencies that help explain the outcome (delegation) per path of the intermediate solution, as well as the deviant cases – those that can be considered members of a path but do not present the outcome. Non-Delegation to Foreign Ministries Let us now consider the negative result, nondelegation. We test which conditions are connected to nondelegation, that is, presidential diplomacy. Appendix 3 displays the resulting truth table. Table 4 presents the results from the sufficiency and necessity analysis. Table 5 shows the distribution of cases in the paths yielded by the intermediate solution. Necessary Conditions Regarding individual causal conditions, professionalisation was deemed almost necessary to nondelegation (consistency score = .88, coverage = .47). The coverage score shows that it is just a little under the minimum threshold of 0.5. Thus, it is just about a nontrivial condition. This suggests that professionalisation is an important condition for nondelegation. A nontrivial condition suggests that the link between it and the cases is strong, and whenever the outcome is present, the necessity condition is present too. As for a trivial condition, the factor occurs so frequently in all the instances under study that its presence does not indicate much. Rightwing president or diplomatic professionalisation received the highest consistency score (0.98). Then, with a consistency score of 0.97, followed the combinations of no elite consensus or nonpresidential expertise, no elite consensus or diplomatic professionalisation, rightwing president or absence of authoritarianism, no elite consensus or rightwing president. Among these necessary conditions, the least trivial of the combinations (coverage = .48) is no elite consensus and low presidential expertise. Sufficient Conditions Diplomatic professionalisation appears in three paths. Absence of international stability and absence of elite consensus appear in two paths. Authoritarianism, absence of authoritarianism, nonrightwing president, and expertise appear only in one path each. The intermediate solution showed a lower consistency score and coverage, which also reflects on the low number of observations covered by the solution. Moreover, only two paths (1 and 3) fully match the hypotheses – that is, conditions with the opposite signs to the ones leading to delegation. These combinations are absence of international stability, no rightwing president and diplomatic professionalisation (~S*~R*P), and absence of authoritarianism, low elite consensus, and high presidential expertise and diplomatic professionalisation (~A*~C*E*P). This is precisely what is expected in fsQCA due to the assumption of asymmetric causation: the same conditions that lead to the occurrence of the outcome cannot be immediately transported to the negation set of the outcome (Ragin, 2008). Journal of Politics in Latin America 12(2)140 Table 3. Presidential Terms Explained by Each Sufficient Path for Delegation. Delegation explained by each path Deviant cases (no delegation) Path 1 (S*C) Arturo Umberto Illia (1963–1966, ARG) Roberto Levingston (1970–1971, ARG) Juan Carlos Onganía (1966–1970, ARG)* Felipe Calderón (2006–2012, MEX)* Alejandro Lanusse (1971–1973, ARG) Getúlio Vargas (1951–1954, BRA) Juscelino Kubitschek (1956–1961, BRA)* Jânio Quadros (1961, BRA)* João Goulart (1961–1964, BRA) Artur da Costa e Silva (1967–1969, BRA) Emílio Médici (1969–1974, BRA) João Figueiredo (1979–1985, BRA) José Sarney (1985–1990, BRA) Itamar Franco (1992–1995, BRA) Adolfo Ruiz Cortines (1952–1958, MEX) Adolfo López Mateos (1958–1964, MEX Gustavo Díaz Ordaz (1964–1970, MEX)* José López Portillo (1976–1982, MEX) Enrique Peña Nieto (2012–2015, MEX) Path 2 (S*R*~P) Pedro Eugenio Aramburu (1955–1958, ARG) Juan Domingo Perón II (1952–1955, ARG) Gustavo Díaz Ordaz (1964–1970, MEX)* Path 3 (S*R*~E) Juan Carlos Onganía (1966–1970, ARG)* Felipe Calderón (2006–2012, MEX)* Leopoldo Galtieri (1981–1982, ARG) Reynaldo Bignone (1982–1983, ARG) João Café Filho (1954–1955, BRA) Juscelino Kubitschek (1956–1961, BRA)* Jânio Quadros (1961, BRA)* Path 4 (~A*C*R) Eurico Dutra (1946–1951, BRA) Vicente Fox (2000–2006, MEX) Juscelino Kubitschek (1956–1961, BRA)* Felipe Calderón (2006–2012, MEX)* Jânio Quadros (1961, BRA)* Note: Period and country of each presidential term are detailed inside the brackets. “*” represents cases that can be explained by more than one path. Countries are abbreviated as follows: BRA = Brazil; ARG = Argentina; MEX = Mexico. Cases are deemed explained by a path, whenever it holds a membership greater than 0.5 in that configuration. Amorim Neto and Malamud 141 Table 4. Results from the Necessity and Sufficiency Assessments for NonDelegation. Path Conjunction Relation Consistency Coverage Raw Unique Necessary condition P←~D 0.88 0.47 – R + P ←~D 0.98 0.44 – ~C +~ E →~D 0.97 0.48 – Intermediate solution 0.70 0.49 – 1~S*~R*P + 0.79 0.17 0.11 2~S*A*~C*P + 0.74 0.06 0.05 3~A*~C*E*P + 0.68 0.32 0.27 Note: S = stability, A = authoritarianism, C = consensus, R = right wing, E = expertise, P = professionalisation, D = delegation. The symbol “+” denotes logical OR, and “*” denotes logical AND. “~” indicates the absence of a condition. Necessity is represented as “←” and sufficiency as “→”. Table 5. Terms Explained by Each Sufficient Path for NonDelegation. Nondelegation cases explained by each sufficient path Deviant cases (delegation) Path 1 (~S*~R*P) Luiz Inácio Lula da Silva I (2003–2006, BRA) Raúl Alfonsín (1983–1989, ARG) Luis Echeverría (1970–1976, MEX) Path 2 (~S*A*~C*P) Jorge Videla (1976–1981, ARG) Carlos Salinas de Gortari (1988–1994, MEX) Path 3 (~A*C*~E*P) Carlos Menem II (1995–1999, ARG) Cristina Fernández de Kirchner II (2011–2015, ARG) Fernando Henrique Cardoso I (1995–1998, BRA) Fernando Henrique Cardoso II (1999–2002, BRA) Luiz Inácio Lula da Silva II (2007–2010, BRA) Dilma Rousseff II (2015–2015, BRA) Ernesto Zedillo (1994–2000, MEX) Note: Period and country of each presidential term are detailed inside the brackets. Countries are abbreviated as follows: BRA = Brazil; ARG = Argentina; MEX = Mexico. Cases are deemed explained by a path, whenever it holds a membership greater than 0.5 in that configuration. Journal of Politics in Latin America 12(2)142 Discussion and Findings As expected in configurational analysis, we found equifinality: many roads led to presidential delegation to foreign ministries in Argentina, Brazil, and Mexico in 1946–2015. Various configurations of international stability, low diplomatic professionalisation, elite consensus, rightwing presidents, and low presidential expertise on foreign affairs were found sufficient for delegation. These five factors are INUS conditions, that is, an insufficient but necessary part of a condition that is itself unnecessary but sufficient for the result (Mahoney et al., 2009). Specifically, there are four alternative paths to delegation: (1) international stability and elite consensus; (2) international stability, rightwing president, and low diplomatic professionalisation; (3) international stability, rightwing president, and low presidential expertise; and (4) absence of authoritarianism combined with elite consensus and rightwing president. Surprisingly, only one of these combinations contradicted the hypothesised conditions: absence of authoritarianism (instead of presence) combined with elite consensus and rightwing incumbent. These results suggest that contextual conditions matter, as is the case of rightwing democratic presidents when there is elite consensus. These were the cases of Brazil’s Dutra (1946–1951), Kubitschek (1956–1961), and Quadros (1961). As for professionalisation, it was deemed almost necessary to nondelegation by itself. This is a most relevant finding, one that corroborates the Weberian hypothesis that it is precisely the strongest bureaucracies that create control problems for their political masters, paving the way for the latter to tame the foreign ministry.11 With regard to nondelegation, the six hypothesised conditions were present in the solution terms as INUS conditions: diplomatic professionalisation, absence of elite consensus, absence of international stability, presidential expertise, absence of authoritarianism, and absence of rightwing president. Only one path showed a factor with the opposite sign as hypothesised for delegation: absence of authoritarianism in path 4. However, it makes intuitive sense that democratic regimes might as well lead to delegation whenever accompanied by facilitating conditions (elite consensus and rightwing presidents). Findings Our findings on presidential expertise, absence of elite consensus, and absence of international stability as conditions of nondelegation corroborate to a large extent the work of Cason and Power (2009) on the causes of the “rollback” of Brazil’s foreign ministry in the 1990s and 2000s. As mentioned, the authors highlighted precisely personal, domestic, and international factors underlying the gradual erosion of this once powerful bureaucracy. How do our findings converse with some key works on presidential appointment strategies and patterns of bureaucratic recruitment in Latin America? First, international stability proved to be an important and frequent condition of delegation in both sufficient and necessary solutions (when combined with other conditions). This dovetails with Geddes’s findings that Latin American presidential Amorim Neto and Malamud 143 appointments are conditional on chief executives’ political survival strategies (Geddes, 1994). To put it in other words, when Latin American presidents face political crises, they have incentives to politicise their bureaucratic appointments (Merke et al., 2020). Symmetrically, in periods of international stability, it becomes less costly for presidents to delegate to diplomats than in periods of international upheaval. Our findings on the of role international stability in presidential delegation to foreign ministries also converge with those of Kaplan (2017). He shows that incumbents’ delegation preferences over economic agencies are conditional on external shocks (economic crises). Likewise, as argued by Dargent (2014), competition for the policymaking arena increases with international turmoil, rendering delegation to social and economic experts less likely due to what he calls an “unbalanced constellation of stakeholders.” We have shown that there is a similar dynamic at work with regard to diplomats. Presidents’ low expertise on foreign affairs also converges with Dargent’s (2014) findings on the asymmetry of information between experts and politicians. Such asymmetry strengthens the propensity of politicians to delegate policymaking authority to technocrats. Dargent (2014: 42) also shows that elite consensus by itself does not lead to the autonomy of the technocracy, but can help to “enhance technical autonomy and raise the likelihood of technical continuity.” Likewise, our fsQCA shows that elite consensus is found to be a sufficient element for delegation to experts only when accompanied by other factors such as international stability and rightwing presidents. Finally, rightwing presidents, when combined with other conditions, were found to be a sufficient solution and a necessary factor for delegation to the foreign ministry. In other words, the link between the condition and the outcome is so strong that a rightist president not only facilitates delegation (sufficient), but is needed in order to have delegation (necessity). This finding converges with those of Schneider (1998) and Ai Camp (1998) on the association between the rise of technocracy and neoliberalism in Latin America. Our findings on presidential ideology also converge with those of Kaplan (2017) on Latin American economic agencies: the profile of those appointed to head the latter differs according to government ideology. Our findings also add a twist to Merke et al.’s (2020) claim that presidential ideology is the single most relevant factor behind foreign policy change in Latin America: apparently, this is more so when presidents are leftists, as they tend to lean less than rightwing leaders on status quobiased bureaucracies. Robustness Checks Møller and Skaaning (2019) recommend that, similarly to other methodologies, robustness checks should be conducted in QCA studies to further prove the results. First, Marx and Dusa (2010) advocate that for a desirable ratio of conditions to cases, for large sample sizes, a 0.2 is the maximum recommended. This study’s conditions to cases ratio is 0.11 (6/54), thus passing the test. Journal of Politics in Latin America 12(2)144 Second, Braumoeller (2017) and Veri (2017) argue that QCA results might be contaminated by ambiguous cases. Thus, we checked whether solutions were being explained by deviant cases. As seen in Tables 3 and 5, this does not seem to be the case. Third, Grauvogel and von Soest (2014) argue that one can compare the different solutions to distinguish core conditions. As seen in Table 2 and in tests not reported here, the resulting paths to delegation share commonalities like stability, which is also present in most of paths of the parsimonious and complex solution. Also, the parsimonious solution shares the same number of solutions and two sets of equal paths (international stability and elite consensus; international stability, rightwing president, and low presidential expertise). The other paths differ only due to the presence of the rightwing INUS condition in path 2 and path 4 in the intermediate solution. Similar to Mello (2014) and Grauvogel and Von Soest (2014), another possible robustness check is to make changes in the sample and see how they affect the results. This was done by rerunning the sufficient and necessary tests for delegation without the Mexican presidents. Mexico was removed because, unlike Argentina and Brazil, all its presidents concluded their constitutional terms, it did not have a military regime, it did not go through the pink wave in the twentyfirst century, and it is located in North America. Paths 2, 3, and 4 are the same in both tests, as seen in Table 2 and in tests not reported here. This is important especially because paths 2 and 3 are in line with the expected hypotheses. Path 1 differs from the one in Table 2 only due to an extra condition (no authoritarianism) and, apart from that, the new sample yielded a fifth path (S*A*R). Necessary conditions performed similarly in the new sample. The yielded consistency scores were also higher than 0.9, indicating that an extra country (Mexico) does not significantly affect the results. As a final robustness check, we used an alternative measure of delegation based on the proposition that the more a president changes the foreign minister, the less authority he or she delegates to the foreign ministry. We estimated the maximum time – as a proportion of the president’s term – that the longest lasting foreign minister held the office during each presidency. This is a simpler measure than the one developed by Amorim Neto and Malamud (2019). However, it captures a dimension – foreign minister turnover – that was absent in their indicators. The sufficiency analysis was conducted in the same manner (i.e. the minimum consistency score was set at 0.9, and the same factors were used) as the initial one. As a result, seven causal conjunctions were generated, two of which were similar to the original analysis: (1) the combination of stability and consensus, and (2) consensus and rightwing. In the latter conjunction, the nonauthoritarian INUS factor is not present as in the original analysis. It is noteworthy that stability, consensus, and rightwing also appear with the same expected (positive) sign as in the original analysis, and thus remain INUS conditions. However, the new findings cast doubts on the role of authoritarianism and expertise, as these factors appear with both signs in the alternative analysis. Also, professionalism now appears with a positive sign (instead of a negative) as an INUS condition. Only three out of seven items now fully corroborate our hypotheses (instead of three out of four). As for the necessity tests, all INUS conditions were confirmed. Our new delegation measure corroborates the original finding that Amorim Neto and Malamud 145 professionalisation is a necessary factor for nondelegation, even if at a lower threshold (consistency = .76 and coverage = .20). Hence, considering that fsQCA models are very sensitive to changes in measures, and that a large part of our results still holds when using a completely different proxy for delegation, our original findings remain relevant. Conclusion This article offers the first systematic assessment of factors connected to presidential delegation to foreign ministries, a phenomenon that has been described by country experts but never explained in depth or tested in a comparative fashion. We contend that there is no reason to exclude foreign ministries from the comparative literature on presidential appointment strategies and patterns of bureaucratic recruitment in Latin America. The study of diplomats essentially reinforces this literature’s findings on economic and other ministries (Camerlo and MartínezGallardo, 2017; Camerlo and PérezLiñán, 2015; Dargent, 2014; Kaplan, 2017), and therefore provides support to the claim that foreign policy should be studied as a dimension of public policy – and not as a special domain where crossparty consensus prevails, as used to be taught in the most prestigious of Latin American diplomatic academies. Overall, our main proposition – that presidential delegation to foreign ministries depends on motivation and opportunity – seems to hold water for Latin America’s three largest powers, particularly in the four decades following the end of World War II. While systemic opportunities (international stability) were similar for Argentina, Brazil, and Mexico, motives to delegate – including elite consensus on foreign policy, diplomatic professionalisation, presidential ideology, and presidential expertise on foreign affairs – varied crossnationally and over time. As for the negative outcome (i.e. nondelegation), professionalisation was deemed almost necessary to nondelegation by itself. The combinations of rightwing president or diplomatic professionalisation and the absence of elite consensus or low presidential expertise came out as weak necessary conditions. Furthermore, the absence of international stability, norightwing presidents, the absence of elite consensus, and presidential expertise and diplomatic professionalisation were found to be INUS conditions of nondelegation. In other words, while there is more than one path to delegation, the aforementioned conditions seem to foster presidential delegation to foreign ministries whenever present. Since they are INUS conditions, they are not enough by themselves for presidents to delegate to the foreign ministries, but they become so whenever combined with other factors. The relevance of these conditions was highly intuitive in light of the literature on presidential appointments and patterns of bureaucratic recruitment in the region (Amorim Neto and Malamud, 2019; Dalbosco, 2014; Faria and Casarões, 2013; Figueroa and Schiavon, 2014), but the impact of different configurations could not be told beforehand. We found some apparently obvious combinations, but also several unexpected ones, that foster or hinder delegation. The analysis identified four causal paths conducive to delegation, and three conducive to nondelegation. Together, they account for thirtyeight cases (combining the sufficient analysis for delegation and nondelegation) out of our total sample of fiftyfour, Journal of Politics in Latin America 12(2)146 as illustrated by Table 3 and Table 5. In each set, one path stood out. With regard to delegation, path 1 (international stability and elite consensus) explains nineteen cases out of twentysix, most of them clustered in the 1950–1990 period. We call it “the golden age of professional diplomacy.” With regard to nondelegation, path 3 (nonauthoritarianism and nonconsensus, plus expertise and mature professionalisation) explains seven cases out of twelve, all of them clustered after 1994. We call it “the rise of presidential diplomacy.” Both groups include presidents of our three national cases. This means that the transition from professional diplomacy to presidential diplomacy depends on context rather than country. One could say that, with regard to foreign policymaking in the largest Latin American powers, history trumps geography. Acknowledgements The authors thank the superb research assistance provided by Pedro Accorsi, Camila Farias, and Sebastián Ronderos. We are grateful to Igor Acacio, Ignacio Arana Araya, Mariana Borges da Silva, Marcelo Camerlo, Ana Covarrubias, Júlio César Cossio Rodriguez, Feliciano de Sá Guimarães, Gregory Michener, Guadalupe González, Antonio OrtizMena, Luis Schenoni, Jorge Schiavon, and Consuelo Thiers for data, comments, and wisdom. The authors also thank two anonymous reviewers for their very useful critiques and suggestions. Amorim Neto acknowledges the support of Brazil’s National Research Council (CNPq). Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/ or publication of this article: Malamud acknowledges the support of Portugal’s Fundação para a Ciência e a Tecnologia (Project UID/SOC/50013/2013). Notes 1. Waltz (1967) argues that, in the US presidential system, key aspects of foreign policy need congressional consent, thus rendering policy-making in this area more collective than in parliamentary UK, where the prime minister enjoys wide latitude. This collective aspect of presidentalism is less pronounced in Latin America, where executive pre-eminence in foreign policy-making remains an outstanding feature. 2. According to Tenpas et al. (2012), incumbent US presidents at the beginning of the postWorld War II period travelled less than their post-Cold War counterparts. The same trend is found among Argentine, Brazilian, and Mexican chief executives. However, as shown by Amorim Neto and Malamud (2019), some Cold War Latin American presidents travelled more frequently than their post-Cold War counterparts. Intentionally or not, and regardless of transportation technology, presidents who travel less end up delegating more authority to their diplomats than those who travel more. Amorim Neto and Malamud 147 3. Note that the appointment of ambassadors needs senatorial consent in Argentina, Brazil, and Mexico. In Mexico, presidential international travel also requires such consent. Foreign aid budgetary allocations, the deployment of troops, and declarations of war depend on congressional approval in the three countries too. These constitutional constraints show that presidents cannot conduct foreign policy regardless of legislatures. However, the non-negligible role of the legislative branch in foreign policy does not invalidate our contention that the conversion of presidents’ preferences into foreign policy depends more on executive control over the bureaucratic apparatus than on legislative majorities. 4. When a case holds a membership of 0.5 in one or more conditions, it cannot be assigned to any of the truth table rows due to a mathematical property of fsQCA. Schneider and Wagemann (2012) recommend avoiding the 0.5 mark whenever possible. 5. See note in Table 1 for the exception in the Mexican case. 6. Our main source is Biografía de Presidentes de Argentina, available at https://www. buscabiografias. com/ biografia/ verDetalle/ 10559/ Presidentes% 20de% 20Argentina 7. On Brazil’s foreign policy and the degree of consensus underlying it, see also Bandeira (2005); Barreto, 2001, 2006); Brigagão and Rodrigues (2006); Cervo (2002); Gonçalves (2011); Lima (2000), Pereira (2003, 2005); Prado and Miyamoto (2010), Saraiva (2011), Sennes (2003); Souto (2003); Vizentini (1998, 2004). 8. An excellent online source for the biography of Brazilian presidents is Getulio Vargas Foundation’s (FGV) Dicionário Histórico-Biográfico Brasileiro (DHBB), available at http:// www. fgv. br/ cpdoc/ acervo/ arquivo 9. On Mexico’s foreign policy and the degree of consensus underlying it, see Bernal-Meza (2007); Domínguez and Castro (2009); González and Pellicer (2011); Grabendorff (1978), Herrera and Santa Cruz (2011); Ojeda (1984); Raat and Brescia (2010); Velázquez (2010). 10. Mexican presidents’ biographical data were culled from Ai Camp (2011b). 11. This finding dovetails with those of Drezner (2000). He argues that the more insulated the US foreign policy bureaucracy, the higher its probability of survival but the lesser its ability to influence the broad contours of policy. Conversely, a more fluid bureaucracy may have greater influence. Thus, on the basis of Drezner, as the foreign ministries of Argentina, Brazil, and Mexico became more professionalised, they became more autonomous and, concomitantly, less able to affect foreign policy. This, in turn, facilitated non-delegation. ORCID IDs Octavio Amorim Neto https:// orcid. org/ 0000000238913361 Andrés Malamud https:// orcid. org/ 0000000188381650 References Ai Camp R (1998) Technocracy a la mexicana, antecedent to democracy. In: Centeno MA and Silva P (eds) The Politics of Expertise in Latin America. Basingstoke: Palgrave Macmillan, pp. 196–213. Ai Camp R (2011a) Mexico: What Everyone Needs to Know. Oxford: Oxford University Press. Ai Camp R (2011b) Mexican Political Biographies, 1939-2009, 4th edn. Austin, TX: University of Texas Press. Alcántara M (2012) El Oficio del Político. Madrid: Tecnos. Journal of Politics in Latin America 12(2)154 Appendix 3 Table 3A. Truth Table Analysis of NonDelegation With Consistency CutOff Set as 0.75 and Frequency Set at 1. S A C R E P Freq. ~D Raw consist. 0 1 1 0 0 1 1 1 1.00 0 1 0 1 1 1 1 1 0.89 0 1 0 1 0 1 1 1 0.88 1 0 0 1 1 1 1 1 0.77 1 0 0 0 1 1 3 1 0.76 0 0 0 0 0 1 2 1 0.76 1 1 0 1 1 0 1 0 0.71 1 1 0 1 1 1 2 0 0.70 0 0 0 1 0 1 3 0 0.69 1 1 1 1 1 1 4 0 0.68 1 1 1 1 0 1 1 0 0.65 0 0 0 1 1 0 1 0 0.64 0 0 1 1 0 1 1 0 0.62 1 0 0 0 0 1 2 0 0.61 0 1 1 1 1 1 1 0 0.59 1 1 0 1 0 1 2 0 0.54 1 0 0 1 0 1 1 0 0.51 0 0 1 1 1 1 1 0 0.49 1 0 1 1 0 1 3 0 0.45 1 0 1 0 0 1 1 0 0.42 1 1 1 0 1 0 1 0 0.38 1 0 1 0 1 1 1 0 0.36 0 1 1 1 0 0 1 0 0.25 1 0 0 1 1 0 1 0 0.23 Note: S = stability, A = authoritarianism, C = consensus, R = right wing, E = expertise and P = professionalisation, ~D = nondelegation. Nondelegation refers to whether the cases of the combination or row are above the consistency cutoff, that is, they are set as a subset of the outcome (No = 0; Yes = 1).