[1]
EXECUTIVE PRIVILEGE
?
Leadership Turnover and Foreign Policy Change in
Democratic Systems.
Pascal Snijders (s1697404)
Bachelor’s Project Foreign Policy in the Trump Era (7562) Supervisor: Dr. M.R. Di Giuseppe
Leiden University, Faculty of Social Sciences Word count excluding bibliography: 6871
[2]
rump, the Insurgent, Breaks With 70 Years of American foreign policy – reads the headline of an article in the New York Times written by Mark Landler on the 28th of December 2017. The article is basically a list of all the decisions made by Trump that, according to the author, break with US foreign policy tradition. The criticism of the new NATO headquarters, the complaints about the unfair financial burden on the US’s shoulders, the recognition of Jerusalem as the capital of Israel, not to mention the various falling outs over trade Trump has had with China, the EU, and neighbouring countries; all this begs the question whether one leader, especially one so polarising, can actually fundamentally change a country’s foreign policy, or are they limited in their ability to do so? The news that the House of Representatives voted on January 22nd to refuse President Trump any federal money to pay the costs of leaving NATO (Editorial board, 2019, Jan. 26), might show that Congress is fighting back against the excesses of President Trump’s rhetoric. Is this a one time, and perhaps only partisan action or is it indicative of larger institutional constraints that are ever present and limiting the President’s manoeuvrability? If these restraints exist, do they exist in other forms of democracy? To see what impact leadership has on foreign policy within and between democracies, I shall research whether a difference between foreign policy preferences of democratic presidential countries and democratic parliamentary countries arises due to leadership turnover.
To lay the foundation for this research the literature shall first be discussed in three segments. First, the role of leaders in International Relations (IR) and foreign policy. Second, the literature on democracy and foreign policy, and corresponding institutional expectations. Third, I shall discuss the relationship between foreign policy and voting in the United Nations General Assembly, and why looking at changes in UNGA voting patterns might be used to indicate domestic constraints. Then, I shall present and discuss my theory and hypothesis. Afterwards, my research design shall be outlined, so I will provide a proper explanation for my decisions. Finally, the empiric results shall be presented and discussed.
IR, Leadership, and Foreign Policy
The debate about the pre-eminence of leaders, or their influence on policy, is a long and ongoing one. In the nineteenth century, the historian Thomas Carlyle described history as being “but a biography of great men,” articulating a belief held by many at the time. Great men, like Lincoln or Otto von Bismarck, might have disagreed with Carlyle, as both placed great importance on the events, which they confess controlled them, rather than them controlling events (Horowitz, 2018). With the onset of the Cold War, IR scholars would mostly side with these Great men
[3]
and argue that leaders are not critical. Robert Jervis (2013) best articulates the critiques: First Jervis argues that all people who might come to power at any given time will roughly share the same values and beliefs, since those that do not rarely rise to the top of the elite, secondly, leaders are socialised while in office, and thirdly, leaders are restrained both domestically and internationally. Leaders have also often been relegated to the realm of ‘errors’, as they were deemed too idiosyncratic for any study (Rosenau, 1966 as cited in Horwotiz, 2018), and even if they did matter, their actions were deemed too hard to generalise (Byman & Pollack, 2001 as cited in Horowitz, 2018).
Additionally, scholars have for a long time argued that foreign policy differed from domestic policy, as foreign policy would be driven by higher international imperatives, such as ‘National Interest’ while domestic policy is the result of group competition over policy, something a leader would have little interest over (Mattes et al., 2015). This outlook has however lost ground, as scholars have put foreign policy firmly within the realm of policy competition. Foreign policy can be explained, not by leaders, but by influential economic actors and their interests vying for a say in foreign policy (Snyder, 1991; Narizny, 2003, 2007). Solingen (2007) adds that the make-up of ruling coalitions defines foreign policy, and even the use of nuclear weapons. Do leaders define the ruling coalitions, or do the ruling coalitions define leaders? It would seem that leaders aren’t governing, but are governed.
Contrary, there are those scholars who claim that leaders do in fact matter as a leader’s level of experience plays an important role in decision-making (Saunders, 2017). Saunders continues by claiming that a seasoned team of government veterans, or advisors, cannot compensate for the inexperience of a leader, which may exacerbate pre-existing biases. Moreover, leaders should be considered a decision-making unit that changes over time and according to Hermann et al. (2001), the personality of the leader in question matters, especially concerning how sensitive a leader is to political constraints.
Furthermore, Horowitz (2018) posits that while the institutional approach offers much value, idiosyncratic leaders should not be ignored, since leaders are the ones that make the final call in crises, but also because leaders can choose to focus, domestically and internationally, on certain policy priorities, and where to spend political capital. By doing so, leaders exert influence on policy within the allotted manoeuvrable room (Horowitz, 2018). Besides this, Horowitz (2018) also argues that leaders influence public perception (domestically and internationally) of the leader’s country, which is an important factor in international relations and foreign policy. Is a country deemed as trustworthy, and does a change in perception follow leadership turnover?
[4]
Building on this notion that leaders should be considered a decision-making unit, Wolford posits that leaders should not only be considered a unit of analysis, but the fundamental unit of analysis, as leader-level variables account for a far larger variance in empirical models than other level variables (Wolford, 2007; Wu & Wolford, 2018). Wu and Wolford (2018) also posit that leaders have incentives to build reputations when they are fresh, they will be tested domestically and internationally, so periods following leadership turnover often depict higher probability of escalation of disputes, and drastic changes in policy.
Overall, it is clear to say that the debate has been ongoing, and may remain ongoing, or perhaps someday the pendulum will no longer swing to the other side, but remain firmly in the middle. Perhaps this synthesis is forming now, the latest phase of the debate is one where the leader is considered as an important actor and decision making unit that can shape reputation, make final calls, and exert his influence in a variety of ways, but this outlook does not neglect the presence of institutional restraints that may limit leaders and their excesses. This is a view that is shared by Mattes, Leeds and Carrol (2015), who built their research on the idea that despite constraints of the international (and domestic) system, leaders retain meaningful choices in foreign policy to pursue. It follows that the make-up of the constraining factors defines how many “meaningful choices” a leader has, so not every system is the same. Leaders might have different levels of manoeuvrability in non-democracies when compared to democracies, or even between democracies. How do democracies stack up in relation to foreign policy?
Democracy, Institutions, and Foreign Policy
Scholarly debate about democracies and foreign policy goes back to at least Alexis de Tocqueville, who at the time argued that America’s democratic system would cripple the successfulness of its foreign policy. Frequent leadership turnover and the need to appeal to public opinion makes democracies unable to remain consistent in foreign policy positions, nor adhere to long-term commitments. Gartzke and Gleditsch (2004) argue that the frequent turnover in “winning coalitions” is what makes democracies unreliable allies. Contrary to this view are those who argue that democratic institutions moderate the volatility of leadership turnover, and that democracies have in fact highly stable foreign policy commitments (Leeds, et al. 2009; Mattes, 2012, Mattes, et al. 2015).
Debs & Goemans (2010) posit that democratic systems carry certain policy expectations, moderating the policy effects of leadership turnover. Likewise Christensen and Redd (2004) argue that the extensive bureaucracy lead to policy stability despite leadership turnovers. There thus seem to be systemic moderations on the influence of idiosyncratic leaders in democracies.
[5]
Tsebelis (2000), and later on Opperman and Brummer (2017), argue that institutional constrictions, domestic and international, arise due to the presence of veto players, and the constellation of said veto players, their numbers and nature, affects policy stability, and leadership manoeuvrability. However, even within such an institutionalist approach, one needs to take into account the characteristics of veto players, and of leaders, as the individual characteristics of leaders like political beliefs and leadership traits render them less or more prone to block policy change, or to initiate it (Opperman & Brummer, 2017). Given the difference in possible veto player make-up, especially across political systems, it would thus follow that different political systems offer greater or lesser freedom for its respective leaders regarding policy. Would it the not be wrong to treat the category of ‘democracy’ as a homogeneous set, as the institutional make-up of democracies clearly differs?
De Tocqueville’s critique focusses on the young United States, since then multiple democracies, besides the U.S. presidential system, have arisen. Just as de Tocqueville compared democracy to the plethora of non-democracies present at the time, modern-day scholars have only researched the difference between democracies and non-democracies regarding their institutional make-up and (foreign) policy. Mattes, Leeds, and Carroll (2015) have shown that democracies have more consistent foreign policy preferences when compared to non-democracies. However, just as democratic systems differ from non-democratic systems in terms of institutional make-up, so do democracies differ from democracies. Mattes, Leeds, and Carroll (2015) support this claim and have stated that the next step in the research of foreign policy should be to differentiate within the subsets of democracies and non-democracies. This differentiation is however not yet resembled in the literature, as comparisons placing them in a larger discussion is lacking, and only in-depth case studies can be ascertained. Thanks to previous research we now know that democracies overall have a tendency to be more consistent, however, are all democracies as consistent, or might there be a difference depending on the make-up of said country? Therefore, I find it imperative to dive into the unknown, and empirically test our theoretical understanding of democratic institutions and their impact on foreign policy. I shall differentiate between the two largest variations of democratic systems within the democratic subset, presidential and parliamentary democracies.
Foreign Policy Constraints and UNGA Voting
The goal of this research is to measure any possible change in foreign policy of democratic nations. To do this, a viable way to measure changes in foreign policy orientations of countries is required. An arena in which such changes might be witnessed is the voting in the United
[6]
Nations General Assembly. It is the only arena wherein all nations, no matter their size, are represented and have the opportunity to voice their views. While larger, and wealthier states have various way through which to make their policy positions known, the tools to do so for smaller and less developed states are limited, so UNGA meetings offer a great opportunity to communicate their foreign policy positions on an equal footing, given the universal membership of the UNGA (Mattes et al., 2015). Therefore UNGA voting data has become the standard for measuring state preferences, as they are clearly comparable and observable (Bailey et al., 2017; Mattes et al., 2015). Mattes, Leeds and Carroll (2015) argue that UNGA voting shows patterns, providing an indirect reflection of a state’s foreign policy positions, naming UNGA voting records as a latent indicator of foreign policy orientations and international alignments. After studying how domestic institutional changes affected states’ alignment with the United States, Voeten concluded that: “the current dominant focus in the IR literature on the effects of domestic institutions is too restricted. The interaction between institutional arrangements and societal preferences produces political outcomes” (as cited in Mattes et al., 2015). Potrafke (2009), and Dreher and Jensen (2013) demonstrate that domestic ideological preferences matter as alignment to the United States was influenced by ruling party ideology of OECD countries (Potrafke, 2009), and when leaders’ ideologies are similar to that of the US President, countries are more likely to vote in line with the US (Dreher & Jensen, 2013). While these studies do indicate the presence of ideological influence of leaders, they do not consider the presence of domestic institutional constraints, which is what I am interested in. Mattes, Leeds and Carrol (2015) share this critique and have partly filled this gap by using UNGA voting data to show that democratic institutions put constraints on leaders, as countries with democratic institutions show less year-to-year change in UNGO voting patterns, as opposed to non-democracies. My study sets out to add to this, by going even deeper and consider the different democratic institutions.
Theory and Hypotheses
When researching the foreign policy preferences, and leadership turnover in democratic systems, I find it inadequate to simply treat the democratic subset as homogeneous. As discussed in the literature review, the make-up of institutions, shapes the playing field for actors in the political system. Not one system is thus alike to the other, even when considered democratic. Given the diversity within the democratic grouping, the varying institutional constellations and their, possibly, varying degrees of moderating capabilities, I shall focus on the distinction between democracies, or more accurately, between presidential and
[7]
parliamentary systems, as a more in-depth comparative analysis is required in order to shed a light on the moderating effects of the different democratic institutions.
Parliamentary systems are characterised by the dependence of the executive on the legislature, as government derives its mandate from parliament, and in order to not risk being ousted by parliament, the government must maintain majority confidence (Tsebelis, 2000). Tsebelis (2000) also argues that the legislature and executive effectively represent the same constituency due to the endogenous nature of government formation. Additionally, Palmer, London and Regan (2004) posit that governments leaning on different parliamentary majorities (left-leaning, or right-leaning) behave differently in the foreign domain, indicating different foreign policy priorities when majority support in parliament shifts, as might happen after elections. Change in policy preferences, either set out by the executive or by the legislature, thus only occurs when the parliamentary majorities drastically change. However, the frequency of this happening in parliamentary systems is not that high as junior partners might change, but senior partners usually remain within coalitions, whose preferences usually prevail (Kaarbo, 1996). The coalition system of parliamentary democracies, makes leadership turnover less frequent, and less impactful.
Additionally, literature shows that legislatures in parliamentary democracies are also no longer content with letting government handle foreign policy on its own, and are increasing permanent control over foreign policy preferences, and policy formulation through the use of standards and committees (Nolan, 1985; Capling & Nossal, 2003; Strong, 2015; Ihalainen & Matikainen, 2016).
Contrary, presidential systems are characterised by an executive that operates independently from the legislature, as it derives its mandate through direct elections, and not from the legislature. De jure, foreign policy prerogative often lies with the legislature (Trimble, 1989; Tsebelis, 2000; Horowitz, 2018), but de facto the executive has the upper hand given the unwillingness, and often inability, of the legislature to make use of its own legal prerogatives (Trimble, 1989). The executive also has the unique control over foreign policy bureaucracies (Trimble, 1989), and inhibits the communicative and articulation functions, all of which are key for setting and executing policy (Peake, 2001). The communicative and articulation functions mentioned by Peake (2001) are also essential for building reputation and perception, as they enable the executive to position and assert themselves abroad, making it vital for foreign policy (Horowitz, 2018; Wu & Wolford, 2018). Politics and diplomacy are, after all, the art of perception. The literature points to a consensus that the executive, especially the chief executive
[8]
in the form of the president, is regarded as the dominant force in foreign policy (Trimble, 1989; Tsebelis, 2000; Peake, 2001, Horowitz, 2018).
Additionally, a change in leadership in the U.S. is often accompanied by a change in ruling party, which substitutes a change in domestic source of leadership support (Mattes, Leeds & Matsumura, 2016), and since a new president represents a different constituency and is elected through a winner-takes-all style election (Tsebelis, 2000), no coalitions are needed, unlike a parliamentary system which makes leadership turnover more frequent, and change (both in leadership and in policy preferences) more pronounced, compared to parliamentary systems. In short, presidential systems are more idiosyncratic than parliamentary systems. A caveat in the literature that needs to be addressed is that the literature on presidentialism and foreign policy focuses exclusively on the United States of America. Any assumptions made by me on the basis of the literature may thus work for the U.S.A., but might not apply in general to all democratic presidential systems. For the purpose of this research, I shall generalise the expectations created by the literature on the U.S. presidential system to all democratic countries labelled a presidential, since they tend to follow the same institutional framework, and electoral systems on which my assumptions are founded.
All in all, given the institutional frameworks of the two political systems, I theorise that presidential systems are highly idiosyncratic and favour the executive, while parliamentary systems are more coalition centred and favour the legislature. Thus, I assume that presidential systems are more susceptible to foreign policy inconsistency than parliamentary systems, and that leadership change in presidential systems leads to greater change in foreign policy preference than in parliamentary systems. This has led me to formulate the following two hypotheses:
Hypothesis 1: Presidential Democracies show more inconsistency in UNGA voting patterns than Parliamentary Democracies.
Hypothesis 2: The impact of changes in domestic sources of leader support on changes in UNGA voting patterns is stronger in Presidential Democracies than in Parliamentary Democracies.
Research Design Dependent Variable
To test my hypotheses, I shall use the already pre-existing yearly voting records in the UNGA from 1975 to 2008. As discussed, UNGA voting data is the standard when measuring foreign
[9]
policy preferences, and any change that might happen. However, for the purpose of this research I am only interested in shifts in policy positions of countries in relation to themselves, and not their relationship to each other states or to great powers. Besides I need to be able to establish continuity across periods in UNGA voting. Bailey, Strezhnev and Voeten (2017) developed the ideal point model, a measure that is suitable for my goals. Bailey et al. (2017) posit several advantages of this model: it allows for more valid comparisons, as changes in voting agendas are no longer falsely attributed as changes in preferences. Additionally, it allows to identify more meaningful shifts in foreign policy orientations, and it allows for more analyses than only dyadic analyses (Bailey et al., 2017). The ‘IdealPoints model’ developed by Bailey et al. (2017) shall serve as the basis for my compiled dataset. Using the ideal point data, I shall generate a continuous measure of yearly changes in foreign policy positions, using the absolute value of the difference between the nation’s ideal points in 1 year relative to the previous year.
Independent Variables
To assess the effect of leadership turnover I need to define what a leadership turnover is and devise a way to measure this. A leadership turnover shall be counted as such when a new Chief Executive enters office. To identify Chief Executives I shall use the LEAD dataset introduced by Ellis, Horowitz and Stam (2015). To code leadership turnovers, years where the Chief Executive changes shall be coded as 1, while years where the Chief Executive remains the same shall be coded as 0. However it is important to consider the nature of leadership turnovers, as Mattes, Leeds and Carroll (2015) pointed out, leadership transitions without a change in source of leadership support (henceforth named SOLS), did not have a statistically significant impact on UNGA voting patterns, while leadership transitions accompanied by a SOLS change were statistically significant. I shall thus use the CHISOLS dataset developed by Mattes, Leeds and Matsumura (2016) to identify SOLS changes. Mattes et al. (2015) conceptualised a leader’s SOLS in democracies as “those who vote for or associate with the leader’s party” (p. 285). A change in party affiliation in the Chief Executive shall thus be treated as a leadership turnover based on a SOLS change. Predesignated successors, such as vice-presidents, or legal next in line, taking over office shall not me counted as a SOLS change. The same goes for leaders elected by the same or similar voting blocs (regional, ethnic, class, etc). Using these rules, two dummy variables are created, one for leadership turnovers based on SOLS changes, called SOLS Change, and one for leadership turnovers not accompanied by any SOLS changes, called Other Leader Transitions. SOLS Change shall be coded 1 if a leadership turnover took place in that country year, and the new leader has a different SOLS. Other Leader Transitions is coded
[10]
1 if a leadership turnover took place in that country year, but the SOLS remained the same. These are mutually exclusive dummy variables.
Second, I am interested in leadership turnover in two specific democratic political institutions, namely presidential and parliamentary systems. This shall be coded based on the variable definitions given by Cruz, Keefer and Scartascini (2018) in their Database of Political Institutions. Cruz et al. (2018) define a presidential system as “systems with presidents who are elected directly or by an electoral college (whose only function is to elect the president.” Systems with no prime ministers are also counted as presidential, and systems with both prime ministers and presidents are subjected to testing by further qualifications: (a) can the president veto, (b) can the president appoint and dismiss the prime minister or other ministers, (c) can the president dissolve parliament (Cruz, et al., 2018). If (a) is true, or if (b) and (c) are true, then the system is classified as presidential otherwise the system is classified as parliamentary. All presidential cases are coded 0. Cruz et al. (2018) define parliamentary systems as systems where the chief executive is elected by the legislature, and are coded as 2. Cruz et al. (2018) also use a third category, semi-presidential where the president is elected by the legislature, but the legislature cannot easily recall the president (2/3 majority needed) or needs to dissolve itself to force the president out of office, these cases are coded as 1.
Since I am interested only in the presidential and parliamentary systems, I have decided to classify semi-presidential systems as parliamentary, because they still follow the earlier set out rule of parliamentary systems being characterised by a legislature electing the executive. While it may be harder to get rid of such a parliamentary elected president, it is still possible. I thus code a dummy variable labelled Presidential Democracy where all countries coded as 0 in the DPI2017 dataset will be coded as 1, and all other cases labelled 1 and 2 according to the DPI2017 dataset will be coded as 0.
Finally, my research is limited to only democratic countries, thus I shall filter out all country years that are not democratic. To decide the democratic level of a country, I shall use the Polity2 scores from the POLITY IV project (Marshall, Jaggers & Gurr, 2010). All country years measuring 6 and above are marked as democratic. All country years measuring a 5 or below shall be dropped This way I shall include only observable country years that are democratic according to this measurement, leaving 2038 country years spread over 103 as democratic.
[11] Control Variables
I shall control for factors that potentially affect both UNGA voting, and leadership turnover (both SOLS and non-SOLS). First I shall control for alliance ties with the United States throughout the entirety of 1975-2008. During the Cold War, alliance ties could possibly offer constraints on both domestic and foreign political change (Mattes et al., 2015). After the Cold War, such affects could still persist, as close alliance ties to the US might carry certain conditions and expectations. This control variable is coded based on the ATOP data from Leeds, Ritter, Mitchell, and Long (2002) and labelled as US Ally. All country years where countries were committed in a defensive pact with the United States are coded as 1, and all other cases are coded as 0.
Secondly, I control for the short term effect of the end of the cold war. As Mattes et al. (2015) point out, the fall of the Iron Curtain was an event that lead to significant international realignments, and to domestic political changes. They warn against falsely attributing changes in UNGA voting to leadership turnover and SOLS changes, when it was this unique event that might cause such changes (Mattes et al., 2015). In the research done by Mattes, Leeds, and Carroll (2015) the end of the Cold War did not appear to have a significant effect on the UNGA voting, however, I am curious to see if this is also the case when only looking at democratic systems. Thus, for the purpose of my research I shall still control for a possible impact the end of the Cold War might have had on both UNGA voting, and leadership turnovers. The dummy variable labelled Cold War End is coded 1 if the observed country year coincides with the period of 1989 through 1991.
Finally, I shall control for changes in economic growth measured in annual percentage change in GDP. As discussed earlier in this paper domestic and international economic interests might be major driver of foreign policy, and policy change (Snyder, 1991; Narizny, 2003, 2007; Solingen, 2007). Economic prosperity, stagnation, or even shrinkage, might thus effect UNGA voting, but also leadership turnovers and SOLS changes, as the economy is often a major factor in any democratic election campaign. The data for economic growth is retrieved from the World Development Indicators from 2018, measured in annual percentage change in GDP. A dummy variable is then created where all values from zero and up are codes as 0, while all values below zero are codes as 1. This dummy variable is consequently labelled as Negative Economic Growth.
[12]
Empirical Results
Since my dependent variable is continuous – the absolute change in the UNGA ideal points from t–1 to t – I shall be using OLS regression. All models include country fixed effects in order to account for unmeasured factors associated with particular countries. In order to test my two hypotheses, I run 6 models, 2 models focussing on Hypothesis 1, and 4 models focussing on hypothesis 2.
Table 1 shows the two models used to test Hypothesis 1. The basic model (Model 1) includes the main theoretical variable indicating regime type, Presidential Democracy, as mentioned in the research design, this is a dummy variable. The control model (Model 2) includes the control variables, Negative Economic Growth, Cold War End, and US Ally.
Models 1 and 2 both indicate that Presidential Democracy has a statistically significant positive effect on UNGA vote change, although the effect is ever so slightly tempered when the control variables are included in model 2, indicating that there is significantly more year-to-year change in the UNGA voting patterns of presidential democracies than parliamentary democracies. This is in line with Hypothesis 1.
The inclusion of the control variables does not affect my conclusion, as the results for my independent variable remain statistically significant. Interestingly, the control variable Cold War End reached conventional levels of statistical significance, The period from 1989 to 1991 is thus rightly thought to be a period of major upheaval in the international system and should be associated with significant positive changes in UNGA voting when looking merely at democratic regimes.
Table 1. Absolute Change in UNGA Voting as a Function of Presidential Democracy, 1975-2008. (1) Basic Model (2) Control Model Presidential Democracy 0.038** (0.016) 0.036** (0.017) Negative Economic Growth 0.000 (0.007)
Cold War End 0.021** (0.008)
US Ally 0.009 (0.017)
Constant 0.058*** (0.017) 0.052** (0.020)
Observations 2038 2038
Number of Countries 103 103
(***p<0.01, **p<0.05, *p<0.1; two-tailed). (All models include country fixed effects).
[13]
The other control variables do not reach conventional levels of statistical significance, the coefficients are also not in the expected directions, as US Ally appears to be associated with positive changes in UNGA voting patterns, instead of being a constraining factor. Likewise, Negative Economic Growth, does not appear to have any impact at all. I have tried to see if the inclusion of positive economic growth, using the raw percentage change in GDP, or using the absolute values of the percentage change in GDP, would render different results, but all remain statistically insignificant, and do not change the coefficient. Thus, when merely looking at the political system, economic figures do not appear to be impactful when explaining UNGA voting patterns.
Table 2 includes the four models used to test Hypothesis 2. Model 3 includes the main theoretical variables of interest: SOLS Change, and Other Leader Transitions. Model 4 adds the control variables: Negative Economic Growth, Cold War End, and US Ally. To establish how the effects of both SOLS Changes and Other Leader Transitions differ between presidential and parliamentary democracies, I estimate two models that interact SOLS Change and Other Leader Transition with Presidential Democracy. Model 3 includes only the main theoretical variables from the first two models and includes the interaction variables. Model 4 includes the control variables to the interactive model.
Table 2. Absolute Change in UNGA Voting as a Function of SOLS Change, Other Leader Transitions, and Presidential Democracy, 1975-2008. (3) Basic Model (No Controls) (4) Basic Model (With Controls) (5) Interactive Model (No Controls) (6) Interactive Model (With Controls) SOLS Change 0.008 (0.006) 0.007 (0.006) 0.007 (0.008) 0.008 (0.008) Other Leader Transitions 0.004 (0.008) 0.004 (0.008) 0.009 (0.010) 0.009 (0.010) SOLS Change × Presidential Democracy 0.001 (0.012) -0.001 (0.012) Other Leader Transitions × Presidential
Democracy
-0.016 (0.018) -0.016 (0.018) Negative Economic Growth -0.001 (0.007) -0.001 (0.007) Cold War End 0.020** (0.008) 0.020** (0.008) US Ally 0.020 (0.016) 0.021 (0.016) Constant 0.088*** (0.012) 0.070*** (0.017) 0.088*** (0.012) 0.070*** (0.017) Observations 2038 2038 2038 2038 Number of Countries 103 103 103 103 (***p<0.01, **p<0.05, *p<0.1; two-tailed).
[14]
Models 3 and 4 indicate that the coefficients for both SOLS Change and Other Leader Transitions do not reach statistically significant levels (SOLS Change p = 0.208 for Model 3 and p = 0.211 for Model 4; Other Leader Transitions p = 0.596 for Model 3 and p = 0.636 for Model 4), and thus do not significantly affect UNGA voting patterns. Therefore, years wherein a leadership transition takes place are not to be associated with a change in UNGA voting patterns. The coefficients do, however, move in the expected positive direction.
Adding the control variables in Model 4 does not change the results of the independent variables, which remain insignificant. I do observe the same behaviour as in Table 1, as Cold War End is once again statistically significant, while the other two control variables do not reach statistically significant levels. The coefficient of Negative Economic Growth does, however, move in an unexpected negative direction in this model. If there was to be any significant impact of economic growth on UNGA voting patterns, I expected negative growth to lead to more change in UNGA voting patterns, not less. Again, I tried different methods of registering economic change to see if it changed the results, but the coefficient did drastically change, nor did it become statistically significant.
When observing models 5 and 6, Hypothesis 2 appears to be unsubstantiated, as none of the theoretical variables; SOLS Change (p = 0.347 for Model 5 and p = 0.307 for Model 6), Other Leader Transitions (p = 0.364 for Model 5 and p = 0.380 for Model 6), and none of the interactive variables; SOLS Change × Presidential Democracy (p = 0955 for Model 5 and p = 0.939 for Model 6) and Other Leader Transitions × Presidential Democracy (p = 0.388 for Model 5 and p = 0.374 for Model 6), reach levels of statistical significance. The inclusion of control variables in Model 6 does seem to change the coefficient of the interaction variable SOLS Change × Presidential Democracy from positive to negative, but the effect remains insignificant. The control variables continue to follow the trends observed in all the previous models.
Leadership turnover in democracies does not appear to have a statistically significant effect on UNGA voting patterns. Changes in domestic sources of leader support do also not appear to have a stronger, or even significant, effect on UNGA voting patterns in presidential democracies compared to parliamentary democracies. These observations have therefore lead me label Hypothesis 2 as unsupported.
Overall, I conclude that Hypothesis 1 is supported, while Hypothesis 2 remains unsupported. The political system of presidential democracies has a statistically significant positive impact on change in UNGA voting patterns, meaning that presidential democracies exhibit significantly more year-to-year change in UNGA voting patterns than parliamentary
[15]
democracies. Meanwhile evidence shows that leadership turnover, and corresponding domestic sources of leader support do not have a statistically significant impact on UNGA voting patterns, nor is this effect any stronger in presidential democracies when compared to parliamentary democracies. If not leadership turnover is responsible for changes in UNGA voting patterns, than there must be something else responsible. Perhaps institutions, and not leaders are responsible for change in foreign policy preferences when looking at democratic regimes.
Discussion
IR scholarship has been dominated by the dominance of countries, and systems as units of analysis, but recently the leader has entered the arena as an additional unit of analysis to consider. Scholars argue that leaders should even be considered the fundamental unit of analysis, not the structure of the international system and of states. The role of leaders in IR has been extensively discussed, and literature has pointed out that leaders have, perhaps unsurprisingly, far more manoeuvrability in non-democratic regimes, than in democratic regimes, but the diversity of the democratic subset, and possible implications of leadership in democratic regimes, has never been properly researched. I sought to remedy this. The basic premise of my paper was that if IR scholars want to understand the impact of leaders and leadership turnover on foreign policy changes, we need to consider the domestic structures leaders operate in.
Therefore, I sought to research the impact of leadership turnover in the two most prominent political systems in the democratic subset, presidentialism and parliamentarism, on foreign policy change, using UNGA voting over time to indicate foreign policy change. My research was modelled after the research done by Mattes, Leeds, and Carroll (2015) on leadership turnover and foreign policy change, wherein democracies and non-democracies were studied. Even though I tried to mirror the previously mentioned research as closely as possible, some limitations should be discussed. While these are interesting points I would like to change in my research given the ideal circumstances, I do not believe these limitations have significantly affected the credibility of my findings.
First, as mentioned in the theory section of this paper, the theoretical evidence for my expectations of the institutional structures of presidential systems, and the role of leadership turnover regarding foreign policy change is entirely based on literature dedicated to the U.S. presidential system. Literature discussing other presidential democracies, such as South American presidential democracies, are all written in languages I do not master. Perhaps a partnership with researchers possessing greater foreign language skills than I might remedy this problem. Other literature discussing presidential systems in general also operate from the
[16]
assumption that the U.S. presidential system can be generalised, and are thus also based on a single case. Regardless, I do not expect this to tarnish the credibility of my research, nor my findings.
Second, my research only spans a limited time period and is not completely up to date. This research spans the period of 1975 up until 2008. Due to the various datasets I used to compile my own data selection, and to ensure the maximum availability of data for every case, I decided to limit myself to this time period. Essential information such as the Database of Political Institutions 2017 – used for identifying political systems – and the CHISOLS 4.0 dataset – used to identify leadership turnover – was missing prior to 1975 (DPI2017) and after 2008 (CHISOLS 4.0). Unfortunately, this means that recent developments in international politics, are not captured in my data.
Third, while my research focusses on the domestic institutional restraints on leaders’ manoeuvrability regarding foreign policy, I have only taken into account the raw political system, and not possible other veto players. Nor have I taken into account the make-up of the legislature. Executive leadership might change, but this does not mean that the leader always enjoys a majority backing in the legislature. Including this into my research would, however, have widened the scope too much.
Interestingly, my findings show that leadership turnovers in democracies do significantly impact foreign policy change, while the presence of a presidential political system does have a significant positive impact on foreign policy changes. These results are in direct opposition to the expectations I had after having reviewed the Mattes et al. (2015) research, wherein the impact of SOLS changes on foreign policy changes were deemed to be statistically significant. Perhaps the exclusion of non-democratic regimes has contributed to the different results. This might indicate that the institutional structures present in democratic regimes might be resilient enough to render the impact of leadership turnover on foreign policy change irrelevant, as opposed to non-democratic regimes where the leader might have a bigger impact on policy and thus leadership turnover in non-democratic regimes would arguably also have a bigger impact on foreign policy. This is corroborated by Mattes et al. (2015, p. 287), as their results show that while SOLS changes in democracies and non-democracies are both statistically significant, the effect is larger in non-democratic regimes, which might depict the constricting influence of institutional structures in democracies on the effects of leadership turnover and the influence of leaders in general.
My expectation that presidential democracies were more inconsistent in their UNGA voting patterns, and thus in their foreign policy, was based on the assumption that foreign policy
[17]
in presidential systems was more susceptible to leadership turnover, so while hypothesis 1 might be corroborated by the data, it is not because of the assumed causal effects. Something else seems to be causing the difference in foreign policy consistency between presidential and parliamentary democracies. More research needs to be done into this.
My speculation on the matter can be traced back to the literature. The longevity of the coalition system at the core of parliamentary democracies – as discussed in the theory section of this paper – might be an explanatory factor for the foreign policy consistency observed in parliamentary democracies. Conversely, presidential democracies might be more prone to foreign policy swings due to the constant battle between the executive and legislative branches over the foreign policy setting prerogative (Trimble, 1989; Tsebelis, 2000; Peake, 2001; Horowitz, 2018). These two opposite views on the exact constitutional nature of the foreign policy agenda can be a contributing factor to policy inconsistency. Whenever one branch gains the upper hand over the other branch, policy preferences might thus switch. As pointed out, both branches might also represent different constituencies, explaining possible differences in foreign policy preferences. This problem does not exist in parliamentary democracies, where government and parliament generally represent the same constituency. Structural, and not idiosyncratic causes might thus lie at the root of the foreign policy inconsistency of presidential democracies. The precise factors causing the inconsistency regarding foreign policy present in presidential democracies requires further research.
Furthermore, my analysis shows that when looking solely at democratic regimes, the end of the Cold War – the period from 1989 to 1991 – does have a significant impact on change in UNGA voting patterns, unlike the models tested by Mattes et al. (2015), where they were surprised to find that the end of the Cold War seemed not to be significant. A significant realignment of the foreign policy priorities in democracies during the end of the Cold War seems likely, as foreign policies of democratic states, most of whom were aligned to the West, were built around the supposed threat of the Soviet Union and its allies. The end of the Cold War changed this status quo, so it makes sense that this is represented in the statistical results.
My research seems to give credence to a structuralist approach regarding the role of leaders in foreign policy, as opposed to the pre-eminence of leaders as the fundamental unit of analysis. The role of leaders can only be understood through their relationship to the domestic structures wherein they operate, especially when looking at democratic regimes, as they can be considered far less idiosyncratic. Future research into the exact nature of this relationship is paramount and might offer a clearer understanding of foreign policy change in democratic
[18]
regimes, which will allow scholars to form a more complete picture of the formation of foreign policy, and the nature of foreign policy change.
Literature:
Bailey, M.A., Strezhnev, A. & Voeten, E. (2017). Estimating dynamic state preferences from United Nations voting data. Journal of Conflict Resolution, vol 61(2), pp. 430-456. Acquired through: https://dataverse.harvard.edu/dataset.xhtml?persistentId=hdl: 1902.1/12379
Capling, A. & Nossal, K.R. (2003). Parliament and the Democratization of Foreign Policy: The Case of Australia’s Joint Standing Committee on Treaties. Canadian Journal of Political Science, vol 36(4), pp. 835-855.
Christensen, E.J. & Redd, S.B. (2004). Bureaucrats Versus the Ballot Box in Foreign Policy Decision Making: An Experimental Analysis of the Bureaucratic Politics Model and the Poliheuristic Theory. Journal of Conflict Resolution, vol 48(1), pp. 69-90. Cesi, C. Keefer, P. & Scartascini, C. (2018). Database of Political Institutions 2017
(DPI2017). Inter-American Development Bank. Numbers for Development. Acquired through: https://mydata.iadb.org/Reform-Modernization-of-the-State/Database-of-Political-Institutions-2017/938i-s2bw
Debs, A. Goemans, H.E. (2010). Regime Type, the Fate of Leaders, and War. American Political Science Review, vol 104(3), pp. 430-445.
Dreher, A. & Jensen, N. (2013) Country or Leader? Political Change and UN General Assembly Voting. European Journal of Political Economy, vol 29, pp 183–196. Ellis, C., Horowitz, M. & Stam, A. (2015). Introducing the LEAD data set.
International Interactions, vol 41(3), pp. 718-741.
Gartzke, E. & Gleditsch, K.S. (2004) Why Democracies May Actually Be Less Reliable Allies. American Journal of Political Science, vol 48(4), pp. 775–795.
Horowitz, M.C. (2018). Leaders, Leadership, and International Security. In A. Gheciu & W.C. Wohlforth (Eds), The Oxford Handbook of International Security (pp. 246-258). Oxford: Oxford University Press.
Ihalainen, P. & Matikainen, S. (2016). The British Parliament and Foreign Policy in the 20th Century: Towards Increasing Parliamentarisation? Parliamentary History, vol 35(1), pp. 1-14.
Jervis, R. (2013). Do Leaders Matter and How Would We Know? Security Studies, vol 22(2), pp. 153-179.
[19]
Kaarbo, J. (1996). Power and influence in foreign policy decision making: The role of junior coalition partners in German and Israeli foreign policy. International Studies
Quarterly, vol 40(4), pp. 501-530.
Landler, M. (2017, December 28). Trump, the Insurgent, Break With 70 Years of American Foreign Policy. The New York Times. Retrieved from:
https://www.nytimes.com/2017/12/28/us/politics/trump-world-diplomacy.html Leeds, B.A., Mattes, M. & Vogel, J.S. (2009) Interests, Institutions, and the Reliability of
International Commitments. American Journal of Political Science, vol 53(2), pp. 461–476.
Leeds, B.A., Ritter, J.M., Mitchell, S.M. & Long, A.G. (2002). Alliance Treaty Obligations and Provisions, 1815-1944. International Interactions, vol 28. pp. 237-260. Retrieved from: http://www.atopdata.org/data.html.
Marshall, M.G., Jaggers, K., & Gurr, T.R. (2010) Polity IV Project: Dataset Users’ Manual. College Park: University of Maryland.
http://www.systemicpeace.org/inscr/p4manualv2017.pdf .
Mattes, M. (2012) Democratic Reliability, Precommitment of Successor Governments, and the Choice of Alliance Commitment. International Organization, vol 66(1), pp. 153–172.
Mattes, M. Leeds, B.A. & Carroll, R. (2015). Leadership Turnover and Foreign Policy Change: Societal Interests, Domestic Institutions, and Voting in the United Nations. International Studies Quarterly, vol 59(2), pp. 280-290.
Mattes, M. Leeds, B.A. & Matsumura, N. (2016). Measuring Change in Source of Leader Support: The CHISOLS Dataset. Journal of Peace Research, vol 53(2), pp. 259-267. Narizny, K. (2003). Both Guns and Butter, or Neither: Class Interests in the Political
Economy of Rearmament. American Political Science Review, vol 97(2), pp. 203–220. Narizny, K. (2007). The Political Economy of Grand Strategy. Ithaca, NY:
Cornell University Press.
Nolan, C.J. (1985). The Influence of Parliament on Human Rights in Canadian Foreign Policy. Human Rights Quarterly, vol 7(3), pp. 373-390.
Oppermann, K. & Brummer, K. (2017). Veto Player Approaches in Foreign Policy Analysis. Oxford Research Encyclopedia of Politics.
DOI: 10.1093/acrefore/9780190228637.013.386
[20]
Constraints on International Conflict Behavior in Parliamentary Democracies. International Interactions, vol 30(1), pp. 1-24.
Peake, J. (2001). Presidential Agenda Setting in Foreign Policy. Political Research Quarterly, vol 54(1), 69-86.
Potrafke, N. (2009) Does Government Ideology Influence Political Alignment with the US? An Empirical Analysis of Voting in the UN General Assembly. The Review of
International Organizations, vol 4(3), pp. 245–268.
Tidmarch, C.M. & Sabatt C.M. (1972). Presidential Leadership Change and Foreign Policy Roll-call Voting in the US Senate. Western Political Quarterly, vol 25(4),
pp. 613-625.
Saunders, E. (2017). No Substitute for Experience: Presidents, Advisers, and Information in Group Decision Making. International Organization, vol 71(1), pp. 219-S247. doi:10.1017/S002081831600045X
Snyder, J. (1991) Myths of Empire: Domestic Politics and International Ambition. Ithaca, NY: Cornell University Press.
Solingen, E. (2007) Nuclear Logics: Contrasting Paths in East Asia and the Middle East. Princeton, NJ: Princeton University Press.
Strong, J. (2015). Interpreting the Syria vote: Parliament and British foreign policy. International Affairs, vol 91(5), 1123-1139.
The Editorial Board. (2019, January 26). Saving NATO. The New York Times. Retrieved from: https://www.nytimes.com/2019/01/26/opinion/sunday/nato-trump.html
Trimble, P. (1989). The President's Foreign Affairs Power. American Journal of International Law, vol 83(4), pp. 750-757.
Tsebelis, G. (2000). Veto Players and Institutional Analysis. Governance- an International Journal of Policy, Administration and Institutions, vol 13(4), 441-474.
Wolford, S. (2007). The turnover trap: New leaders, reputation, and international conflict. American Journal of Political Science, vol 51(4), 772-788.
World Bank, World Development Indicators. (2018). GDP growth (annual %) [Data file]. Retrieved from: https://data.worldbank.org/indicator/ny.gdp.mktp.kd.zg
Wu, C. X., & Wolford, S. (2018). Leaders, States, and Reputations. Journal of Conflict
