The International and Domestic Sources of Bipartisanship in U.S. Foreign Policy

Abstract

Foreign policy scholars have long emphasized bipartisanship in foreign policymaking, particularly in the context of presidential appointments to the foreign policy bureaucracy, emphasizing the role of international crises and national security in affecting bipartisanship. In spite of their enduring nature, few systematic analyses of these claims exist. This study addresses these gaps using new data on appointees to the foreign policy bureaucracy. The results challenge the conventional wisdoms emerging from the foreign policy literature regarding the importance of factors like war, instead pointing to the importance of domestic political factors in affecting bipartisanship in appointments to the foreign policy bureaucracy.

Keywords

foreign policy bipartisanship bureaucracy Cold War consensus

In the summer of 1940, President Franklin Roosevelt asked Henry L. Stimson to be his Secretary of War. Stimson, a seventy-three-year-old lawyer from New York, had a lengthy history of government service. Having previously served as Governor General of the Philippines, Secretary of State, and a previous tour as Secretary of War, Stimson was unquestionably qualified for the position. His acceptance was conditional upon two things. First, that Roosevelt’s nominee for secretary of the navy, Frank Knox, also accepts his own nomination. Second, Stimson should retain the power to appoint his own subordinates in the War Department.¹ These conditions having been met, Stimson began one last tour of government service. Beyond his rich experience in government, the most important of Stimson’s features was that he was a Republican. Like Stimson, Frank Knox and Stimson’s eventual subordinates—Robert Lovett and John McCloy—were all Republicans and were all appointed to administer a major war effort in a Democratic administration.

Although the importance of bipartisanship in foreign policy has long been highlighted by scholars, little systematic work on the subject exists. This study expands upon previous work in three key ways. First, while many of the most important episodes of bipartisanship in foreign policymaking have involved executive appointees, we still know little about how presidents use such appointments in a systematic sense. Stimson’s appointment in particular is viewed by many scholars as marking the beginning of an era characterized by especially high levels of bipartisan cooperation in foreign policy.² Such appointments functioned as important tools in shaping executive relations with Congress and the public during World War II and the early years of the Cold War. Stimson and many other appointees helped to create the foundations for the liberal internationalist policies that formed the basis of American foreign policy for decades. Appointees like Robert Lovett, John McCloy, and Dean Acheson all helped to secure domestic political support for policies that marked a dramatic turn in U.S. foreign policy. The use of bipartisan appointments in this process was key—Kupchan and Trubowitz (2007, 13) state that President Roosevelt “sought to make Republicans stakeholders in his foreign policy by appointing members of the opposition to important foreign policy posts.” Accordingly, this study has clear implications for better understanding the strategies that presidents can employ to affect executive–legislative relations.

Second, our understanding of bipartisanship is largely rooted in the foreign policy literature, and most studies on the subject typically fall into two groups. The first is more qualitative, relying on anecdotal evidence rather than systematic analyses. This group also places substantial emphasis on bipartisanship in the context of the foreign policy bureaucracy (e.g., Destler, Gelb, and Lake 1985; Halberstam [1972] 1992; Isaacson and Thomas 1986; Neustadt [1960] 1991). The second, more recent, group has focused on Congressional voting patterns (Chaudoin, Milner, and Tingley 2010; Kupchan and Trubowitz 2007; McCormick and Wittkopf 1990; Meernik 1993), executive–legislative relations (Kupchan and Trubowitz 2007; McCormick and Wittkopf 1990; Meernik 1993), or public/elite opinion (Holsti and Rosenau 1984; Kupchan and Trubowitz 2007). While this second group has improved our understanding of bipartisanship through the analysis of systematic patterns and relationships, they offer little insight into the conditions influencing bipartisanship in presidential appointments to the foreign policy bureaucracy. Accordingly, this study applies a more rigorous and systematic analysis to an area of study that, while central in shaping our understanding of bipartisanship, has been characterized primarily by anecdotal evidence.

Third, although the importance of bipartisanship has been highlighted by the foreign policy literature, it focuses on how international crises or major international events, such as Vietnam and the end of the Cold War, have affected bipartisanship, relegating domestic political conditions to a more minor position.³ However, there is reason to believe that domestic forces also matter. Trubowitz and Mellow (2005) have found that bipartisanship in Congress fluctuates according to the strategic incentives faced by legislators and varies in response to the composition of government. Research has also found that the processes governing presidential appointments are impacted by the strategic calculations of the president and legislators (Bond, Fleisher, and Krutz 2009; Krutz, Fleisher, and Bond 1998; McCarty and Razaghian 1999). Accordingly, I control for domestic political factors in addition to the factors traditionally emphasized by the foreign policy literature, providing us with a more complete understanding of the forces determining bipartisanship.

This article proceeds as follows. First, I review the factors that have been argued to affect bipartisanship. Second, using new data set, I test several hypotheses to evaluate competing domestic and foreign policy-based explanations underlying presidential decisions to be more or less bipartisan in their appointments to the foreign policy bureaucracy. Contrary to conventional wisdom, I find little evidence that Vietnam or the end of the Cold War is associated with a decline in the level of bipartisanship exhibited by presidents. All else being equal, I find that Republicans and Democrats have actually converged in their tendency to make bipartisan appointments over time. I also find little evidence that periods of conflict are associated with higher levels of bipartisanship. However, the results do indicate that longer conflicts see a decline in bipartisanship. Alternatively, I find evidence that bipartisanship is correlated with the level of support a president enjoys in Congress and that this effect is conditional upon the level of political polarization. Finally, I conclude with a discussion of the results and suggest avenues for future research.

Sources of Bipartisanship

The factors argued to influence the level of bipartisanship exhibited by presidents can be divided into two general categories. First, foreign policy scholars have long highlighted the role of international circumstances. Many authors have pointed to major events such as World War II, the Vietnam War, and the end of the Cold War in affecting bipartisanship across all areas of American political life (e.g., Destler, Gelb, and Lake 1985; Holsti and Rosenau 1984; Isaacson and Thomas 1986; Kupchan and Trubowitz 2007). Thus, bipartisanship is primarily the product of international pressures that allow for more or less latitude in the degree to which partisan preferences influence foreign policy. Second, the president’s decision to make bipartisan appointments can also be viewed as a strategic choice, intended to maximize the president’s political leverage vis-à-vis Congressional opposition, for example (Berger 1975; Trubowitz and Mellow 2005). In contrast to the foreign policy literature, these stimuli are the product of domestic political circumstances.

International Sources

Foreign policy scholars have long argued that international threats and crises (e.g., war) motivate greater bipartisanship. There is a general view that it can be politically risky for politicians to be patently partisan in international affairs—particularly in matters of national security. During these periods, national leaders set aside partisan interests in an effort to demonstrate national unity. The use of military force in particular is held to lead to a rallying effect, where the public and politicians attempt to demonstrate their support for military personnel and that they are placing the national interest above more narrow partisan interests.⁴ As commander-in-chief of the military, and as the only office elected by the country as a whole, presidents have been viewed as especially sensitive to the need to demonstrate unity during times of crisis. Consequently, presidents may be more bipartisan for the sake of national unity when faced with significant international threats and crises (Halperin, Clapp, and Kanter 2007; Hilsman 1987; Neustadt 1970; Trubowitz and Mellow 2005).

There is anecdotal evidence indicating that presidents have responded to such events by emphasizing greater bipartisanship—often through appointments to the foreign policy bureaucracy. Many studies have highlighted the Roosevelt administration’s response to the emerging war in Europe as an example of such behavior. Congress had taken steps to block U.S. involvement in the escalating European conflict. Many legislators were wary of becoming involved in another major European war, with the Western element of the Republican Party being particularly opposed to U.S. involvement (see Herring 2008; Trubowitz 1998). However, Roosevelt sought to build domestic support for U.S. action by appointing Republicans to lead the administration’s war effort. Henry Stimson and Frank Knox were appointed to head the War Department and Navy Department, respectively, thus placing both branches of the military in the hands of members of the Republican opposition. Bipartisan appointments continued to be an important feature of the Truman administration as it sought to deal with the emerging Soviet threat. Many members of Congress did not initially share the Truman administration’s sense of the danger posed by the Soviet Union, assuming that the United States would return to pre-war levels of military spending and force posture. Foreign aid programs in particular were viewed as wasteful by many Congressional Republicans—many of whom referred to the provision of aid to Greece and Turkey as “Operation Rathole” (Bonds 2002). Truman’s response was to delegate substantial authority to key bureaucratic officials who were charged with convincing a skeptical Congress of the need to increase spending on aid and the military. Truman’s appointment of Lovett, a Republican, has been noted as a particularly important gesture of bipartisanship. Lovett worked closely with members of Congress—particularly Republican Senator Arthur Vandenberg, Chairman of the Senate Foreign Relations Committee—to rally support for these foreign policy innovations (Bonds 2002; Briggs 1994). President Barack Obama’s decision to retain George W. Bush’s Secretary of Defense, Robert Gates, during a critical juncture in the wars in Afghanistan and Iraq provides a more recent example of such behavior. These examples lead me to my first hypothesis:

Hypothesis 1: International crises are positively correlated with the likelihood of observing a bipartisan appointment.

Foreign policy scholars also argue that particular watershed events have substantially reduced bipartisan cooperation in foreign policymaking. Beginning with the general notion that foreign policy was guided by a bipartisan foreign policy “Establishment” during the early years of the Cold War,⁵ this literature argues that the bipartisan consensus has been damaged over time. Several scholars have argued that the Vietnam War demolished the foundations of bipartisan cooperation. Perhaps most prominently, Destler, Gelb, and Lake (1985, 19) state that “the endless and seemingly hopeless agony of the Vietnam War destroyed the consensus, sprayed power out from the center toward the political extremes, and made forging majorities a very trying affair.” Ultimately, the war is viewed as causing a growing rift among political elites and the mass public regarding the best strategies for engaging in world affairs (Destler, Gelb, and Lake 1985; Hilsman 1987; Holsti and Rosenau 1984; Wittkopf 1990). Others argue that the end of the Cold War caused a breakdown in bipartisanship by eliminating the raison d’être for bipartisan cooperation and removing the greatest constraint on freer U.S. action in the international arena. This allowed presidents to pursue more partisan policies with less risk of repercussions in the international sphere, and without appearing to place partisan interests above national security at home. Specifically, Republicans have been viewed as pursuing an increasingly unilateral and militant foreign policy strategy, while Democrats have tended to favor more multilateral and less militaristic policies (Ikenberry 2002, 2011; Kupchan 2003; Kupchan and Trubowitz 2007).⁶

Hypothesis 2: The periods following watershed events in foreign policymaking should see a lower likelihood of observing a bipartisan appointment than the preceding time periods.

Domestic Sources

There is reason to believe that domestic political conditions also influence presidential decisions to make bipartisan appointments to the foreign policy bureaucracy. Evidence suggests that the president’s party may be linked to such decisions, with Democratic presidents appearing to have embraced bipartisanship more than their Republican counterparts. During the early Cold War period, the Democratic Party was more uniformly in favor of an internationalist foreign policy, while the Republican Party was divided between its Northeastern internationalists, and its Western element, which opposed international activism. Even internationalist Republicans required coaxing by Democrats. Alternatively, as the Republican presidential nomination was usually secured by the party’s Northeastern wing, Republican presidents had to focus on securing support from within their own party, rather than reaching across party lines to Democrats, whose economic interests often predisposed them to support international engagement (Trubowitz 1998). Second, the early Cold War period also saw fears of communism lead the Eisenhower administration to purge many State Department bureaucrats viewed as having communist sympathies. Against the backdrop of McCarthyism, appointing individuals with even remotely liberal political beliefs was politically dangerous (see Acheson 1969; Gaddis 2011; Isaacson and Thomas 1986). The emphasis on bipartisan appointments returned under President Kennedy, who consulted closely with senior Republican statesmen like Robert Lovett and filled key administration positions with Republicans. These partisan differences have likely persisted over time. Although Vietnam is argued to have driven Republicans and Democrats further from the center, scholars have suggested that Democrats were still more prone to bipartisanship than Republicans. Destler, Gelb, and Lake (1985, 119) write that

while the liberal–left in the Carter administration could not govern without conservatives, the conservative right in the Reagan administration thus could and did get along without any liberals . . . The striking feature of the Reagan administration was its ideological purity.

And while some scholars have argued that the consensus did not truly break down until after the end of the Cold War, they generally echo the basic shifts in partisan preferences toward foreign policy (Ikenberry 2011; Kupchan and Trubowitz 2007).

Previous empirical work provides some support for the possibility that such partisan differences exist. Fleisher and Bond (1988) and Meernik (1993) find that Democratic legislators tend to be more likely to support the foreign policy initiatives of Republican presidents than vice versa. Although not synonymous with party, political psychologists have found evidence indicating that liberals tend to be more open-minded and open to diversity than conservatives. This may translate into increased bipartisanship among Democrats (e.g., Carney et al. 2008).

Hypothesis 3: Democratic presidents are more likely to make bipartisan appointments than Republican presidents.

Several studies have shown how domestic political and economic conditions affect the foreign policy choice of presidents (Clark 2000; Fordham 1998; Howell and Pevehouse 2007; Kriner 2010). Research on the effects of domestic political and institutional forces on bipartisanship in foreign policy is more limited. Meernik (1993) finds mixed support for the influence of domestic variables, such as presidential popularity and presidential positions on legislation in motivating Congressional bipartisanship. Trubowitz and Mellow (2005) have also shown domestic forces, such as divided government, to affect bipartisanship in Congressional voting more broadly. Although these studies have focused on Congressional voting, much of their theoretical logic can also apply to presidential decisions to make bipartisan appointments. Presumably, presidents want to advance their policy agenda, and institutional factors may constrain or enable presidents to varying degrees. Among such factors, relations with Congress are an important influence. A president who enjoys greater support in Congress is more likely to be able to push favored policies through. For example, a major obstacle to the Truman administration’s European aid package was the fact that the eightieth Congress was under Republican control, leading Truman to make bipartisan appointments to strengthen executive–legislative relations (Bonds 2002; Briggs 1994).

Previous research has shown that the level of partisan support a president enjoys in Congress does affect presidential actions, such as constraining the president’s ability to use military force as a foreign policy tool (Clark 2000; Howell and Pevehouse 2007; Kriner 2010). Work by McCarty and Razaghian (1999) indicates that divided government is correlated with longer nomination processes, which can hinder a president’s ability to effectively develop and implement policy.⁷ However, Trubowitz and Mellow (2005) argue that periods of divided government provide moderates in Congress with greater influence as they represent an important swing group. When the president’s party is in the minority, or only holds the majority by a small margin, the ability to work with moderate opposition legislators is key. Thus, when a president faces less support in Congress, he may be inclined to make bipartisan appointments in an effort to credibly signal his willingness to work with Congress and to build moderate coalitions. Furthermore, given that members of Congress often rely on the expertise of bureaucrats, such appointments may function similarly to what Epstein and O’Halloran (1995, 237) refer to as “confirmatory signals” whereby the bipartisan nature of an appointee makes information coming from the executive via that appointee more credible. This should help presidents to build support when facing greater legislative opposition.

Hypothesis 4: The level of partisan support a president enjoys in Congress is negatively correlated with likelihood of observing a bipartisan appointment.

Greater support in Congress should make it easier for the president to form a coalition to support his legislative initiatives. However, Poole and Rosenthal (1997) and McCarty, Poole, and Rosenthal (2006) have provided evidence that Congress has become increasingly polarized since the 1970s. Research has also found that the nominations of presidential appointees are more likely to be delayed as polarization increases (Bond, Fleisher, and Krutz 2009; McCarty and Razaghian 1999). Kupchan and Trubowitz (2007) also provide some evidence indicating that increasing polarization has been accompanied by a decrease in the number of moderates in Congress.⁸ If bipartisan appointments are intended to enhance the president’s ability to pursue his legislative agenda through the construction of moderate coalitions, then we should expect that a larger population of moderates will facilitate this process. Thus, there is less reason to expect a president will make a bipartisan appointment when polarization is high as the probability of such appointments helping to facilitate moderate coalitions is diminished.

Hypothesis 5: Congressional polarization should be negatively correlated with the likelihood of observing a bipartisan appointment.

We might also expect the effect of partisan support in Congress to be conditional upon the level of polarization. We should expect presidents to respond to lower levels of support in Congress by making bipartisan appointments in an effort to build coalitions of moderate legislators. However, the ability of the president to rally support from moderate opposition party members should be easier when polarization is low and the opposition party’s median is closer to the president’s own. As polarization increases, we should expect it to be harder to find moderates with whom coalitions can be formed, because the proportion of moderates declines with increasing polarization. Under such conditions, the expected utility of making a bipartisan appointment to address diminished Congressional support should be lower as compared with periods of lower polarization, when such appointments may be more effective.

Hypothesis 6: The effect of a decline in Congressional support on the likelihood of observing a bipartisan appointment should be larger under conditions of low polarization.

Data and Analysis

Dependent Variable

As most previous studies of bipartisanship have focused on Congressional voting patterns, a new method of evaluating bipartisanship in presidential appointments to the foreign policy bureaucracy is required to test the hypotheses listed above. Accordingly, I collected new data on appointments to the foreign policy bureaucracy. The data used herein consist of 2,430 office-year observations, running from 1948 to 2011. The agencies included are the Department of State, Department of Defense, Department of the Treasury, the Central Intelligence Agency (CIA), the National Security Advisor (NSA), and the Council of Economic Advisors (CEA). Because the Departments of State and Defense are the two primary agencies in the foreign policymaking process, data were collected on individuals holding positions at the Secretary, Deputy Secretary, Under Secretary, and Assistant Secretary levels. Given their historical importance, the Director of Policy Planning and the Ambassador to the United Nations were also included. For the Treasury, NSA, and CIA, I coded individuals at the Principal and Deputy levels. Finally, all three members of the CEA were also coded.⁹ This time frame was chosen for the relative stability of the foreign policy bureaucracy during this time period. The National Security Act of 1947 created the Department of Defense and the Central Intelligence Agency. Furthermore, many positions at the Under Secretary and Assistant Secretary levels did not exist until after this time. Thus, any observations before 1948 would be based on a bureaucratic landscape that was quite different from that found after 1948.¹⁰

For each individual position that was observed, I coded what, if any, party affiliation the office’s occupant had. The dependent variable in the following analysis is a dichotomous variable that indicates whether the office is occupied by a bipartisan appointee during a given year(1 = Yes, 0 = No).¹¹ An appointee is coded as bipartisan if they are a member of the opposition party, for example, a Democrat serving in a Republican administration. Any other officials are coded as “0.” As the dependent variable is dichotomous, I estimate a series of probit models to test the hypotheses listed above.¹²

Independent Variables

I include several key independent variables to test the hypotheses outlined above. First, I include a dummy variable to identify those years when the United States is at war to evaluate Hypothesis 1. The foreign policy literature expects politicians to put narrow partisan interests aside during times when national security issues are on the line. War involvement provides a useful measure as it indicates a significant, and potentially protracted, period of armed conflict. These larger conflicts are the kinds of events we should expect to produce greater unity, and so we should expect this variable to yield a positive coefficient. This variable is generated using the Correlates of War Project’s Interstate War Data Set (Sarkees and Wayman 2010).¹³

To evaluate Hypothesis 2, I control for the impact of the watershed events described above by using two separate dummy variables for the post-Vietnam and the post-Cold War periods. These variables are coded “0” for each year prior to an event and “1” for each year after.¹⁴ We should expect a negative coefficient for both of these variables.

Hypothesis 3 suggests that we should expect differences between Republicans and Democrats in the probability of observing bipartisan appointments. I use a dummy variable to identify whether or not the president is a Democrat (1) or Republican (0). Theory also suggests that, while the watershed events listed above should be expected to dampen bipartisanship generally, partisan differences may exist across the entire time period under consideration. I interact the president’s party variable with each of the watershed event dummies. This will allow me to compare the conditional means of both parties against one another and across time periods, allowing for a fuller examination of the impact that these events had on the likelihood of bipartisan appointments.

Hypotheses 4 to 6 concern the influence that domestic institutional conditions ought to have on presidential decisions to make bipartisan appointments. As per Hypothesis 4, the level of partisan support a president enjoys in Congress should be negatively correlated with bipartisan appointments. Following Kriner (2010), I control for support by calculating the mean percentage of both houses of Congress that are controlled by the president’s party.¹⁵ Hypothesis 5 suggests a negative relationship between Congressional polarization and bipartisan appointments. To capture the level of Congressional polarization, I use the mean polarization score for the House and Senate (McCarty, Poole and Rosenthal 2006; Poole 2012; Poole and Rosenthal 1997). I also include an interaction term between the Congressional support variable and the polarization variable to evaluate the hypothesized conditional relationship between the two, as per Hypothesis 6.

I also control for other variables that may have an impact on the presence of bipartisanship. Previous work suggests that election timing can influence the partisan makeup of presidential administrations (Best 1981; Brown 1982). The year following an election might see a lower rate of bipartisan appointments as presidents fill vacancies with political loyalists. Krutz, Fleisher, and Bond (1998) also find that presidential nominations are more likely to succeed earlier in a president’s term. It is likely that presidents will use these opportunities to make appointments that are closer to their own ideal points. To control for these dynamics, I include a dummy variable for election years, lagged one time period because such appointments will not occur until after a president is inaugurated. Trubowitz and Mellow (2005) find that poor economic conditions are linked to lower levels of bipartisanship in Congressional voting. As economic conditions worsen, redistributive pressures intensify and the stakes of any given deal may be higher as compared with periods of economic prosperity. I include a measure of the average annual unemployment rate to control for this possibility (Bureau of Labor Statistics 2011). Presidential approval may also impact the likelihood of bipartisan appointments. Meernik (1993) finds that presidents enjoy greater bipartisan support in Congress when presidential popularity is high. High popularity is also associated with higher probabilities that the president’s nominations will be approved by the Senate (Krutz, Fleisher, and Bond 1998). Accordingly, bipartisan appointments may be less useful when approval is high, and so I control for the mean annual level of presidential job approval (Woolley and Peters 2011). Finally, the bureaucracy has grown over time, thereby affecting the relative opportunity for presidents to make bipartisan appointments. I control for the total number of positions in a given year to account for changes in opportunity.

Results

Table 1 shows the results for the first set of probit models. Model 1 includes only the variables for Vietnam, and Model 2 introduces the variables for the Cold War. First, we find—somewhat surprisingly—that the war variable is not statistically significant in either Model 1 or 2. According to Hypothesis 1, we should expect this coefficient to be statistically significant and positive. The foreign policy literature has long suggested that periods of war prompt citizens and policymakers to “rally round the flag” as a display of national unity. However, this is a fairly blunt test of how conflict affects the propensity toward bipartisanship, and I examine these dynamics further below.

Table 1.

Bipartisan Appointees—Base Model.

Hypothesis	Variable	(1)	(2)
H1	War	−0.0282 (0.0735)	−0.0388 (0.0753)
H2	Post-Vietnam	0.769*** (0.165)	0.754*** (0.142)
H2	Post-Cold War		0.0295 (0.0924)
H3	Democratic president	1.154*** (0.191)	1.300*** (0.195)
H2 and H3	Democratic president × Post-Vietnam	−1.020*** (0.130)	−0.921*** (0.0979)
H2 and H3	Democratic president × Post-Cold War		−0.282 (0.173)
H4	President’s support in Congress	−0.0815*** (0.0254)	−0.104*** (0.0277)
H5	Polarization	−4.507* (2.401)	−5.699*** (2.179)
H6	Support × Polarization	0.0971** (0.0382)	0.124*** (0.0422)
Controls	Election	−0.133** (0.0675)	−0.116* (0.0670)
	Unemployment rate	−0.00884 (0.0253)	−0.00976 (0.0247)
	Approval rate	0.00691** (0.00320)	0.00760** (0.00305)
	Total appointees	−0.0283** (0.0126)	−0.0283** (0.0125)
	Constant	2.599** (1.299)	3.533*** (1.299)
	Log likelihood	−658.4	−657.9
	Observations	2,430	2,430

Robust standard errors clustered by administration in parentheses.

Two-tailed significance tests used: *p < .1. **p < .05. ***p < .01.

Hypothesis 2 suggests that watershed events should be associated with a decline in the likelihood that presidents make bipartisan appointments. Model 1 includes the post-Vietnam dummy variable, while Model 2 adds the post-Cold War dummy variable. Both models include the interaction terms between the post-event dummies and the Democratic president dummy variable. Hypothesis 3 suggests that we should expect Democratic presidents to be more inclined to make bipartisan appointments than Republicans. These interactions help us evaluate whether or not these events have led to a structural decline in the likelihood of presidents making a bipartisan appointment, but also the extent to which there are differences between the two parties, and whether or not presidents of each party have responded differently to these events. In Model 1, the post-Vietnam variable is positive and significant at the .01 level. This finding holds up in Model 2, where I introduce the post-Cold War variable, which is itself not statistically significant. Alternatively, the Democratic president variable is positive and significant at the .01 level in both models. As these variables are dummy variables, we can begin to draw some basic inferences. First, there is little support for Hypothesis 2, as neither the post-Vietnam nor the post-Cold War variables appear to be associated with a decrease in the probability of presidents making bipartisan appointments. The positive and significant coefficient on the Democratic president variable does provide some evidence that Democratic presidents have been more likely to make bipartisan appointments than Republican presidents. However, to fully evaluate Hypotheses 2 and 3, we must consider the post-Vietnam and post-Cold War variables along with their respective interaction terms. The positive and statistically significant coefficient on the Democratic president variable indicates that Democratic presidents had a statistically significantly higher probability of making bipartisan appointments than Republicans during the early Cold War period. However, the significance on the post-Vietnam variable and its interaction term with the Democratic president variable suggest that the differences between the parties have changed. Democratic presidents have seen a decline in the probability that they make a bipartisan appointment in the post-Vietnam period. The coefficient on the post-Cold War interaction term does not reach statistical significance. I discuss these issues in greater detail below.

Hypotheses 4 to 6 concerning the level of support a president enjoys in Congress, the negative effect of polarization, and the conditioning effect of polarization on the support variable enjoy fairly robust support across Models 1 and 2. The presidential support variable is negative and statistically significant at the .01 level in both models. This finding suggests that as presidential support in Congress increases, the probability of seeing a bipartisan appointment drops. This finding is in line with Hypothesis 4. Similarly, the results indicate polarization is negatively correlated with the probability of a president making a bipartisan appointment—as the gap between the two parties increases, and the population of moderates in Congress declines, the utility of such appointments declines. This finding supports Hypothesis 5. However, the interaction term suggests that the magnitudes of these effects are not static, as I discuss more below.

The results shown in Table 1 provide a first cut at the issues addressed here and require further examination. The lack of a clear relationship between the war variable and bipartisanship is at odds with conventional wisdom. However, the variable used in these models is a rather blunt instrument—the dummy variable is coded “1” for every year during which the United States is at war. This coding scheme only tells us if periods of war are different from periods where the United States is not at war and may be masking substantial variation within conflicts. Particularly long conflicts—Vietnam and the wars in Iraq and Afghanistan, for example—became especially contentious issues, with members of the opposition party seeking to distance themselves from the administration’s policies, and often becoming vocal critics. Alternatively, presidents and their administrations have often responded by closing ranks.¹⁶ Previous quantitative studies by Meernik (1993) and Trubowitz and Mellow (2005) have also suggested that prolonged conflicts can have a negative effect on bipartisanship in Congressional voting. Accordingly, as the duration of a conflict increases, we might expect presidents to be less likely to make bipartisan appointments. This also suggests that any rally effect should be most likely to occur early in a conflict.¹⁷

The models in Table 2 introduce two new variables to tease out this potential intra-conflict variation. The “War Onset” variable is coded “1” for the first year of a war, and the “War Duration” variable counts the number of years since a war began.¹⁸ This new specification allows us to see if wars are immediately met with an increase in the predicted probability that presidents will make a bipartisan appointment, and whether or not the probability of observing a bipartisan appointment changes over the course of a conflict.

Table 2.

Bipartisan Appointee—War Onset and Duration.

Hypothesis	Variable	(1)	(2)
H1	War onset	0.0785 (0.0530)	0.0765 (0.0548)
H1	War duration	−0.0668*** (0.0195)	−0.0671*** (0.0181)
H2	Post-Vietnam	0.977*** (0.114)	0.954*** (0.0995)
H2	Post-Cold War		−0.0411 (0.0667)
H3	Democratic president	1.178*** (0.150)	1.317*** (0.190)
H2 and H3	Democratic president ×Post-Vietnam	−1.108*** (0.108)	−1.015*** (0.0643)
H2 and H3	Democratic president × Post-Cold War		−0.246 (0.171)
H4	President’s support in Congress	−0.131*** (0.0203)	−0.150*** (0.0286)
H5	Polarization	−8.736*** (1.806)	−9.492*** (1.948)
H6	Support × Polarization	0.191*** (0.0322)	0.212*** (0.0399)
Controls	Election	−0.195** (0.0794)	−0.174** (0.0803)
	Unemployment rate	−0.0226 (0.0224)	−0.0250 (0.0203)
	Approval rate	0.00593*** (0.00190)	0.00677*** (0.00194)
	Total appointees	−0.0401*** (0.0110)	−0.0391*** (0.0106)
	Constant	5.365*** (0.951)	6.036*** (1.244)
	Log likelihood	−654.3	−653.9
	Observations	2,430	2,430

Robust standard errors clustered by administration in parentheses.

Two-tailed significance tests used: *p < .1. **p < .05. ***p < .01.

The results from Table 2 are informative. War onset fails to reach statistical significance in either model. Accordingly, there is no indication that the onset of a war is met with a higher probability of presidents making a bipartisan appointment. This poses some further challenges to Hypothesis 1 and the conventional wisdom concerning the rallying effect of conflict. Alternatively, War Duration is statistically significant and negative in both Models 1 and 2. This negative coefficient indicates that as the duration of conflicts increases, the predicted probability of a president making a bipartisan appointment declines. Figure 1 illustrates this dynamic across the range of the war duration variable observed in the estimation sample. There is a baseline predicted probability of .10 that a president makes a bipartisan appointment, decreasing to a low of approximately .025 at the highest observed value of the war duration variable. The latter predicted value is statistically significantly lower than the predicted value at the low end of the X-axis. Such long conflicts are unusual, however. The dashed line represents the mean for the war duration variable—only 2.9 years. So while the predictions at either end of the X-axis in Figure 2 are statistically significantly different from one another, the predicted values at the low end of the X-axis and those at the mean value are not statistically significantly different. Accordingly, these results indicate that while we do not see an increase in bipartisanship as a result of a war’s onset, we can expect to see a decline in bipartisanship if a war drags on long enough. This is consistent with the idea that partisans close ranks as war duration increases and it becomes the subject of political wrangling, and supports the findings of previous studies finding a negative relationship between international crises and bipartisanship (Meernik 1993; Trubowitz and Mellow 2005).

Figure 1.

Predicted probability of a bipartisan appointment as a function of war duration.

Figure 2.

Predicted probability of observing a bipartisan appointment by the president’s party and time period.

The results shown in Table 2 for the variables evaluating Hypotheses 2 and 3 are generally consistent with the results from Table 1. The post-Vietnam variable, the Democratic president variable, and their interaction all remain statistically significant and in the expected directions. However, the interaction between the Democratic president variable and the post-Cold War variable is still not statistically significant in Model 2. As I discussed above, these results require some care in interpreting. Figure 2 shows the predicted probabilities of a bipartisan appointment by the president’s party in the pre- and post-Vietnam period based on Model 1 from Table 2. Figure 2 indicates that in the pre-Vietnam period, the predicted probability of a bipartisan appointment was approximately .12 for Democratic presidents and .01 for Republican presidents. While the difference between the two parties in the pre-Vietnam is statistically significant, the post-Vietnam period actually sees an increase in the probability of a bipartisan appointment for Republican presidents. In fact, post-Vietnam Republicans are statistically indistinguishable from post-Vietnam Democrats. The results also indicate that the predicted probability of a bipartisan appointment under Democratic presidents has remained relatively stable over time.

These results are striking. First, they indicate partial support for Hypothesis 3 and the idea that Democratic presidents have been more prone to make bipartisan appointments than Republican presidents but suggest that this difference was confined to the early Cold War period. Second, there is no indication that the post-Vietnam period was marked by a general decline in the probability of observing a bipartisan appointment, as several scholars have claimed (for one prominent example, see Destler, Gelb, and Lake 1985). However, the findings here suggest that there was actually a convergence over time, with Republican presidents becoming more likely to make bipartisan appointments, bringing them in line with Democratic administrations. There are a few possible explanations for these discrepancies. First, the probability of Republican presidents making bipartisan appointments during the early Cold War period may have been suppressed by an unusual combination of factors. The Eisenhower administration was the first Republican administration in two decades, and there was likely substantial pressure to give bureaucratic positions to fellow Republicans as a result. In addition, many targets of the McCarthy hearings were bureaucratic appointees and the Eisenhower administration was under pressure to avoid appearing soft on communism, perhaps causing them to shy away from appointing more liberal Democrats. Republican assertions that the Democrats under Truman “lost” China to the communists also made bipartisan appointments difficult to sell (see Acheson 1969; Gaddis 2011; Isaacson and Thomas 1986). Alternatively, the increased probability of Republican presidents making bipartisan appointments could also reflect the idea that waning support among Democrats for the more militant varieties of liberal internationalist policies prompted Republican presidents to become more bipartisan in an effort to secure support for their preferred policies.

The models in Table 2 also yield robust support for Hypotheses 4 through 6. The president’s level of Congressional support, polarization, and their interaction all perform as in Table 1. Again we see that the support variable is negative and statistically significant at the .01 level in Models 1 and 2, indicating that an increase in the president’s level of support in Congress is negatively correlated with the probability of making a bipartisan appointment. This is in line with Hypothesis 4. Providing further support for Hypothesis 5, polarization also remains negative and significant at the .01 level in both models, indicating that increasing polarization is correlated with a lower probability of observing a bipartisan appointment. The interaction term continues to be significant at the .01 level across both models.

Figure 3 plots the marginal effect of a change in the support variable across the range of polarization observed in the estimation sample. I plot the marginal effect of a four percentage point decrease (from 52% to 48%) in the level of Congressional support the president enjoys. This range was chosen as it represents a more substantively meaningful change—one where the president’s party goes from controlling, on average, a majority of the seats in Congress, to one where the president’s party controls a minority of seats. As we expect bipartisan appointments are used to compensate for lower levels of Congressional support by helping to build moderate coalitions, it makes intuitive sense to look at this in terms of how presidents respond to a decline in Congressional support. At the lowest levels of polarization, this four percentage point decrease in support is correlated with an increase of nearly three percentage points in the predicted probability of a bipartisan appointment. However, as polarization increases, the magnitude of this effect decreases in size, eventually becoming negative. When polarization is equal to .60, there is still an increase in the predicted probability of a bipartisan appointment, but the magnitude of this effect has shrunk to about one percentage point. The marginal effect of this decrease becomes statistically insignificant when polarization reaches an approximate value of .64 on the X-axis, as indicated by the shaded box in the graph. The effect becomes statistically significant again when polarization exceeds .76. However, when polarization is high, this decline in Congressional support leads to a decrease in the predicted probability of a bipartisan appointment. The decreasing magnitude of this effect makes sense—when polarization is low, moderates with whom bipartisan appointees can work are plentiful, but when polarization is high, moderates are scarcer. High polarization environments may also be lacking in the more moderate types of potential appointees that presidents could draw on to fill posts within the foreign policy bureaucracy. Figure 3 yields support for Hypothesis 4, but with one caveat: The hypothesized relationship between Congressional support and presidential decisions to make bipartisan appointments holds, but only when polarization is low and the supply of moderates is high.

Figure 3.

Marginal effect of a four percentage point change in the level of Congressional support (from 52% to 48%) across the range of polarization.

Two additional points should be made. First, the marginal effect shown in Figure 3 experiences a statistically significant change in magnitude within the positive range (i.e., when polarization ≤ .64). This point matters in the context of the second point: Changes in polarization of the magnitude that are represented in Figure 3 unfold over the course of several years. Accordingly, these results indicate that while we do see a statistically significant change in the magnitude of the effect of a decrease in Congressional support, this change will not necessarily be clear from one Congress to the next. It is also difficult to draw further inferences regarding the extent to which we should expect to see change in the magnitude of the negative effect (i.e., when polarization ≥ .76). The large confidence intervals at the upper end of the X-axis are partially driven by the fact that there are fewer observations at these higher values of polarization. So while we can say that there is a statistically significant negative effect for the support variable in this higher range of polarization, we cannot conclude that this effect has necessarily been getting “more” negative.

Here I will briefly review the results of the control variables. The election variable is negative and significant in all of the models, suggesting that presidents are less likely to make bipartisan appointments immediately following elections. The variable for total appointees is negative and statistically significant in all models. This would indicate that presidents have not continued to make bipartisan appointments at the same rate as they have other appointments, suggesting that relative bipartisanship has declined over time. The unemployment rate variable and the approval rate variable both fail to match expectations in all four models. The unemployment rate variable is not statistically significant in any model. Accordingly, we cannot draw any inferences as to it having a consistent effect one way or the other. And while presidential approval is statistically significant in all models, it is not in the expected direction.

Conclusion

This project has advanced our understanding of the international and domestic sources of bipartisanship in foreign policymaking on several fronts. First, by drawing on the observable implications of conventional wisdoms that have existed for decades in the foreign policy literature, but have remained largely untested, this study has generated new data and testable hypotheses on bipartisan appointments to the foreign policy bureaucracy. Subjecting these claims to systematic empirical scrutiny, however, points to some deficiencies in the stylized truths that the qualitative foreign policy literature has typically embraced. Alternatively, by drawing on arguments from the American politics literature, this study has been able to expand upon our knowledge of the systematic relationships between domestic political forces and bipartisanship in foreign policymaking.

First, there is no evidence that war leads to an increase in the probability of a bipartisan appointment to the foreign policy bureaucracy. In fact, what inferences we can draw from the preceding analysis suggest that the probability of bipartisan appointments actually declines as wars drag on. To put it differently, while conflict does not appear to help bipartisanship, it can certainly harm it. Robustness checks using alternative measures of U.S. conflict involvement yield either no correlation, or a statistically significant negative correlation with bipartisanship (see the supplemental appendix at http://prq.sagepub.com/supplemental/).¹⁹ These findings are less surprising than they may first seem. As I discuss above, there is evidence indicating that conflicts become increasingly politicized as they drag on, prompting presidents to close ranks and creating opportunities for the opposition party to attack the administration’s handling of the conflict (Isaacson and Thomas 1986; Meernik 1993; Trubowitz and Mellow 2005). In reality, this suggests that presidents will be less willing to make bipartisan appointments, and opposition party members are less inclined to accept such invitations, should they be extended. Furthermore, appointing individuals to the bureaucracy can be a timely affair and is perhaps not the fastest or most efficient means of demonstrating national unity. Those cases that have been held up as examples of these dynamics—the appointment of Henry Stimson, for example—are especially high-profile cases and are not necessarily representative of any systematic processes. Stimson’s appointment is actually somewhat anomalous when we consider that American intervention into the emerging European conflict was an issue that evolved over the years preceding America’s actual entry into World War II. Stimson was Secretary of War for over a year before the attack on Pearl Harbor, and in that time, he helped build support for increasing American involvement. Wars often do not provide presidents with this kind of lead time, and this may have made the use of bipartisan appointments a more viable option for building national unity in this particular case.

Similarly, the results fail to provide support for the oft-cited collapse of bipartisanship following watershed events such as Vietnam and/or the end of the Cold War. This poses a challenge to conventional wisdom, as these findings indicate that bipartisanship in Republican and Democratic administrations actually converged in the post-Vietnam period. There is also no evidence that the post-Cold War period saw a decline in bipartisanship. This could be the result of a few factors, such as the long absence of Republicans from the White House and the politics surrounding McCarthyism during the early Cold War period, as well as the growing need for Republican presidents to generate Democratic support for international activism following Vietnam. However, these issues also suggest an important, yet often overlooked, point: While the existence of a bipartisan consensus is certainly an issue of “if,” it is also an issue of “when” and “where.” Measurement issues, as well as decisions concerning the appropriate unit of analysis and scope of a study, can influence researchers’ findings. Recent work by Kupchan and Trubowitz (2007) and Chaudoin, Milner, and Tingley (2010) illustrates that this subject remains much in dispute, at least in part because of such issues. For example, differences in these studies’ findings concerning changes in public opinion could be attributable to the use of different survey questions. In this sense, this study also represents a different approach as compared with previous work, conducting a systematic analysis of bipartisanship in an area where previous quantitative studies have not explored. Most existing quantitative work has been largely limited in its focus to Congressional voting patterns and public opinion (e.g., Chaudoin, Milner, and Tingley 2010; Holsti and Rosenau 1984; Kupchan and Trubowitz 2007; McCormick and Wittkopf 1990; Meernik 1993; Wittkopf 1990). These discrepancies should lead us to be cautious in drawing overly broad inferences. Bipartisanship may be a more multifaceted concept than previously recognized. Viewed more narrowly, the results of this study do contradict some important claims made by the foreign policy literature, but this should not be taken as conclusive. Only through further research and considering studies collectively can we draw conclusions as to the general state of bipartisanship.

This study has also produced insights into how the domestic political environment affects bipartisanship in presidential appointments to the foreign policy bureaucracy. The results generally support the hypotheses concerning the influence of the level of partisan support a president enjoys in Congress. When polarization is low and moderates in Congress are more plentiful, presidents may be more likely to respond to decreases in their own party’s base of support in Congress by making bipartisan appointments. However, as polarization increases the probability of presidents responding in this way declines. This suggests that increasing polarization has effectively closed one avenue for greater cooperation between the executive and Congress. Given that polarization changes over long periods of time, it is unlikely that this trend will be reversed anytime soon. More broadly, this analysis builds upon previous work by American politics scholars on how Congress affects the appointments process (e.g., Krutz, Fleisher, and Bond 1998; McCarty and Razaghian 1999), as well as work by international relations scholars showing how Congress affects the executive in the foreign policymaking (e.g., Clark 2000; Howell and Pevehouse 2007; Kriner 2010).

There are several possibilities for future research. The results are consistent with the theoretical notion that presidents use bipartisan appointments to build moderate legislative coalitions to advance their policy goals, but they do not speak to the actual efficacy of such a strategy. The fact that presidents are more likely to make bipartisan appointments in the face of lower Congressional support does not necessarily mean that executive–legislative relations will improve, or that the president will enjoy greater legislative success. Future studies could use these data to examine the actual effect that these high-ranking appointees have on executive–legislative relations. Patterns in legislative gridlock, for example, may yield insights into the effect on legislative activity. Similarly, if bipartisan appointees are working with Congressional leadership to moderate legislation, we might expect them to be associated with higher levels of bipartisanship in Congressional voting patterns. Additional work in these areas can improve our understanding of how bipartisanship actually affects the policymaking process, or whether it is simply a symbolic act with little substantive consequence. The idea of a bipartisan consensus as a more multifaceted phenomenon also opens the door to other interesting questions. For example, what does it suggest if bipartisanship in Congressional voting has held constant but has declined in public opinion polls? Some recent research has begun to address these sorts of issues (e.g., Busby and Monten 2012) and future research can seek to further explore how bipartisanship varies across different issue areas and elite–mass divisions. Future projects can use these data to address these questions by looking at the kinds of positions bipartisan appointees are given. For example, do we see bipartisan appointees given influential positions, or are they removed from key policymaking posts? Finally, this study contributes to mounting evidence that international crises and conflict do not have the positive impact on bipartisanship that the foreign policy literature has typically claimed. Future research should seek to address this issue more fully to better understand when and how military conflict affects bipartisan cooperation.

Footnotes

Acknowledgements

The author thanks Michael Allen, Colin Barry, Dave Clark, Chad Clay, Ben Fordham, William Howell, Katja Kleinberg, Michael McDonald, Greg Robinson, Olga Shvetsova, and Julie VanDusky–Allen for their comments and suggestions.

Author’s Note

Previous versions of this article were presented at the 2011 Annual meeting of the American Political Science Association in Seattle, Washington, and at the Binghamton University World Politics Workshop. All remaining errors are my own.

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) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Acheson

Dean

. 1969. Present at the Creation: My Years in the State Department. New York: W. W. Norton.

Berger

Henry W.

1975. “Bipartisanship, Senator Taft, and the Truman Administration.” Political Science Quarterly 90 (2): 221–37.

Best

James J.

1981. “Presidential Cabinet Appointments: 1953–1976.” Presidential Studies Quarterly 11 (1): 62–66.

Binder

Sarah

. 1999. “The Dynamics of Legislative Gridlock, 1947-1996.” American Political Science Review 93 (3): 519–33.

Bird

Kai

. 2000. The Color of Truth: McGeorge Bundy and William Bundy: Brothers in Arms. New York: Touchstone.

Bond

Jon R.

Fleisher

Richard

Krutz

Glen S.

2009. “Malign Neglect: Evidence that Delay Has Become the Primary Method of Defeating Presidential Appointments.” Congress & the Presidency 36:226–43.

Bonds

John B.

2002. Bipartisan Strategy: Selling the Marshall Plan. Westport: Praeger.

Briggs

Philip J.

1994. Making American Foreign Policy: President-Congress Relations from the Second World War to the Post-Cold War Era. New York: Rowman & Littlefield.

Brown

Roger G.

1982. “Party and Bureaucracy: From Kennedy to Reagan.” Political Science Quarterly 97 (2): 279–94.

10.

Bureau of Labor Statistics. 2011. “Unemployment Rate. Technical Report.” Washington, DC: United States Government. http://data.bls.gov.

11.

Busby

Joshua W.

Monten

Jonathan

. 2008. “Without Heirs? Assessing the Decline of Establishment Internationalism in U.S. Foreign Policy.” Perspectives on Politics 6:451–72.

12.

Busby

Joshua W.

Monten

Jonathan

. 2012. “Republican Elites and Foreign Policy Attitudes.” Political Science Quarterly 127 (1): 105–42.

13.

Carney

Dana R.

Jost

John T.

Gosling

Samuel D.

Potter

Jeff

. 2008. “The Secret Lives of Liberal and Conservatives: Personality Profiles, Interaction Styles, and the Things They Leave Behind.” Political Psychology 29 (6): 807–40.

14.

Chaudoin

Stephen

Milner

Helen V.

Tingley

Dustin

. 2010. “The Center Still Holds: Liberal Internationalism Survives.” International Security 35 (1): 75–94.

15.

Clark

David H.

2000. “Agreeing to Disagree: Domestic Institutional Congruence and U.S. Dispute Behavior.” Political Research Quarterly 53 (2): 375–400.

16.

Destler

I. M.

Gelb

Leslie

Lake

Anthony

. 1985. Our Own Worst Enemy: The Unmaking of American Foreign Policy. New York: Simon and Schuster.

17.

Epstein

David

O’Halloran

Sharyn

. 1995. “A Theory of Strategic Oversight: Congress, Lobbyists, and the Bureaucracy.” The Journal of Law, Economics, & Organization 11 (2): 227–55.

18.

Fleisher

Richard

Bond

Jon R.

1988. “Are There Two Presidencies? Yes, But Only for Republicans.” Journal of Politics 50 (1): 747–67.

19.

Fordham

Benjamin O.

1998. “Partisanship, Macroeconomic Policy, and U.S. Uses of Force, 1949-1994.” The Journal of Conflict Resolution 42 (4): 418–39.

20.

Gaddis

John L.

2011. George Kennan: An American Life. New York: The Penguin Press.

21.

Ghosn

Faten

Palmer

Glenn

Bremer.

Stuary

2004. “The MID3 Data Set, 1993–2001: Procedures, Coding Rules, and Description.” Conflict Management and Peace Science 21:133–54.

22.

Halberstam

David

. (1972) 1992. The Best and the Brightest. Ballantine Books. New York, NY.

23.

Halperin

Morton

Clapp

Priscilla

Kanter

Arnold

. 2007. Bureaucratic Politics and Foreign Policy. 2nd ed. Washington, DC: Brookings Institution Press.

24.

Herring

George C.

2008. From Colony to Superpower: U.S. Foreign Relations since 1776. Kindle ed. New York: Oxford University Press.

25.

Hilsman

Roger

. 1987. The Politics of Policy Making in Defense and Foreign Affairs: Conceptual Models and Bureaucratic Politics. 3rd ed. Englewood Cliffs: Prentice Hall.

26.

Hodgson

Godfrey

. 1973. “The Establishment.” Foreign Policy 10 (Spring): 3–40.

27.

Hodgson

Godfrey

. (1990) 1992. The Colonel: The Life and Wars of Henry Stimson. Random House. New York, NY.

28.

Holsti

Ole R.

Rosenau

James

. 1984. American Leadership in World Affairs: Vietnam and the Breakdown of Consensus. Unwin Hyman. Winchester, MA.

29.

Howell

William G.

Pevehouse

Jon

. 2007. While Dangers Gather: Congressional Checks on Presidential War Powers. Princeton: Princeton University Press.

30.

Ikenberry

G. John

. 2002. “America’s Imperial Ambition.” Foreign Affairs 81 (5): 44–60.

31.

Ikenberry

G. John

. 2011. Liberal Leviathan: The Origins, Crisis, and Transformation of the American World Order. Princeton: Princeton University Press.

32.

Isaacson

Walter

Thomas

Evan

. 1986. The Wise Men: Six Friends and the World They Made. New York: Simon and Schuster.

33.

Kriner

Douglas L.

2010. After the Rubicon: Congress, Presidents, and the Politics of Waging War. Chicago: The University of Chicago Press.

34.

Krutz

Glen S.

Fleisher

Richard

Bond

Jon R.

1998. “From Able Fortas to Zöe Baird: Why Some Presidential Nominations Fail in the Senate.” American Political Science Review 92 (4): 871–81.

35.

Kupchan

Charles A.

2003. The End of the American Era: U.S. Foreign Policy and the Geopolitics of the Twenty-First Century. New York: Vintage Books.

36.

Kupchan

Charles A.

Trubowitz

Peter L.

2007. “Dead Center: The Demise of Liberal Internationalism in the United States.” International Security 32 (2): 7–44.

37.

Mackenzie

G. Calvin

Light

Paul

. 1985. “Presidential Appointees, 1964–1984.” http://doi.org/10.3886/ICPSR08458.v1.

38.

Mann

James

. 2004. Rise of the Vulcans: The History of Bush’s War Cabinet. New York: Viking.

39.

Marquis Who’s Who. 2011. “Marquis Who’s Who Online Database.” http://www.marquiswhoswho.com/.

40.

McCarty

Nolan

Poole

Keith T.

Rosenthal

Howard

. 2006. Polarized America: The Dance of Ideology and Unequal Riches. Cambridge: The MIT Press.

41.

McCarty

Nolan

Razaghian

Rose

. 1999. “Advice and Consent: Senate Responses to Executive Branch Nominees 1885–1996.” American Journal of Political Science 43 (4): 1122–43.

42.

McCormick

James M.

Wittkopf

Eugene R.

1990. “Bipartisanship, Partisanship, and Ideology in Congressional-Executive Foreign Policy Relations, 1947–1988.” The Journal of Politics 52 (4): 1077–100.

43.

Meernik

James

. 1993. “Presidential Support in Congress: Conflict and Consensus on Foreign and Defense Policy.” The Journal of Politics 55 (3): 569–87.

44.

Mueller

John E.

1970. “Presidential Popularity from Truman to Johnson.” American Political Science Review 64 (1): 18–34.

45.

Neustadt

Richard E.

1970. Alliance Politics. New York: Columbia University Press.

46.

Neustadt

Richard E.

(1960) 1991. Presidential Power and the Modern Presidents: The Politics of Leadership from Roosevelt to Reagan. Revised ed. New York: Free Press.

47.

Poole

Keith T.

2012. “Voteview.” http://voteview.com/house55.htm.

48.

Poole

Keith T.

Rosenthal

Howard

. 1997. Ideology and Congress. New Brunswick: Transaction Publishers.

49.

Roberts

Priscilla

. 1992. “‘All the Right People’: The Historiography of the American Foreign Policy Establishment.” Journal of American Studies 26 (3): 409–34.

50.

Sarkees

Meredeith R.

Wayman

Frank

. 2010. Resort to War: 1816–2007. Washington, DC: CQ Press.

51.

Trubowitz

Peter

. 1998. Defining the National Interest. Chicago: The University of Chicago Press.

52.

Trubowitz

Peter

Mellow

Nicole

. 2005. “‘Going Bipartisan’: Politics by Other Means.” Political Science Quarterly 120 (3): 433–53.

53.

Wittkopf

Eugene R.

1990. Faces of Internationalism: Public Opinion and American Foreign Policy. Durham: Duke University Press.

54.

Woodward

Bob

. 2002. Bush at War. New York: Simon and Schuster.

55.

Woolley

John

Peters

Gerhard

. 2011. “The American Presidency Project.” http://www.presidency.ucsb.edu/data/popularity.