Roundtable 2-4 on “Is Liberal Internationalism in Decline?”75 min read

This roundtable extends the debate on the future of liberal internationalism (LI)  started by Charles Kupchan and Peter Trubowitz (K&T) three years ago which provoked responses by Steven Chaudoin, Helen Milner, and Dustin Tingley (CMT), and by Joshua Busby and Jonathan Monten (B&M).[1]  This was the subject of a panel at a recent meeting of the American Political Science Association, and the memos written by the three teams form the basis for this Roundtable.  Despite – or because – of my lack of involvement in these issues I was asked to chair the panel and so here will make a few introductory remarks, and after those involved have had their say will outline a few ideas for further research.


H-Diplo | ISSF Roundtable, Volume II, No. 4 (2011)
Diane Labrosse and Thomas Maddux, H-Diplo/ISSF Editors
George Fujii, H-Diplo/ISSF Web and Production Editor
Commissioned for H-Diplo/ISSF by Christopher Ball
H-Diplo/ISSF Roundtable Review on “Is Liberal Internationalism in Decline?”
Published by H-Diplo/ISSF on 27 February 2011
Introduction by Robert Jervis, Columbia University. 2
Essay by Josh Busby, University of Texas-Austin and Jonathan Monten, London School of Economics and Political Science. 7
Essay by Stephen Chaudoin, Princeton University, Helen V. Milner, Princeton University, and Dustin Tingley Harvard University. 16
Essay by Brian C. Rathbun, University of Southern California. 33

Introduction by Robert Jervis, Columbia University

This roundtable extends the debate on the future of liberal internationalism (LI)  started by Charles Kupchan and Peter Trubowitz (K&T) three years ago which provoked responses by Steven Chaudoin, Helen Milner, and Dustin Tingley (CMT), and by Joshua Busby and Jonathan Monten (B&M).[1]  This was the subject of a panel at a recent meeting of the American Political Science Association, and the memos written by the three teams form the basis for this Roundtable.  Despite – or because – of my lack of involvement in these issues I was asked to chair the panel and so here will make a few introductory remarks, and after those involved have had their say will outline a few ideas for further research.

Building on public opinion research by Robert Shapiro and his colleagues,[2] K&T argue that in neither elite nor mass opinion is there a consensus supporting the LI that underpinned American policy during the Cold War.  LI was built on the combination of multilateralism and the willingness to use force, and Republicans have now rejected the former, while Democrats have abandoned the latter.  The result will be paralysis, vacillation, or perhaps a new but much less ambitious compromise.   K&T see the causes as residing in both the less compelling external environment in which the United States finds itself and, even more, in changes in American political, social, and economic life.  CMT define LI more broadly, do not find a decline in bipartisan support for it either in public opinion or in Congress, and argue that the substance (what they call outputs) of American foreign policy has changed much less than would be the case were the K&T argument correct.  B&M’s research leads them to take a position in between, but closer to CMT.  Looking at a variety of indicators, they too find that the K&T picture is overdrawn, if not incorrect.  But they also detect shifts away from LI, especially among Republicans.

The papers written for this Roundtable are clear and succinct enough so that further summary on my part is not needed.  Clearly, much depends on the definition of LI, and one reason why K&T see it as no longer viable is that define it narrowly in terms of multilateralism and the willingness to use force while CMT define it more broadly as support for continuing involvement with the world.  Definitions of course are not right or wrong, but I can see difficulties with both definitions.  Short of isolationism, few policies will fall outside the CMT definition and it is quite possible that there could be a consensus on their form of LI without there being agreement on more specific actions.  I am more sympathetic to K&T’s approach, but we need to ask whether opposition to some forms and some instruments would put one outside of their definition of LI.  Would opposition to the International Criminal Court mean that one would not be in favor of multilateralism; would opposition to a heavy American commitment to counter-insurgency in Afghanistan mean that one did not have sufficient faith in military force to qualify as an internationalist? That said, I cannot take the next step and offer what I believe would be a more useful definition.

The fact that this debate is about the extent of changes over time should make it of particular interest to historians.  In a common form of presentism, arguments of this kind often fall into the trap of imagining a “Golden Age” when things were much better.  K&T are careful to say that foreign policy never stopped at the water’s edge and that partisanship often was very high, even–or perhaps especially–in the first years of the Cold War.  Nevertheless, concentrating on the present disharmony runs the danger of distorting the past.

The final question I would like to raise is whether we have a problem with partisan polarization or a problem with Republicans.  One has to take the latter possibility seriously when 31% of Republicans say that President Barack Obama is Muslim.[3] While it is true that one can find a significant number of Americans who are willing to believe all sorts of crazy things (a third of Democrats blame the Jews for the financial crisis and about as many said that George W. Bush knew about 9/11 in advance),[4] the demonization of Obama does seem to be a symptom of something deeper and is accompanied by the opposition of many Republicans to the anodyne New START treaty with Russia, despite its endorsement by the leading hard-line Republican officials.  But we must be especially careful here.  Both the K&T and B&M teams note the post-Vietnam defections by Democrats with the understanding that the international order requires force as well as multilateralism, and here as elsewhere the fact that most social scientists who study these questions are Democrats may affect our research, a topic discussed in a recent Roundtable.[5]


Many discussions simply peter out.  I hope that does not happen here because the issues are important for both scholarship and public policy.  To facilitate further discussion, let me note a few issues that I think call for additional research.  Procedurally, this area seems ripe for what Philip Tetlock has called “adversarial collaboration.”[6] People with different ideas about the subject could collaborate on what evidence they think would help answer the questions and even conduct joint analysis of the data.  This does not and should not mean that the disputes will be easy to settle–people should not be too quick to abandon positions for which there is significant logic and evidence–but in principle it should be possible for people who hold different views to reach at least some agreement on what sorts of evidence would be particularly relevant.

Turning to substance, all three teams agree that it is most important and difficult to establish measures (be they quantitative or qualitative) of American foreign policy.  If the argument is about the past and future of LI, we need not only an abstract definition, but judgments about what policies do and do not fit with it.  This is a notoriously difficult problem, and one in which our own political preferences may inevitably play a role.  For some, Bush’s stress on spreading democracy abroad makes him a liberal internationalist, while the lack of enthusiasm for such policies among Democrats indicates that it is not only in their hesitancy to use force that they have abandoned this tradition.[7] Others of course disagree.  Nevertheless, it is hard to see how we will make progress on this topic without tackling the question of how we are to categorize foreign policy outputs.

Other fruitful areas for research are more specific, and perhaps more tractable.  One is the extent to which generational changes are involved.  B&M have raised this question, but more work is needed at both the elite and mass level.  We are well beyond the World War II generation, and people who have no memories of the Cold War are beginning to enter positions of power.  Do they have fundamentally different outlooks? Of course it can be difficult to separate generational from positional effects (i.e., younger members of the elite may adopt different views as they gain greater power).  This is not likely to occur with the mass public, however.  Although it would be a mistake to simply project current views into the future, seeing whether there are significant generational changes is certainly worthwhile.

Another change may be more apparent than real.   The polarization in Congress and in the general public is in part the product of the migration of conservatives out of the Democratic party and, to a lesser extent, a movement of liberals out of the Republican ranks, especially in the Northeast.  The parties have become more homogeneous, just as political scientists in the early 1950s thought was desirable.[8] To the extent that this is at work, a polarization of parties and mass opinion is not a sign of a change of attitudes on either level.  Although the change still could be significant because compromise was easier and rhetoric was less harsh when the differences were within as well as between parties, we want to know the extent to which this kind of realignment is behind the polarization.  Another question, probably harder to get at, also relates to the possibility for agreement.  This is whether we have seen a waning of the desire for and ability to construct compromises.  Without lapsing into Golden Age thinking, it does seem as though it used to be easier for political leaders to talk quietly together and to join with opinion leaders in the private sectors to develop widespread support for policies.  As both a cause and an effect of the change, the fragmentation of the media has been much discussed, as has the replacement of broadly trusted leaders in business, labor, and education with the rise of celebrities.  No one could write about the Council on Foreign Relations as a focus for the development, articulation and legitimization of foreign policy anymore, for example.  So we need to explore whether and why American political life has become generally more contentious.

Finally, some of the debate turns on the significance of data on Congressional roll calls.  The complex details are covered in the substantive papers of this Roundtable, and I just want to note that this information is filled with both potential and problems.  One difficulty is that Congressional voting is highly strategic as members of Congress maneuver to place themselves in a good light and their opponents in a bad one.  Procedural issues can be used to avoid being seen as taking unpopular positions, and what may be most important are issues that never get to the floor, a possibility that complicates attempts to measure changes over time.  These questions are crucial to students of Congress as well as to those working on foreign policy, and historians who want to study changes over time also have a stake in using this information.  So this is an area for fruitful discussion among and within fairly diverse scholarly communities.

Much remains unclear or in dispute, but I hope others will find this topic as interesting and important as I do and will join the research and discussion.


Robert Jervis is Adlai E. Stevenson Professor of International Politics at Columbia University.  He has written a number of books including American Foreign in a New Era (Routledge, 2005) and Why Intelligence Fails: Lessons from the Fall of the Shah and Iraqi WMD (2010).  He was President of the American Political Science Association in 2000-01 and in 2006 received the national Academy of Sciences’ tri-annual award for contributions of behavior science toward avoiding nuclear war.   He is co-editor of the Cornell Studies in Security Affairs.

Essay by Josh Busby, University of Texas-Austin and Jonathan Monten, London School of Economics and Political Science

“Off-Center: Misplaced Emphases in Debates about Liberal Internationalism”

We read with great interest the recent exchange in International Security on the subject of whether liberal internationalism (LI) in U.S. foreign policy is in inexorable decline. We give credit to Charles Kupchan and Peter Trubowitz (K&T) for staking out a clear position in this debate and in many ways initiating the research agenda on this question. We credit Steven Chaudoin, Helen Milner, and Dustin Tingley (CMT) for advancing the discussion by bringing new evidence to bear on the question of LI’s fate. In this essay, we address three issues. First, we summarize the evidence from our own work on this issue, and what we think this means for the overall question of whether LI is in decline. Second, we raise several questions about the evidence presented in CMT’s recent International Security paper. While this exchange (including K&T’s 2007 “Dead Center” article and their recent reply) advanced the debate on this question, we feel that both miss the mark in important ways.[9] Third, we outline what we think is an important question not answered by this research literature so far. Specifically, what happened to support for multilateralism among elites within the Republican Party?

In order to determine whether LI has suffered a permanent decline, we first have to understand the concept and the conditions under which it was initially successful. Because our colleagues understand LI and its bases of support differently, they ultimately clash over an area they both agree was historically identifiable with LI, bipartisanship. While bipartisanship has become an important historic symbol of LI, it was not the substance of it, so a debate focusing on whether bipartisanship has gone up or down provides only a partial picture of the role of liberal internationalist ideas in American foreign policy.

What then is LI? LI was the set of animating ideas that governed U.S. foreign policy in the Cold War. LI emerged out of the searing experience of the Great Depression and World War II. In the aftermath of the Second World War, supporters of LI recognized that peace and prosperity could only be restored with an America engaged in support of the new liberal order, which would require support for free trade, foreign assistance to rebuild Europe, the creation of new international organizations, and, as the intentions of the Soviet Union became clear, a willingness to back this order through military spending to keep pace with or surpass the power of the Soviet Union. These elements roughly can be grouped into a cooperative and coercive component. Politically, two groups rallied behind the idea of LI: left of center groups that supported the cooperative side of international engagement (free trade, international organizations, and foreign assistance) and centrist internationalists, which included both Democrats and a number of Republicans, that supported cooperative tools of engagement but also were willing to use force, spend resources on the military, and were prepared to match the military power of the Soviet Union.

The structural challenge of the Soviet Union made this combination functional for the success of the postwar liberal order. We agree with K&T: while not structurally determined, the international system exercised strong pressures that channelled U.S. decision-makers in the direction of supporting both cooperative and coercive elements of engagement, with limits to the extent to which the United States could rely on either on its own without risking Soviet incursions into western Europe (if the United States solely relied on soft power resources) or without triggering a third world war (if the United States became too bellicose, unilateral, and intemperate in its relations with the Soviets). At the same time, the lessons of the Depression and World War II were not lost on the public, which rallied around the idea that America needed to stay engaged internationally to prevent global economic collapse and world war.

Summary of Our Previous Findings

Is LI, or, in our parlance, “establishment” internationalism in decline?[10] In our first paper on this question, we tried to assess this claim by identifying a set of measures to capture trends in support for internationalism. If LI were eroding, we thought this would be evident in the level of internationalist content in presidential State of the Union speeches and political party platforms; polarization in congressional roll-call voting as measured by a now dovish interest group, the Americans for Democratic Action (ADA), and a hawkish interest group, the American Conservative Union (ACU); and in a data set we created on the social backgrounds of a cross-section of U.S. foreign policy leaders since 1945. We found that internationalist content had not substantially decreased in the public statements made in State of the Union addresses by either Republican or Democratic presidents, or in Republican and Democratic Party platforms. When we looked at important substantive votes on foreign policy identified by the ACU and ADA, we found that Republican and Democratic lawmakers have become increasingly polarized in voting at ideological extremes. But, this has been true for some time, as Democrats turned against military spending after Vietnam, and Republicans became increasingly hostile to multilateral institutions and foreign assistance after the Cold War. Finally, when looking at the social backgrounds of political appointees, we found no statistically significant change in where foreign policy elites grew up or went to university, but we did find a decrease in the percentage of elites with military experience. We also found that the younger generation of policymakers for whom the Vietnam War was the first formative foreign policy experience of their twenties and early thirties still represented a relatively small proportion of elites (about 27%) who served in upper foreign policy positions in the post Cold War period 1992-2006.

We drew two conclusions from this evidence. First, the tenets of LI still are important rhetorical touchstones for both parties. Second, while K&T may be correct that LI is in terminal decline, either this is not captured by the measures we chose, or the full impact of these forces has yet to occur.

A second paper, forthcoming in Political Science Quarterly, looks at a different version of this question, focusing on trends in public and elite opinion.[11] The George W. Bush administration departed from the liberal brand of internationalism in many of its key foreign policy decisions. However, this shift began (both within the Bush administration itself and for the congressional Republicans that came to power in the 1990s) before September 11th, frequently cited as the “shock” that caused the United States’ dramatic turn towards a policy of unilateralism and the aggressive use of national power. We seek to answer two questions: were these policies broadly out of step with foreign policy attitudes of the wider U.S. public, or were these policies broadly supported by the public?

To answer this question, we looked at survey data from the American National Election Studies and the Chicago Council on Global Affairs (CCGA) from 1982 to 2004. We used questions from the CCGA surveys (asked in all years and in both the elite and public surveys) to construct scales measuring support for the hard power dimensions of internationalism (support for defense spending, maintaining military superiority, and deploying military forces in various conflict scenarios), and the soft power dimensions of internationalism (support for free trade, foreign aid, and the United Nations). Our goal was to evaluate whether foreign policy attitudes among Republican Party elites are becoming less internationalist in comparison to the wider U.S. public.

Three results emerged from this analysis. First, foreign policy issues were not particularly salient to the public during the 1990s when congressional Republicans first came to have majority status. That these Republicans were elected largely on domestic grounds, but were then able to impose a more unilateralist foreign policy agenda through control over congressional veto points such as committee chairmanships, lends some support to the line of argument that suggests Republicans imposed these policy preferences despite public opposition. Second, public foreign policy attitudes did not sharply change in response to the end of the Cold War, although on several dimensions elite attitudes did. Finally, the question of whether the gap between the foreign policy attitudes of Republican elites and the public has increased is more nuanced than it first appears. Republicans did diverge from the public by moving in a more hard-line direction beginning in the early 1990s (that is, they supported the hard power elements of our scale in greater numbers, particularly in deploying forces in conflict scenarios). However, Republican elites have also been more consistently internationalist than the public in terms of support for free trade and foreign assistance. On the soft power dimensions of internationalism, the key exception is multilateralism. Here, support for the United Nations among Republican elites has decreased dramatically in comparison to the wider public. This last point is further confirmed by looking at other measures of multilateralism asked in the CCGA surveys since 2002, where support among Republican elites for new multilateral institutions is substantially lower when compared to both the general public and the Republican public.

Does the Evidence in “The Center Still Holds” Hold?

While creative in their effort to develop new measures, we think the evidence in CMT has several potential flaws. First, their focus on bipartisanship leads the authors to track phenomena such as congressional gridlock, legislation co-sponsorship, and roll call voting that do not directly measure the underlying concept of LI. Even if they are correct that the preponderance of evidence supports the view that polarization in foreign policy is not occurring, it may nonetheless be the case that the consensus foreign policy view is moving away from LI. They acknowledge this when writing “Our analysis here focused on the degree of bipartisanship in foreign policy, but not on whether particular bills were pro- or anti-liberal internationalism” (92).

In terms of public opinion, we agree with K&T’s critique that CMT’s measure of public opinion, which shows consistently high support for the idea that the U.S. should play an “active part in world affairs” since the end of the second World War, merely registers the public’s basic tolerance for international engagement but says little about their preferences for liberal vs. other forms of internationalism.[12] In addition, on at least one key issue – support for participation in multilateral institutions – there appears to be clear evidence of partisan polarization over the last decade. Shapiro and Bloch-Elkon, for example, find using the CCGA survey data that Republican and Democratic respondents have diverged on a large number of foreign policy issues since 1998.[13] In our own analysis, we find that the opinion gap between Republican and Democratic elites has increased on many foreign policy questions: on the question on whether to strengthen the United Nations, the gap between party elites increased to over 40 percentage points in 2004, and the difference was over 60 percentage points on whether to support the International Criminal Court (ICC), although party elites remained close on several other issues, such as support for free trade. CMT rightly point out that K&T use two separate questions in measuring polarization in public opinion but do not contest the fact that the question on military spending shows a clear trend in polarization over whether the US should increase military spending.[14] It seems plausible that support for international engagement has remained high among both parties, but that polarization has increased over the precise form this active role should take.

Is “Dead Center” on Target?

In their 2010 response, K&T concluded that bipartisan support for LI is dead and that accordingly, “It is far preferable for the United States to pursue a more modest grand strategy that enjoys domestic support than to pursue an overly ambitious statecraft that further polarizes the nation and leaves an uncertain world without the benefit of measured and steady U.S. engagement” (109). They may be right, but we wonder whether their ultimate conclusions about LI’s demise are still premature. What they describe as the pragmatic tempering of U.S. global ambitions to ward against more sweeping public disaffection with foreign entanglements may be an overreaction. The public continues to tolerate substantial military involvements in Iraq and Afghanistan as well as significant foreign assistance for global causes such as AIDS treatment programs. Table 1 shows that the public continues to support participating in new multilateral institutions, such as the ICC (76% in 2004), a climate agreement (71% in 2004), and a nuclear test ban treaty (87% in 2004).[15] Of course, a general expression of approval does not capture how salient these issues are for the public as policy priorities or whether the public is willing to incur costs to achieve them, and support for many of these institutions may have softened during the recent financial crisis. Nonetheless, we wonder whether K&T’s call for a modest retrenchment from international engagement is anticipating a public backlash rather than a reaction to it.

The second question raised by K&T’s analysis relates to their proposed grand strategy. The international environment is characterized by increasing economic interdependence and complex transnational problems that require global solutions. In this environment, a strategy of retrenchment potentially exposes the country to systemic costs. Less international ambition potentially increases our vulnerability to the negative externalities wrought by globalization and the vicissitudes of the decisions of other actors.

It is also not clear that retrenchment does anything to solve the problem of partisan drift. What would the coalition supporting this strategy look like and what are the core policies around which it would converge? In their 2007 Foreign Affairs piece, K&T recommend a suite of policies that they say is more politically sustainable including support for regional institutions, rebuilding hard power, and pursuing energy independence.[16] At the same time, they recognize that some global problems will continue to require formal institutionalized cooperation. While we tend to agree with their diagnosis of the problem, many of their prescriptions themselves are extremely contentious politically. No winning coalition has yet to form on energy policy, and Secretary of Defense Robert Gates’ proposals for cutting military spending have engendered intense opposition. In general, the Democrats already moved to become less of a party of what K&T call “partnerships” but not power (though its pacifist wing might be more ascendant again after Iraq and the financial crisis), but Republican elites have not yet come to terms with the idea that formal institutionalized cooperation continues to be necessary to solve key global problems.

Next Steps: Republicans and Multilateralism

In our view, changes in public opinion are not the most important factors driving a shift away from LI. Indeed, the main change we observe is that a subset of Republican elites has lost confidence in multilateral approaches to foreign policy. Throughout the post-war period, Republican leaders represented a consistent pillar of support for international engagement, not only by supporting high levels of defense spending but also by consistently supporting active U.S. leadership in international organizations and multilateral forums. Beginning in the 1990s, however, Republican politicians increasingly questioned the value of these multilateral institutions and whether the U.S. should participate in them. A new wave of conservative intellectuals began arguing that these institutions undercut U.S. sovereignty and freedom of action, and are in many respects illegitimate.[17] Few put their disdain for international organizations as pointedly as John Bolton, who stated in 1994: “There are 38 floors to the U.N. building in New York. If you lost 10 of them, it wouldn’t make a bit of difference.”[18] This faction reflexively opposed international organizations (like the United Nations) and agreements (like the Law of the Sea Treaty, the Comprehensive Test Ban Treaty, a comprehensive climate agreement, the International Criminal Court, the Landmines Ban Treaty, and the new START treaty on arms control). What had been minority views within the Republican Party on arms control, multilateral treaties, and international institutions became the dominant position among Republican elites in both Congress and the executive branch. While there may be good reasons to oppose individual treaty instruments, those who oppose all of these instruments have gained the upper hand in Republican elite circles and displaced traditional supporters of international engagement, such as Senator Dick Lugar, Henry Kissinger, Brent Scowcroft, and Colin Powell.[19] The Senate December 2010 vote in support of the new START treaty by a margin of 71-26, including thirteen Republicans, suggests it is also possible to overstate the degree of unilateralism among Republicans.

We think a key part of the story of LI’s demise is the rise of a unilateralist faction within the GOP and how this faction was able to displace traditionalist moderate internationalists within the party. One possible explanation is that in the 1990s the absence of systemic pressures, coupled with public disinterest in foreign policy, gave policymakers room to pursue idiosyncratic policies that catered to the interests of the handful of activists who remained engaged on their pet causes. Documenting and explaining the displacement of Republican moderate internationalists by neoconservatives, unilateralists and isolationists remains an important research question.

Table 1: Support for Multilateralism

  1994 2002 2004
Participate in UN Peacekeeping
All Leaders 87.37 79.9 83.56
Republican Leaders 88.10 68.87 71.96
Public 74.13 65.22 77.51
Republican-Public Gap             13.97                  3.65           -5.55


Ban Nuclear Tests
All Leaders 83.12 83.56
Republican Leaders 60.19 59.81
Public 84.77 86.96
Republican-Public Gap              -24.58        -27.15


Landmines Ban
All Leaders 75.06 78
Republican Leaders 50.96 52.34
Public 80.21 79.77
Republican-Public Gap              -29.25        -27.43


ICC (International Criminal Court)
All Leaders 66.15 68.89
Republican Leaders 39.05 34.58
Public 75.97 75.55
Republican-Public Gap              -36.92        -40.97


Kyoto Protocol
All Leaders 64.18 70.89
Republican Leaders 23.81 27.1
Public 75.12 71.35


Chi square difference between Public and Leaders statistically significant for all years at the p<.01 level, except for the landmines ban in 2004 and the nuclear test ban in 2002 and 2004 (not significant); 2002 landmines significant at the p<.05 level. Chi square difference between Public and Republican Leaders statistically significant for all years at the p<.01 level, except for peacekeeping in 2002 (not significant).

Joshua W. Busby is an assistant professor at the LBJ School of Public Affairs and the Crook Distinguished Scholar at the Robert S. Strauss Center for International Security and Law at the University of Texas at Austin. He holds a PhD in political science from Georgetown University. He is the author of Moral Movements and Foreign Policy (Cambridge: Cambridge University Press, 2010). Prior to coming to UT, Dr. Busby was a research fellow at the Center for Globalization and Governance at Princeton’s Woodrow Wilson School (2005-2006), the Belfer Center for Science and International Affairs at Harvard’s JFK School (2004-2005), and the Foreign Policy Studies program at the Brookings Institution (2003-2004). His works have appeared in Perspectives on Politics, Security Studies, International Studies Quarterly, Current History, and Problems of Post-Communism, among other publications.

Dr. Jonathan Monten is LSE Fellow in Global Politics in the Department of Government at the London School of Economics and Political Science. He received his PhD from the Department of Government at Georgetown University.

Essay by Stephen Chaudoin, Princeton University, Helen V. Milner, Princeton University, and Dustin Tingley Harvard University

“The Contested State of U.S. Foreign Policy: Liberal Internationalism and American Politics”

In a recent paper, we re-evaluated the claims of Charles Kupchan and Peter Trubowitz (K&T), and many others, that “liberal internationalism” is finished as a foreign policy grand strategy since polarization on foreign policy issues has increased so much over time.[20] In their original article, K&T suggested two “break points” that accelerated this increase in polarization: the end of the Vietnam War and Cold War.[21] They provided public opinion and roll call voting data from the U.S. House of Representatives to support their claims, though they did not conduct any statistical analysis. As in the original article, we also looked at public opinion and roll call data. Neither data source supported the conclusion that bipartisanship in foreign policy has declined since the end of the Vietnam or Cold Wars. We also presented two new pieces of evidence—about congressional gridlock and co-sponsorship patterns—to further refute the claim that polarization over foreign policy has increased. Our goal as social scientists was to test their proposition empirically using systematically constructed and analyzed data. Like others, we found little support for their conclusions.[22]

K&T offered a rebuttal of our arguments and now claim sweeping new support for their original conclusions amongst the community of analysts without citing the people to whom they are referring or providing data to further support their claim.[23] Our original critique focused on replication and reanalysis of their original arguments concerning bipartisanship in Congressional roll call votes on foreign policy and used data that they chose in their initial article. K&T’s main rebuttal is that we “do not address the actual conduct of U.S. foreign policy in making [our] case for the continued vitality of liberal internationalism” (96). This line of defense is puzzling since the  same could be said of “Dead Center.” If our analysis does not adequately address U.S. foreign policy, then their original article also shares this insufficiency. Assuming analysis of data like Congressional activity and public opinion does let us study the conduct of foreign policy we continue our sustained skepticism of their claims. Where we reanalyzed data sources chosen by K&T, we explained our reasons for skepticism clearly and showed how their data did not support their conclusions. Where we presented new data sources, we carefully described their importance in determining if polarization had resulted in the demise of LI.

In light of our recent debate, we think it would be useful to develop a way of evaluating these types of debates as well as to revisit some of the specifics of the debate. How should we think about the overall construction and evolution of United States foreign policy? What is LI? How do we know if a policy is part of the LI program? How do we know when a country’s grand strategy changes? What role does domestic politics play in the development and evolution of a grand strategy? Is polarization rising over foreign policy issues in the U.S., and does this matter for foreign policy choices? What kinds of data and approaches would be most useful for evaluating claims in this debate?

In this article, we outline briefly the three arguments that we develop in greater depth in another paper. First, we discuss the concept of LI. Second, we develop a heuristic for how to think about the construction of U.S. foreign policy and link this to a discussion about the types of data necessary for evaluating claims like K&T’s. We decompose foreign policy into three steps: inputs, legislative and political processes that aggregate and manipulate these inputs, and policy outputs. Bipartisanship is about how process affects outputs such as LI policies; like K&T, we think that Congress’ behavior matters to American foreign policy. We show how this heuristic sheds light on debates about the changes in foreign policy by more tightly linking the analysis with the broader questions asked here and by better informing data collection. Third, we revisit the particular debate about LI with this heuristic in mind. We continue to maintain our original position: polarization on foreign policy has not increased over the last thirty years, and there is little evidence pointing to systematic changes in U.S. foreign policy like those hypothesized by K&T. When evaluating each piece of evidence with our broader heuristic in mind, this point becomes even more clear.


While there is debate, some common elements of LI can be identified. First, most scholars associate it with Woodrow Wilson’s ideas and foreign policies.[24] A core idea is that the spread of democracy at the domestic level and the development of an international order made up of democracies would lead to world peace. Fostering democracy abroad is thus of great importance. The LI project is associated with support for democracy and human rights across the globe. A second element of this conception is the emphasis on openness in the international economy, especially through trade. International commerce, as opposed to mercantilism or autarchy, was a central element of the liberal vision for world peace. Enhancing international economic exchange was then a component of LI. Finally, the third element of LI that is often mentioned deals with multilateralism. Much as in Wilson’s vision of a League of Nations and collective security, LI is associated with international cooperation, often in the institutionalized form of multilateral organizations designed to coordinate states’ behaviors around agreed upon rules and norms.

In their original article, K&T claim that the “liberal internationalist compact that guided the United States for much of the second half of the twentieth century…. entailed a commitment to both power and cooperation: the United States would project its military strength to preserve stability, but it would seek to exercise its leadership through multilateral partnership rather than unilateral initiative. It was the coupling of U.S. power and international partnership that gave the nation’s foreign policy such a distinctive character in the decades following World War II” (pg. 8). [25] In their follow-up article, K&T write “Following a substantial literature on this topic, we define the LI compact as the combination of heavy investment in military force with a commitment to international institutions.”[26] For them, the death of LI implies that foreign policy will gyrate between the two elements of LI: “It follows that in a post-bipartisan era, U.S. foreign policy should oscillate between ideological alternatives when power changes hands. When Republicans are in office, U.S. foreign policy should favor military power over international partnership. When a Democrat occupies the White House, the administration should be more inclined to invest in partnership.”[27]

It is interesting to note that they never mention the important liberal idea of a community of democracies as part of LI, nor the support for democracy globally that this entailed in LI. Since the end of the Cold War at least, however, a common element in the policies of many U.S. administration has been the international support for democracy and democratization. George H. W. Bush, Bill Clinton, George W. Bush, and Barack Obama have had democracy promotion as part of their foreign policy.[28] Once the U.S.S.R. disappeared, the common theme of enhancing democracy globally has permeated American foreign policy in all administrations. And each administration has committed American force to do that, albeit in different circumstances and parts of the world.

Moreover, the second element of traditional conceptions of LI has been neglected in K&T’s definition. The construction of an open international economy supported by international institutions to manage it has been a central and common element of foreign policy across many American administrations. Support for trade liberalization, the World Trade Organization, free trade agreements, the World Bank and the International Monetary Fund has been constant among presidents in the United States. Even Clinton, a Democrat, moved forward on trade liberalization in his term, as Obama now seems to be trying to do in the wake of the financial crisis. These central elements of LI have endured, but they are overlooked in K&T’s definition.

Instead K&T focus on the exercise of American force and multilateralism. The idea of American hegemony and its use of force was not a primary element in the long tradition of LI. It is an element that was added to the LI program as the U.S. emerged from the Second World War in an unusually powerful position. LI is, if anything, ambivalent about the use of force in world politics. In its conception, promoting democracy and an open international economy will render the system more stable and peaceful, so that force will not be necessary. But the inherent tension between the desire to use American power to foster LI and the importance of multilateral cooperation is recognized in the LI tradition. So in some ways the K&T notion of LI seems different than the traditional conception. Its focus on the use of force does not seem to fit with standard notions of LI.


It is helpful to think about three different stages to the construction of foreign policy: inputs, political processes, and policy outputs. These three stages track the process of U.S. foreign policymaking, from the preferences of the relevant actors to the policy choice that is the final result. Such a model is heuristic in that it can help one think about how to systematically analyze questions about changes in U.S. foreign policy. In the following sections, we discuss each piece of evidence individually and place each within the broader context of foreign policymaking. When evaluated in this context, it becomes clearer that the data do not show a collapse in LI stemming from increased domestic polarization.




We begin with the final stage of the construction of foreign policy and consider the outputs. What are the actual choices that the U.S. makes with respect to its foreign policy? Next we empirically and theoretically connect the earlier two stages of foreign policy making to the final policy output stage. Empirically, we emphasize that there is little evidence in support of K&T’s claim about dramatic changes in U.S. foreign policy at the output stage. We observe neither systematic declines in LI policies over time nor dramatic oscillations across administrations. In our longer paper, we demonstrate that the main indicators of U.S. foreign policy do not show any systematic decline in their LI components: overall foreign aid and military assistance the percentage of U.S. foreign aid spending that is multilateral, U.S. military personnel and procurement, U.S. funding for the IMF and World Bank, U.S. military deployments, peacekeeping expenditures, and democracy promotion around the world. There is little evidence that since the end of the Vietnam War or Cold War, the U.S. has experienced a substantial, secular change in any of these components of its foreign policy.[29] Theoretically, we show how trends are unconnected with the types of mechanisms discussed by K&T.


We would emphasize that our original measure of gridlock was the only output measure considered in any of the papers. While we used gridlock data as a direct response to K&T’s claims about the percentage of moderates in Congress, the gridlock measure was one way to evaluate if the supposed rising ideal point extremism was inhibiting the creation of foreign policy “outputs” that were demanded by the pressing issues of the day (“inputs”). Congressional gridlock on foreign policy has not increased over time, a finding which K&T do not challenge. Regardless of its bearing on the discussion of Congressional moderation, the gridlock data show that Congress has continued to address foreign policy issues in need of action.[30]




Like K&T, we focus on the legislative part of foreign policy making. Why focus on Congress? First, polarization in Congress might reflect larger changes in inputs and preferences about foreign policy among elites, interest groups and the public. If we found increased legislative polarization over time, it might reflect a larger underlying change in American attitudes toward international politics. This argument presumably drives some of K&T’s focus on voting patterns. Second, Congress might really matter. William Howell and Jon Pevehouse show that even in the domain where presidents have the most unilateral capacity, military intervention, presidents must worry about Congress, and Congress can check the executive and alter policy choices.[31] We think Congress matters domestically as well as internationally. To the extent that a foreign policy requires a credible commitment by the U.S. for its success, more congressional opposition renders the policy increasingly ineffective.


Has behavior in the legislative process changed so as to render LI impossible because of polarization, as K&T fear? We see little evidence for this and make three arguments. First and most directly, partisanship on foreign policy decisions has not become more polarized. In our research, we found no support for K&T’s conclusion that bipartisanship since Vietnam has steadily declined and that this decline was further hastened with the end of the Cold War. We provided analyses of all House foreign policy votes, as well as more nuanced analyses of different types of votes, such as procedural, amendment and final passage votes. Perhaps our original exposition was not clear enough. Our data show there was no statistically significant decline in bipartisanship beginning with the Vietnam War using all foreign policy votes, regardless of whether they were procedural, amendment, or final passage votes (see figure 5 in our “The Center Still Holds”, and an updated version in Figure 1 below that includes data up to 2008 confirms this). Indeed, post Cold War levels of bipartisanship showed evidence of a slight increase in bipartisanship, not a decline as K&T maintain.


Figure 1

ISSF-Roundtable-2-4-Figure 1

Predicted probabilities of House bipartisan foreign policy vote as a linear function of year using probit regression. Only procedural votes for foreign policy had a significant decline over time.

Second, our claims in “The Center Still Holds” carefully dealt with issues about vote types by examining trends in procedural, amendment, and final passage votes, rather than stopping at an examination of aggregated data. We only found support for K&T’s claims when looking at procedural, and to a lesser degree, amendment votes.[32] As we report, the main area where there exists a decline in bipartisanship beginning with the end of Vietnam War and continuing through the Cold War is in procedural votes only (see figure 5 of our article). This significant pattern is not seen for amendment, final passage, or the set of all foreign policy votes. The trends in frequency of procedural and amendment voting were not the result of the Vietnam War or the end of the Cold War and they did not herald the large scale changes in U.S. foreign policy that K&T claim. Rather, they are a product of domestic political changes, like voting rules and changes in party leadership strategies, that are unrelated to support for and the conduct of U.S. foreign policy. These declines coincided with unrelated changes in domestic politics, and are not indicative of the broader trends argued by K&T.

Their response to our findings is essentially that procedural and amendment votes matter in terms of  substance and that we claimed they did not. We had carefully anticipated this argument in our original paper.

We do not argue that amendment votes contain no foreign policy substance or importance. Instead, the sharp increase in amendment voting combined with the remarkably steady level of bipartisanship on amendment voting suggests that procedure is the culprit for the apparent decline in bipartisanship, rather than a fundamental disagreement over substance. If bipartisanship on amendment voting had decreased significantly over time, then this would be more troubling, but additional statistical analysis shows this is not the case.

Even if some procedural and amendment votes matter for policy (something that we agreed with, and the reason why they were included in our analysis), this would not change the fact that a spike in the number of procedural and amendment votes drives the trends that K&T use to make their claims. They are thus in a tough position to explain why support for a key part of their thesis should rest on changes in procedural voting patterns only. Furthermore, in our longer paper we show that none of K&T’s hypothesized changes are present in the Senate. In particular, we show that there has been no systematic decline in bipartisanship on foreign policy votes since the end of Vietnam in the Senate. Thus looking to other key institutions in the U.S. political system further casts doubt on K&T’s claims.

Third, we cast further doubt that on their claim that changes in voting patterns reflect some groundswell change in LI by presenting new analyses comparing domestic voting with foreign policy voting. Explanations that link changes in underlying preferences with bipartisanship in Congress are much stronger for domestic policy than for foreign policy. Visible trends like rising income inequality have been linked to more polarized preferences over important domestic issues, like taxation and redistribution policy.[33] K&T link these trends to increasing polarization over foreign policy. But this same dynamic need not be at work on important foreign policy issues. For instance, few people view questions of troop deployment or United Nations resolutions as redistributional issues. Empirically, the link between these trends and preferences over domestic policy is clear, but the link between things like income inequality and increased polarization over foreign policy is much less so.

Figure 2 graphs the average levels of bipartisanship in each year for all House votes by foreign versus domestic policy. Even when we update this to the end of the 110th Congress, when looking at all domestic policy votes versus all foreign policy votes, the levels of bipartisanship are generally higher in foreign policy voting. In the vast majority of years, bipartisanship is higher in foreign policy than in domestic policy.

Figure 2

ISSF-Roundtable-2-4 Figure 2

Percentage bipartisan votes by Congressional session. Dashed line for domestic policy votes and solid lines for foreign policy votes. In most years bipartisanship is higher for foreign policy votes (solid line) than domestic policy (dashed line). In a majority of Congressional sessions these differences are significant, including post-Cold War sessions.

In figure 3 we show that the predicted probability of all domestic policy votes being bipartisan over time is declining, compared to foreign policy (figure 1), where the amount of bipartisanship has not changed at all. Regardless of vote type, the decline in bipartisanship for domestic policy votes has been steeper than the decline for foreign policy votes. The same patterns hold for the Senate. In figure 4 we see virtually no systematic changes in the level of bipartisanship beginning with the end of the Vietnam or Cold Wars, no matter what type of votes we analyze.

Figure 3

ISSF Roundtable-2-4 Figure 3

Predicted probabilities of House bipartisan domestic policy vote as a linear function of year using probit regression. A significant decline in procedural, amendment, and all votes was observed.

Figure 4

ISSF Roundtable-2-4 Figure 4

Average Senate bipartisanship levels on all foreign policy votes for each Congress. Bipartisanship increased in the beginning of the 20th century but did not systematically decline following the end of the Vietnam and/or Cold Wars.

Roll call voting is one important metric for studying the behavior of Congress. Some have criticized the use of roll call votes since they are not a random sample of all bills.[34] In CMT, we sought out alternative “testable implications” of K&T’s argument. Cosponsorship patterns were one idea. Our conjecture was if there is a decline in bipartisan foreign policy, this should be evident in cosponsorship patterns, as legislators would be less likely to want to work together on legislation. Our data show that cosponsorship patterns on foreign policy bills have not changed;[35] Democrats and Republicans reach across the aisle just as much now as they did in the past twenty years to cosponsor legislation on foreign policy. The big issue is if the divisive trends that K&T describe are so strong, then why do these trends not also show up in cosponsorship patterns? Why if partisanship is driving political behavior now more than ever, is it not making legislators shun visible cooperation on bills with members of the other party?

Their response was to claim that the received wisdom in American Politics is that cosponsorship is costless and uninformative of legislators’ preferences. This is an inaccurate assertion, and further, it misses the point of analyzing cosponsorship. Our reading of this literature (including their cited sources, along with research published in this decade) does not support K&T’s  portrayal.[36] Cosponsorship is a meaningful way to study trends in legislative behavior for several reasons. First, cosponsorship is hardly costless. Little that “signal[s] to voters, committee members, and other audiences” is inconsequential to a legislator.[37] Further, opponents can publicize cosponsorship and voters can reward or punish legislators for it, as the American politics literature argues.[38] Secondly, even if cosponsorship is less costly, legislators believe that it matters, and make cosponsorship choices accordingly. The effort that most politicians put in to highlighting their sponsorship and cosponsorship records show that politicians believe that voters do pay attention to them. Politicians consciously choose which bills to cosponsor, and they choose bills whose substance they agree with. Koger (2003) uses interviews with Congress members and their staffs to demonstrate that the people making cosponsorship decisions do believe that they perform an important signaling role for their constituents.[39] This undermines K&T’s rebuttal. Even if cosponsorship is costless or uncorrelated with legislative outcomes,[40] it is certainly not random. Legislators make cosponsorship decisions that reflect their preferences, and cosponsorship patterns have not changed on foreign policy issues over time.


Finally, we turn to the issue of inputs into the policy process. Where do the preferences for certain types of policies come from? For us the issue is where does support and opposition to LI come from in American politics. A main input into the policy process in a democracy like the U.S. is the expressed preferences of the public. Have members of the public have become increasingly divided over foreign policy? Such divisions could then exacerbate partisanship in the legislative process and lead to a shift away from LI. Our earlier data showed that the U.S. public was still largely supportive of international engagement, which is a necessary element of LI. Here we first revisit the details of our recent exchange and then present new data.

Our primary concern with K&T’s original analysis is that they used very different survey questions conducted decades apart from one another to tap what they claimed to be the central element of LI. They claim that their questions dealt with a central issue of LI: confrontation of threatening states. We have argued above why this is not a central element of LI. LI is agnostic on such issues.

Concerning their own public opinion analysis, K&T acknowledge in their response that their original plots did not actually display any strong differences between respondents from different parties, an argument that we pointed out in our paper. They equivocate by arguing that the Korean War likely caused a spike in partisanship in the 1950’s and that the isolated shock of the Korean War likely explains why we do not actually see large partisan spreads, even when looking at their plots. They make this argument in a footnote, but we highlight it here because it is troubling for their own arguments. Much of the fear of a collapse of “coherent foreign policy” as a result of polarization is an artifact of a recent foreign policy development that has been much more polarizing than the Korean War: the Iraq invasion in 2003. Most of K&T’s arguments about polarization during the Obama administration are likely consequences of the Iraq war and its polarizing effect on politics. The vast majority of evidence about recent public polarization over foreign policy points undoubtedly at the Iraq war as the cause. If it is acceptable to point to polarizing international conflicts, then we will concede that the Korean War potentially explains their null results when looking at graphs of public opinion questions, while still standing on the strength of our own analysis. But we would extend K&T’s  own argument to suggest that the Iraq War is responsible for the vast majority of the foreign policy polarization that so worries K&T. We cannot simply dismiss as anomalous the data that do not conform to our arguments, especially when an important motivation for this entire discussion begins with whether or not the G. W. Bush administration was an anomaly.

As an improvement, we wanted a survey question that asked something close to LI or at least concerned a necessary component of it and that was asked repeatedly over a broad span of time. Our question about support for international engagement got very close to the most important (and necessary) component of LI and was consistently asked over a meaningful span of years. Whether or not the public supports U.S. engagement abroad is fundamental to LI by any definition, and any definition that says otherwise should be suspect. K&T did not like this question because other policy orientations that might be considered non-LI could also entail engagement (pg. 99), such as “assertive unilateralism.”[41] Joshua Busby and Jonathan Monten (B&M) echo this argument in their essay.

We agree with the sentiment of their arguments, but maintain that our analysis was still the more informative and conceptually accurate. In considering how to evaluate the preferences of a group that has input over time (so as to test K&T’s theses about changes over time) we faced a tradeoff that we explained our original response. While one might pick apart the meaning of any opinion question and its connection to the theoretical concept (here LI), the “engagement” question is the only one presented in this debate that is comparable over large periods of time and that speaks to a necessary component of LI. It is uninformative– whether one’s goal is to evaluate preferences for LI or anything else– to visually compare partisan spreads on questions about the Korean War in the 1950’s with questions about defense spending from after the Cold War.

Our original work focused on engagement as a necessary component of LI. Our updated work focuses on analyzing particular policies that fit into the broader concept of LI. We have some longitudinal data on elite attitudes toward the United Nations and NATO, both LI endeavors. Did the members of the parties increasingly disagree on support for U.N. or NATO? Our data, spanning 1975-2004, show that the gap between Republicans and Democrats for supporting NATO is nearly non-existent both before and after the end of the Cold War. For the United Nations, the gap between the Republicans and Democrats has widened over time. However, this gap is due mostly to Democrats who have become somewhat more supportive of the U.N., whereas average Republican attitudes have not significantly changed. On this multilateralist dimension of LI—using consistently measured questions over time—elites in the U.S. show little change in polarization between the parties.

B&M’s arguments focus more on public and elite opinions. In relating their own work to the debate at hand, they offer that:

On at least one key issue– support for participation in multilateral institutions – there appears to be clear evidence of partisan polarization over the last decade. Shapiro and Bloch-Elkon, for example, find using the CCGA survey data that Republican and Democratic respondents have diverged on a large number of foreign policy issues since 1998. In our own analysis, we find that the gap between Republican and Democratic elites has increased on many foreign policy questions: on the question on whether to strengthen the U.N., the gap between party elites increased to over 40 percentage points in 2004, and the difference was over 60 percentage points on whether to support the International Criminal Court (I.C.C.), although party elites remained close on several other issues, such as support for free trade (5).

We disagree with B&M that this evidence supports K&T’s original claims. We never contested that polarization exists, or that it often appears highly trenchant.[42] We did, however, contest that polarization had increased over time. Looking at the most recent plots of B&M’s data also suggest that gaps have not been widening over time on questions posed to elites about foreign aid, tariffs, support for the use of “carrots” or “sticks,” defense spending, matching power, or troops to Korea. In fact, the U.N. question is the only one where the gap between the opinions of party elites appears to have increased.[43] In their words, “we observe a persistent divide rather than a growing disconnect between Republican and Democratic leaders on a number of dimensions of internationalism, including foreign aid, support for the U.N., defense spending, and matching military power.” It is very important to remember that our debate is about changes over time, not a snapshot of partisanship now.


A common theme in our responses to K&T’s rebuttals above is as follows: when analyzing any one part of the foreign policy making process, from inputs to outputs, and everything in between, it is crucial to keep in mind how each part fits together to produce policy. Snapshots of any one piece of evidence do not weigh in on the questions at hand unless they are placed firmly in the broader context of the making of foreign policy. What pieces of evidence do the best job of this?

Our arguments about gridlock do the best job on this count, and the results from analysis of gridlock have remained uncontested. Sarah Binder’s gridlock data measure whether or not Congress addressed the important issues of the day, or whether Congress was unable to take meaningful action because of gridlock.[44] She argues that her measure taps the success of the legislature in treating public problems. As new challenges like international terrorism or the spread of infectious diseases have arisen in the global commons, determining whether or not Congress has succeeded in doing something about pressing issues is a particularly meaningful way to evaluate K&T’s original arguments.

These features make gridlock more appropriate for analyzing the politics of LI than K&T’s count of the percentage of moderates. K&T argue that “polarization [does] not preclude radical shifts in foreign policy,” referencing President McKinley’s term, and we would agree with this argument. But this argument applies more to their original count of the number of moderates than to our use of gridlock data. K&T argue that polarization necessarily precluded LI, but this need not be the case. To see why their own arguments are indeterminate, consider: if a president supported power and cooperation and enjoyed sufficient legislative support to pass the necessary legislation, then the vehemence with which the other side objected would be of no import. Our gridlock arguments do not rely at all on these indeterminate causal chains. Our gridlock data starts with a measure of “what Congress needs to get done” as a result of outside forces and the pressing issues of the day and then examines whether or not something actually gets done. Of course, the original gridlock measure may be imperfect for other reasons, but an argument that it has no bearing on the conduct of U.S. policy seems odd. In short, our analysis of gridlock is the best (and indeed only) analysis bearing on whether or not ideological extremism has resulted in the demise of LI because of rising partisanship.


One limitation of the preceding analyses is that they do not extend to the Obama administration because the data necessary to conduct systematic analysis are not available. K&T claim that “In three telling respects, Obama’s presidency has confirmed our prediction that the secular decline in bipartisan support for LI would shape the conduct of U.S. foreign policy regardless of which party controls the White House”.[45] We do not see evidence of the dramatic “swing in the pendulum” that they portray Obama as pursuing.

First, it is important to consider the role played by “inputs” into the selection of Obama as a president. Obama was elected in no small part because of his opposition to the unilateralist foreign policies of the Bush administration. Second, in thinking broadly about “process”, K&T’s  characterization of Obama taking a non-centrist (left wing) position on military deployment is odd, given his decision to build up forces in Afghanistan and his selection of Secretary Robert Gates and other Bush era military officials for top positions. Equally striking is Obama’s expansion of overseas anti-terrorist operations.[46] Claims about Obama’s trade policy are also indeterminate. The fact that Obama has managed to avoid rampant protectionism against the backdrop of the current global economic climate is powerful evidence of his support for international engagement. The recent bipartisan vote on sanctions against China for its protectionist exchange rate policy (a lopsided majority of 348 to 79 in the House) illustrates how Obama can protect openness from unfair trade policies by major trading partners. It also illuminates the importance of bipartisanship in foreign policy, a bipartisanship that remains in Congress. The U.S. threat to China is much more credible with such a large vote margin than it would be if it were a bare majority, especially in an election season. One could cite the Senate deliberations over the START treaty as evidence of high polarization. But numerous treaties have failed to pass in the Senate over time. Some have passed with small majorities; hence, it is not clear that this is a case of rising polarization over time. Furthermore, the treaty passed through the Senate Foreign Relations Committee on a 14 to 4 bipartisan vote. And the Obama administration wants to achieve a bipartisan vote so that the treaty has greater international credibility and stability than it would otherwise, which is one reason the administration is moving slowly to build support for it. The first two years of the Obama administration do not provide support for the idea that foreign policy has become more polarized than it ever was.

In conclusion, we think the Obama administration has made enormous strides in maintaining LI in a very difficult political and economic environment. Hence we conclude, in the [paraphrased] and immortal words of LL Cool J, “Don’t call it a comeback. [LI] has been here for years”.[47]


Stephen Chaudoin is a Ph.D. Candidate in the Politics Department at Princeton University and received an M.A. from Emory University in 2006.  His research interests include international political economy and international institutions.  His dissertation research examines the ways that international institutions transmit information among member states and legalized dispute settlement.  His dissertation focuses on the World Trade Organization, and he has related work on the International Criminal Court and the severity of civil conflicts.

Helen V. Milner is the B. C. Forbes Professor of Politics and International Affairs at Princeton University and the director of the Niehaus Center for Globalization and Governance at Princeton’s Woodrow Wilson School. She is currently also the chair of the Department of Politics. She has written extensively on issues related to international political economy, the connections between domestic politics and foreign policy, globalization and regionalism, and the relationship between democracy and trade policy.

Dustin Tingley is Assistant Professor of Government at Harvard University. He received my Ph.D. in Politics from Princeton in 2010 and B.A. from the University of Rochester in 2001. His research interests include international relations, international political economy, and experimental approaches to political science. He is currently working on new experimental projects on bargaining, new methods for the statistical analysis of causal mechanisms, and a book about the domestic politics of U.S. foreign policy.

Essay by Brian C. Rathbun, University of Southern California

Was There Ever a Bipartisan Ideological Consensus?

Charles Kupchan and Peter Trubowitz (K&T) have engaged in a provocative and insightful debate with Stephen Chaudoin, Helen Milner and Dustin Tingley (CMT) as to whether a bipartisan consensus on liberal internationalism (LI) persists in American foreign policy. The former lament the increase in ideological polarization in recent years, whereas the former find no evidence of any significant change over the post-World War II period. They do, however, to agree on one essential fact – that there was a bipartisan consensus based on shared ideological beliefs in the early decades following America victory in World War II and the beginnings of the Cold War. But was this actually ever the case? We do not have any firm evidence of this. The elite surveys that document the increasing polarization of American foreign policy were first fielded in the 1970s, just as Vietnam and détente began to divide the country. I argue that this vision of an early postwar consensus on LI that united the two parties is an idealization of the past. American grand strategy was fraught with ideological and partisan conflict even immediately after World War II. Dean Acheson, American Secretary of State in the late 1940s, called bipartisanship the “magnificent fraud.”[48]

Part of the problem in establishing whether there were significant ideological divisions over American foreign policy in the early postwar period, and whether those divisions have worsened over time, is the very fuzzy concept of “internationalism.” The term implies American engagement overseas, but tells us very little about the nature of that engagement. The dominant model for explaining the two-dimensional structure of ideological opinions among American elites, offered by Wittkopf, makes room for two different types of internationalisms that are negatively related to one another, indicating that the concept is broad enough to include contentious if not contradictory viewpoints.[49]

Internationalism’s opposite is isolationism, traditionally the belief that the United States should limit its relations with other countries to economic and trade relations and avoid political and especially military commitments or involvement in the wider world. But isolationism has not been a viable force in American foreign policy since World War I. The primary axis of foreign policy contention during the League of Nations fight was not isolationism vs. internationalism, but rather between internationalists about the particular form that international engagement would take. True isolationist critics in the Senate numbered less than ten, not enough to amount to the margin of victory of “conservative internationalists” over “progressive internationalists” when the League of Nations covenant was considered. The former beat the latter without their help. Even in 1914, isolationists were an anachronistic bunch whose primary concern was staying true to Washington and Monroe’s admonishments to stay clear of European affairs.[50]

LI is a vapid phrase that should be jettisoned for more meaningful terms. The League showed that the real point of contention was the extent to which the United States would make multilateral commitments to international organizations to provide their security, or whether they would choose to go it alone. As Ruggie noted, those multilateral commitments were more meaningful if they asked the United States to pledge in advance to particular actions such as going to the aid of others in case they were attacked or resolving to solve disputes through third party intermediaries in international organizations.[51] President Woodrow Wilson and his Democratic allies wanted a more automatic security guarantee and mandatory conflict resolution procedures in which parties would not vote on, and could not veto consideration of, their own disputes. Republicans insisted on the ability to decide unilaterally, on a case by case basis, whether to use force or economic sanctions to coerce others to preserve peace and would not allow others any institutional role in resolving their disputes.[52]

The conventional wisdom, one that we see implied in both K&T and CMT, is that World War II marked a decisive break with this pattern of partisan ideological conflict in favor of a bipartisan consensus on multilateralism (or as they put it, “liberal internationalism”). The story is that the war convinced isolationists of the error of their ways, that the United States could not sit on the sidelines as it would be drawn into the fray eventually. But as I have argued, there were practically no more isolationists anyway. And the ideological differences between predominantly multilateral Democrats and predominantly unilateral Republicans persisted after World War II. Democratic planners in the State Department envisioned a new United Nations with a security guarantee that was more automatic than that of the League. Even the four great powers would not have a veto over the use of their armed forces to enforce the peace in their early papers. In gatherings during the war, Republicans had difficulty even committing to the creation of an organization, much less one with any type of real commitments to global security.  These ideological differences persisted even during consideration of the North Atlantic Treaty (N.A.T.), when the growing systemic pressure of the Cold War should have pushed the parties closer together. Whereas the Democratic administration was willing to make as binding a security commitment to the Europeans as the Constitution would allow, Republicans were initially opposed to any sort of institutional or treaty commitment at all. Even the moderate Republicans Arthur Vandenberg and John Foster Dulles insisted initially that the arrangement simply be one of standardizing equipment and pooling military supplies. [53]

At this point, one might object. The U.N. Charter and the N.A.T. passed overwhelmingly in the Senate. Certainly this demonstrates bipartisan consensus and ideological convergence over American grand strategy, which distinguished the post-WWII from the post-WWI period. I argue instead that ideological differences were masked by the processes by which the Roosevelt and Truman administrations, in contrast to that of Wilson, went about ensuring ratification. Whereas the latter pushed forward a treaty based on his vision of multilateral collective security and hoped to force it through the Congress on a wave of favorable public opinion over the objections of conservatives, the former cultivated conservatives and moderate Republicans so as to reach a compromise before these treaties reached the floor of the Senate. This was the true lesson of the failure of the Covenant in 1919, one about process rather than substance.

Republican objections centered on restoring American unilateralism. They are responsible for the great power veto and the lack of any automatic security guarantee in the U.N. Charter. In the N.A.T., they also insisted on qualifications to the security guarantee as well as provisions that ensured that the United States would only help others who helped themselves. The roll call votes on these two instruments therefore contain little information about ideological convergence or partisan consensus. They only indicate contentment with a compromise. Only a qualitative analysis of the process by which these foundations of  LI came about can elucidate the extent of American agreement on fundamentals of postwar order. The differences between the parties after World War II were not as great as those after World War I, likely because the early 20th century was another time of great ideological polarization. Yet there were significant and similar points of contestation.

This brings me to important methodological points that have implications for our understanding of the evolution of bipartisan consensus over time. An aggregate analysis of Congressional roll call data can tell us some things but not others. If we extrapolate the methodological lesson of the bipartisan “consensus” on the U.N. Charter and the N.A.T., it tells us that merely knowing how Senators or Congressman vote tells us very little unless we know what they are voting on. Most foreign policy bills will not be contentious at all, partisan polarization or not. This is because foreign policy, particularly security policy, doesn’t have large distributional consequences for the constituencies of American political parties. We judge whether foreign policy is polarized by whether there are major differences on major issues of policy, such as whether to use torture to fight the war on terrorism or whether to negotiate with the Soviet Union. The minutiae of foreign policy, like whether to increase funding for the embassy in Madagascar, might elicit no controversy, and we wouldn’t expect it to. And if it makes up 98% of what Congress deals with, it will overwhelm the things we will really care about in any aggregate quantitative analysis. We will draw false conclusions about ideological consensus as a result.

Aggregate quantitative data can tell us other things, however. CMT separate the final votes on Congressional bills from procedural votes and amendments. I accept their point that procedural votes are highly partisan for reasons that might not have anything to do with ideological polarization and might indicate a substantive division where there is none. The same cannot be said of amendments. These are efforts by various sides to bring a bill closer to their ideal version. We would expect that the number of amendments for any bill would rise over time as parties become (even) more ideologically divided since they are farther apart to begin with. Any bill is more contentious in such a bipolar distribution of legislators. And this is exactly what the authors find. In other words, CMT’s evidence speaks for K&T’s argument. There is a clear analogy to the process I discovered in my cases studies – two sides far apart, changing a bill, and meeting somewhere in the middle with enough support from both sides.[54] The only difference is that the U.N. and N.A.T. process occurred behind-the-scenes and not on the floor of the Congress. This process might indicate bipartisanship, but only in the trivial sense that two parties are both voting for a bill. It does not indicate ideological convergence, the crucial element whose absence K&T lament. I do as well, but let us not over-idealize the past. Even when there was more bipartisanship there was not necessarily a lot of agreement.


Brian Rathbun is an assistant professor of international relations at the University of Southern California. He received his Ph.D. in Political Science from the University of California, Berkeley. He is the author of Partisan Interventions: European Party Politics and Peace Enforcement in the Balkans (Ithaca and London: Cornell University Press, 2004) and Trust in Cooperation: International Security Institutions, Domestic Politics and American Multilateralism (Cambridge: Cambridge University Press, forthcoming).


© Copyright 2011-2015 The Authors.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 United States License.



[1] Charles A. Kupchan and Peter L. Trubowitz, “Dead Center: The Demise of Liberal Internationalism in the United States,” International Security, Vol. 32, No. 2 (Fall 2007),  7–44; Stephen Chaudoin, Helen V. Milner and Dustin Tingley, “The Center Still Holds: The Potential for Liberal Internationalism Survives.” International Security, vol. 35, (Summer 2010), 75-94; Charles A. Kupchan and Peter L. Trubowitz. “The Illusion of Liberal Internationalism’s Revival.” International Security, vol. 35, no. 1 (Summer 2010), 95-109; Joshua Busby and Jonathan Monten, “Without Heirs? Assessing the Decline of Establishment Internationalism in U.S. Foreign Policy,” Perspectives on Politics, vol. 6 (September 2008), 451-472; Joshua Busby and Jonathan Monte, “With Us or Against Us? Public Opinion and Republican Elite Attitudes on U.S. Foreign Policy after the Cold War,” Political Science Quarterly¸ forthcoming.

[2]  Robert  Y. Shapiro and Yaeli Bloch-Elkon, “Ideological Partisanship and American Public Opinion toward Foreign Policy,” in Morton H. Halperin, Jeffrey Laurenti, Peter Rundlet, and Spencer P. Boyer, eds., Power and Superpower: Global Leadership and Exceptionalism in the 21st Century (New York: Century Foundation Press, 2007).; Jack Snyder, Robert Y. Shapiro and Yaeli Bloch-Elkon, “Free Hand Abroad, Divide and Rule at Home,” World Politics 51, no. 1 (January 2009), 155-87.

[3] Pew Research Center. “Growing Number of Americans Say Obama is a Muslim.”, accessed 10 September 2010.

[4] Ilya Somin. “An Inconvenient Truth,”, accessed 13 September 2010.

[5] H-Diplo | ISSF Roundtable on Politics and Scholarship, Vol. 1, No. 2 (June 2010)

[6] Philip E. Tetlock and Gregory Mitchell, “Implicit Bias and Accountability Systems,” Research in Organizational Behavior, vol. 29, (2009), 3-38.

[7] Here I am drawing on Robert Lieber’s remarks at the APSA panel.

[8] American Political Science Association Committee on Political Parties, Toward a More Responsible Two-Party System (Rinehart & Company, 1950).

[9] Steven Chaudoin, Helen Milner, and Dustin Tingley, “The Center Still Holds,” International Security 35, 1 (Summer 2010), 75-94; Charles A. Kupchan and Peter L. Trubowitz, “The Illusion of Liberal Internationalism’s Revival,” International Security 35, 1 (Summer 2010), 95-109; Charles A. Kupchan and Peter L. Trubowitz, “Dead Center: The Demise of Liberal Internationalism in the United States,” International Security 32, 2 (Fall 2007), 7–44.

[10] Joshua Busby and Jonathan Monten, “Without Heirs? Assessing the Decline of Establishment Internationalism in U.S. Foreign Policy,” Perspectives on Politics, 6, 3 (September 2008), 451-472. Though scholars frequently refer to this broad compact of support for coercive and cooperative means of international engagement as “liberal internationalism,” “liberal” has come to have a specific meaning in the political world, obscuring the fact that the coalition that supported internationalism in the wake of World War II included conservative elements. For this reason, we prefer the less laden term “establishment internationalism,” though we will refer to the broad compact as liberal internationalism to be consistent with the two papers we are responding to.

[11] Joshua Busby and Jonathan Monten, “With Us or Against Us? Public Opinion and Republican Elite Attitudes on U.S. Foreign Policy after the Cold War,” Political Science Quarterly, forthcoming.

[12] Although not all polls support even this conclusion, such as a recent Pew survey showing a sharp increase the in the proportion of the public that thinks the US should “go [its] own way” and “mind its own business” since 2002. Pew Research center for the People & the Press, “America’s Place in the World 2009,” December 3, 2009.

[13] Robert Y. Shapiro and Yaeli BlochElkon, “Foreign Policy, Meet the People” National Interest 97 (September/October 2008), 37-42.

[14] Kupchan and Turbowitz, “Dead Center,” 38, Fig.6

[15] More recent polls show diminished but still strong support for the ICC and a climate change agreement. A 2008 poll by the Chicago Council found 68% of the American public supported the Rome Statute creating the ICC. In December 2009, a USA Today/Gallup poll found that 55% of the public still supported an international treaty to reduce greenhouse gases (though with sharp partisan divides). Chicago Council on Foreign Relations, “New Poll Shows Americans Support Major Changes in U.S. Foreign Policy.” September 22, 2008. Gallup, “Americans Favor U.S. Signature on Copenhagen Treaty,” December 15, 2009

[16] Charles A. Kupchan and Peter L. Trubowitz, “Grand Strategy for a Divided America” Foreign Affairs 8 (July/August 2007): 71-84

[17] Jeremy A. Rabkin, The Case for Sovereignty: Why the World Should Welcome American Independence (Washington, DC: AEI Press, 2004); John R. Bolton, “Should We Take Global Governance Seriously?” Chicago Journal of International Law 1 (Fall 2000); Jack Goldsmith, The Terror Presidency: Law and Judgment inside the Bush Administration (New York: Norton, 2007).

[18] Democracy Now!, “John Bolton in His Own Words: Bush’s UN Ambassador Nominee Condemns United Nations,” March 31, 2005.

[19] Jacob Heilbrunn makes a similar argument in a 2010 Foreign Policy piece, “The End of the Establishment,” Foreign Policy, July/August 2010. This prompted vigorous online criticism by a number of Republican foreign policy practitioners including  Peter Feaver, Will Inboden, and Dov Zakheim.

[20] Stephen Chaudoin, Helen V. Milner, and Dustin H. Tingley, “The Center Still Holds: Liberal Internationalism Survives.” International Security 35, no. 1 (2010): 75-94.

[21] Charles Kupchan, and Peter Trubowitz. “Dead Center: The Demise of Liberal Internationalism in the United States,” International Security 32, no. 2 (2007): 7-44; Kupchan and.Trubowitz, “The Illusion of Liberal Internationalism’s Revival,” International Security 35, no. 1 (2010): 95-109.

[22] Joseph Parent and Joseph Bafumi. “Correspondence: Of Polarity and Polarization,” International Security 33 (2008), 170.

[23] Kupchan and Trubowitz, “The Illusion of Liberal Internationalism’s Revival,” 96.

[24] E.g., Francis Fukuyama, The End of History and the Last Man, 1st ed. (New York: Free Press, 1992); G. john Ikenberry, Thomas J. Knock, and Anne-Marie Slaughter. The Crisis of American Foreign Policy: Wilsonianism in the Twenty-First Century (Princeton: Princeton University Press, 2009); Colin Dueck,  Reluctant Crusaders: Power, Culture, and Change in American Grand Strategy (Princeton: Princeton University Press, 2008); Scott Burchill, “Liberalism” in Theories of International Relations, edited by Scott Burchill (New York: Palgrave, 2001).

[25] Kupchan and Trubowitz, “Dead Center,” 8.

[26] Kupchan and Trubowitz, “The Illusion of Liberal Internationalism’s Revival,” 98.

[27] Ibid, 98, Kupchan and Trubowitz, “Dead Center,” 9.

[28] Jonathan Monten, “The Roots of the Bush Doctrine: Power, Nationalism, and Democracy Promotion in U.S. Strategy,” International Security 29, no. 4 (2005): 112-56.

[29] Kupchan and Trubowitz (K&T) suggested that the lack of moderates would undermine LI without offering an explicit mechanism linking changes in the presence of moderates and LI.  We attempted to fill in this missing link by looking at gridlock.

[30] Puzzlingly, K&T’s rebuttal makes arguments about how polarization is not a perfect predictor of shifts in LI, citing the McKinley and Bush II presidencies.  This argument applies to their attempts to link polarization with LI, but not to gridlock.  Gridlock is a consistent measure of whether or not Congress is addressing the foreign policy challenges at a particular time, regardless of political climate.

[31] William G. Howell and Jon C. Pevehouse, While Dangers Gather: Congressional Checks on Presidential War Powers (Princeton: Princeton University Press, 2007).

[32] K&T claim that we do not include amendment votes in our analysis (pg. 102); but they clearly are included.

[33] Nolan McCarty, Keith Poole, and Howard Rosenthal. Polarized America: The Dance of Ideology and Unequal Riches (Cambridge: MIT Press, 2006).

[34] Clifford J, Carrubba, Matthew Gabel, Lacey Murrah, Ryan Clough, Elizabeth Montgomery, and Rebecca Schambach, “Off the Record: Unrecorded Legislative Votes, Selection Bias, and Roll-Call Vote Analysis,” British Journal of Political Science 36, no. 4 (2006): 691-704; Laurel Harbridge, “Bipartisanship in a Polarized Congress,” working paper, Northwestern University (2009); James Fowler, “Connecting the Congress: A Study of Cosponsorship Networks,” Political Analysis 14, no. 4 (2006): 456-87.

[35] To clarify a data disagreement: K&T claim that we do not include amendment votes in our section on cosponsorship (“Illusion,” 102).  This is a non issue, because there are no instances of cosponsorship on amendment bills concerning foreign policy during this time period.

[36] Daniel Kessler and Keith Krehbiel. “Dynamics of Cosponsorship,” American Political Science Review 90, no. 3 (1996): 555-66; Wilson, Rick K., and Cheryl D. Young. “Cosponsorship in the U.S. Congress,” Legislative Studies Quarterly 22, no. 1 (1997): 25-43.

[37] Kupchan and Trubowitz (2010), 102.

[38] Douglas R. Arnold, The Logic of Congressional Action (New Haven: Yale University Press, 1990); Gregory Koger, “Position-Taking and Cosponsorship in the U.S. House,” Legislative Studies Quarterly 28, no. 2 (2003): 225-46.

[39] Koger, “Position-Taking and Cosponsorship,” 231.

[40] K&T’s use of the argument that “cosponsorship has little effect on policy” (102) demonstrates how severely they miss the point here.  Cosponsorship is uncorrelated with the passage of legislation because of the immensely complicated strategic interactions that go on before any bill ever reaches a roll call vote.  This is precisely the reason why studying cosponsorship is valuable, because it is a “roll-call independent” way for legislators to signal their preferences (Harbridge 2009) and avoids the problem that roll call votes are not a random sampling of all bills.

[41] Kupchan and Trubowitz, “The Illusion of Liberal Internationalism’s Revival,” 99; K&T criticize our question about international engagement saying it does not tap liberal internationalism (LI) directly. We agreed, emphasizing that lacking a single question measured over time we opted for a proxy of a necessary condition of LI. Strikingly, K&T went on to use survey questions nearly identical to our own to illustrate their preferred interpretation of contemporary public opinion.  This criticism of our article is surprising to us because in their section on public opinion during the Obama administration, they are comfortable using public opinion responses to the question of whether or not the United States should “mind its own business” which is closer to the question we used in our response.  Neither that question nor any of the other questions K&T analyze capture the “dual nature” of liberal internationalism.  Of all the questions analyzed, ours speaks most directly and clearly to the questions at hand and, due to the temporal nature of their hypotheses, is best suited given its extensive over time coverage.  Note also that this criticism probably applies more K&T’s use of questions about defense spending, which is a much narrower part of LI, than to our question.

[42] Although we would note that it is often the case that Republican and Democratic opinions do not display statistically significant differences in B&M’s analysis.  The fact that their opinions are not more divergent is evidence enough in support of our argument.

[43] For the U.N. question we are concerned with their recoding of response options into a dichotomous category. Our analysis of the same questions which keeps the original response options does not show the same pattern.

[44] Sarah Binder, “The Dynamics of Legislative Gridlock, 1947-96,” American Political Science Review 93, no. 3 (1999): 519-33.

[45] K&T, “The Illusion of Liberal Internationalism’s Revival,” 102.

[46] Scott Shane, Mark Mazzetti, and Robert Worth. “Secret Assault on Terrorism Widens on Two Continents,” New York Times August 15 2010.

[47] LLCool J. “Mama Said Knock You Out,” Mama Said Knock You Out,  Def Jam Records (1990).

[48] Truman Library, Oral History Project.  Online: .

[49] Eugene R. Wittkopf, Faces of Internationalism : Public Opinion and American Foreign Policy ( Durham: Duke University Press, 2004).

[50] See Brain C. Rathbun, Trust in International Cooperation: The Formation of International Security Organizations and the Domestic Politics of American Multilateralism  (Cambridge: Cambridge University Press, forthcoming).

[51] John G. Ruggie, “Multilateralism: the Anatomy of an Institution,” International Organization 46:3 (1992), 561-598.

[52] I tell this story in Rathbun,  “Before Hegemony: Generalized Trust and the Creation and Design of International Security Organizations,” International Organization, forthcoming.

[53] Rathbun, “The ‘Magnificent Fraud: Trust, International Cooperation and the Hidden Domestic Politics of Postwar American Multilateralism.”  International Studies Quarterly, forthcoming.

[54] See footnotes 3, 5, and 6.