Roundtable on War and Democratic Constraint: How the Public Influences Foreign Policy

Occasionally, the long timelines of academia have an upside. Matthew Baum and Philip Potter’s War and Democratic Constraint was published in 2015, and these reviews were set in motion prior to Election Day. But President Donald Trump’s surprise victory has, among other things, refocused attention on the nature—and fragility—of democratic institutions. Although Baum and Potter’s book had both scholarly and policy relevance before 8 November 2016, it has taken on new significance and urgency in the election’s aftermath.

H-Diplo | ISSF Roundtable, Volume IX, No. 18 (2017)

H-Diplo/ISSF Editors:  Thomas Maddux and Diane Labrosse
H-Diplo/ISSF Roundtable and Web/Production Editor:  George Fujii

Matthew A. Baum and Philip B.K. Potter. War and Democratic Constraint:  How the Public Influences Foreign Policy.  Princeton University Press, 2015.  ISBN:  978-0691165233 (paperback, $28.92).

Published on 26 June 2017



© Copyright 2017 The Authors

Introduction by Elizabeth N. Saunders, George Washington University

Occasionally, the long timelines of academia have an upside. Matthew Baum and Philip Potter’s War and Democratic Constraint was published in 2015, and these reviews were set in motion prior to Election Day. But President Donald Trump’s surprise victory has, among other things, refocused attention on the nature—and fragility—of democratic institutions. Although Baum and Potter’s book had both scholarly and policy relevance before 8 November 2016, it has taken on new significance and urgency in the election’s aftermath.

Baum and Potter’s argument, as ably summarized by Sarah Croco, is that “some democratic governments act more in accordance with their citizens’ preferences than others on matters of foreign policy” because they have “a robust opposition comprised of a large number of political parties and a domestic media that is both accessible to citizens and diverse in the views it represents.” The presence of both effective, diverse opposition and a strong and accessible media environment is necessary for democratic publics to constrain their leaders. Apart from the book’s scholarly contributions, in the wake of Trump’s election and concerns about democratic “backsliding” in the United States, Croco’s point that Baum and Potter “[clarify] what we really mean when we say things like ‘opposition’ and ‘free press’” is especially notable. War and Democratic Constraint puts two essential democratic institutions and their effects on foreign policy in wider context even within the family of “democracies.”

In more purely scholarly terms, the reviewers in this roundtable highlight several contributions. Croco praises the book for showing “unequivocally that democracies cannot be treated as a homogenous group.” She also emphasizes Baum and Potter’s extensive data-gathering effort and “herculean” analysis of news sources in the empirical portions of the book. Christopher Gelpi notes that Baum and Potter are “ambitious on a variety of fronts,” both theoretical and empirical, and that the book delivers on most of its promises. “Any subsequent work on the subject,” he writes, “will need to grapple with the parsimonious framework and persuasive empirical evidence that Baum and Potter set forth.” Thomas Scotto calls War and Democratic Constraint “an impressive and eye-opening attempt to examine how cross-national differences in institutions shape the relevance of public attitudes towards war and conflict when it comes to constraining elite behaviour.” He notes that “what is so exciting about this work to a scholar of public opinion” is that it “looks at when leaders are most likely to exercise constraint in response to public attitudes concerning conflict.”

Although the reviewers are generally united in their admiration for the book, they also raise several quibbles, omissions, and directions for future research. Croco wishes that Baum and Potter had done more to study what she calls “back end” constraint, or the “public’s ability to constrain a war already in progress.” Gelpi offers two “corollaries:” first, that the public can gain information from media reports about wartime casualties even in the absence of elite partisan conflict; and second, that a partisan media environment may erode the mechanisms of media constraint the book proposes. Scotto suggests several avenues for future research, and notes that “an implicit frame in the text is that the public exists to constrain the impulse of elites to rush into military action,” but that the public itself is sometimes the “bloodthirsty” actor.

I will only add one additional comment on a portion of the book not emphasized in these reviews that again takes on new, post-election urgency. In light of the potential upending of American foreign policy in the Trump era, many have pointed to Chancellor Angela Merkel of Germany as the new keeper of the flame of world order. But Germans will go to the polls in the fall of 2017. Chapter 7 of Baum and Potter’s book provides fascinating insights into the nature of German media coverage in a multiparty environment. While undoubtedly more familiar to European readers, this chapter will be of interest to scholars who are less well-versed in the details of German politics, at a time when the domestic politics of German foreign policy will be increasingly relevant.

This roundtable review suggests that the book Baum and Potter have produced will be a staple of the study of democracy and foreign policy for years to come. Perhaps Gelpi’s words summarize it best: they have written a book many of us probably “wish [we] had written.”


Matthew A. Baum (Ph.D., UC San Diego, 2000) is the Marvin Kalb Professor of Global Communications and Professor of Public Policy at Harvard University’s John F. Kennedy School of Government and Department of Government. His research has appeared in over a dozen leading scholarly journals, such as the American Political Science Review, American Journal of Political Science, and the Journal of Politics. His books include Soft News Goes to War: Public Opinion and American Foreign Policy in the New Media Age (Princeton University Press, 2003), War Stories: The Causes and Consequences of Public Views of War (Princeton University Press, 2009, co-authored with Tim Groeling), and War and Democratic Constraint: How the Public Influences Foreign Policy (Princeton University Press, 2015, co-authored with Phil Potter). He has also contributed op-ed articles to a variety of newspapers, magazines, and blog sites in the United States and abroad. Before coming to Harvard, Baum was an associate professor of political science and communication studies at UCLA.

Philip Potter is Associate Professor of Politics and Public Policy at the University of Virginia. He is the principal investigator for a Minerva Initiative project to map and analyze collaborative relationships among terrorist organizations. His recent work has appeared in International Organization, Journal of Politics, International Studies Quarterly, Presidential Studies Quarterly, and The Journal of Conflict Resolution. He is currently completing a volume entitled Elections and American Foreign Policy. He received his Ph.D. from UCLA and his BA from McGill University.

Elizabeth N. Saunders is an Associate Professor of Political Science and International Affairs at George Washington University. Her research and teaching interests focus on international security and U.S. foreign policy. Her book, Leaders at War: How Presidents Shape Military Interventions, was published in 2011 by Cornell University Press and won the 2012 Jervis-Schroeder Best Book Award from APSA’s International History and Politics section. Her research has also been published in journals including International Organization, International Security, International Studies Quarterly, the American Journal of Political Science, and Security Studies.

Sarah Croco is an assistant professor in the Department of Government and Politics and a faculty associate at the Center for International Development and Conflict Management at the University of Maryland. She received her Ph.D. in political science from the University of Michigan in the spring of 2008. Her research interests include: international conflict; the process by which citizens assign leaders responsibility for international wars; the value of policy consistency in elections; territorial disputes and civilian targeting. Her work has appeared in The American Political Science Review, The American Journal of Political Science, The Journal of Politics, International Studies Quarterly, and World Politics, among others.

Christopher F. Gelpi (Ph.D., University of Michigan, 1994) is Chair of Peace Studies and Conflict Resolution at the Mershon Center for International Security and Professor of Political Science at The Ohio State University. His primary research interests are the sources of international militarized conflict and strategies for international conflict resolution. He is currently engaged in research on American public opinion and the use of military force, and on statistical models for forecasting military conflict and transnational terrorist violence. He has also published works on American civil-military relations and the use of force, the impact of democracy and trade on international conflict, the role of norms in crisis bargaining, alliances as instruments of control, diversionary wars, deterrence theory, and the influence of the international system on the outbreak of violence. He is author of The Power of Legitimacy: The Role of Norms in Crisis Bargaining (Princeton University Press, 2002), co-author (with Peter D. Feaver) of Choosing Your Battles: American Civil-Military Relations and the Use of Force (Princeton University Press, 2004). and co-author (with Peter Feaver and Jason Reifler) of Paying the Human Costs of War: American Public Opinion and Casualties in Military Conflicts (Princeton University Press, 2009).

Thomas J. Scotto is a Professor in the School of Government and Public Policy at the University of Strathclyde. His research on comparative foreign policy attitudes of the publics in established democracies is funded by the Economic and Social Research Council (ESRC) of the United Kingdom. His articles on the topic appear in outlets such as International Studies Quarterly, Political Behaviour, International Relations of the Asia-Pacific, the British Journal of Politics and International Relations, and Canadian Foreign Policy.


Review by Sarah E. Croco, University of Maryland

In writing War and Democratic Constraint: How the Public Influences Foreign Policy, Matthew Baum and Philip Potter have made an important and most welcome contribution to the literature on public opinion and conflict. Their book tackles the interesting puzzle of why some democratic governments act more in accordance with their citizens’ preferences than others on matters of foreign policy. This is certainly a question worth posing as we have seen a remarkable degree of variation in this regard within democracies in recent international crises, most notably the war in Iraq in 2003. Baum and Potter’s answer to this question centers on the interaction of two factors: a robust opposition comprised of a large number of political parties and a domestic media that is both accessible to citizens and diverse in the views it represents. In democracies where these two factors are present, leaders will exercise caution before engaging in international conflict. They do so because they know that any missteps will be highlighted by the opposition and easily consumed by the domestic audience via the media. In polities where one of these is absent, leaders have more leeway due to either the lack of a formidable opposition to serve as a whistleblower or a public that is less likely to be well informed of political events.

Baum and Potter demonstrate their theory’s considerable explanatory power by examining three conflict-related behaviors: initiation, reciprocation, and coalition formation. Their proxy for the presence of a meaningful opposition is the effective number of parliamentary parties (ENPP), with a larger number of parties having a larger constraining effect. They capture media access (among other ways) with the number of televisions per 1,000 people. Across all three outcomes, they find that the combination of a high number of political parties and media access has a sizeable effect on leader behavior. Democratic leaders who operate in high-party, media-rich environments are less likely to use military force; more likely to appear credible during disputes, thereby lowering the probability their opponent will reciprocate; and less likely to join wartime coalitions.

While most authors would be content having marshaled such an impressive amount of empirical evidence, Baum and Potter do not stop here. Instead, using a truly extraordinary array of news sources and applying a herculean coding effort, they demonstrate that the news coverage in states with a large number of political parties “offered more policy-oriented news, more criticism of government policy and more diverse coverage” (164) than states with fewer parties. They round out the book with a set of excellent, convincing case studies of four states (the UK, Spain, Poland, and Germany) and their decisions as to whether to join the ‘Coalition of the Willing’ in the war in Iraq. These states represent the four possible combinations of political opposition and media access while allowing the authors to hold constant other key attributes (e.g., the war in question, region, etc.)

While reading the book, two contributions struck me as especially important. First, Baum and Potter demonstrate unequivocally that democracies cannot be treated as a homogenous group. The link between citizen preferences and government policy will be much stronger in high-party systems, causing leaders to behave in a more constrained fashion with regards to risky foreign intervention. Leaders in two-party systems, by contrast, will be hindered far less by public opinion, and more able to influence the tenor of the media’s coverage of their decisions. As their book makes clear, this distinction is crucial since public opinion will have different effects in high and low party states. Appreciating this point is also important for theoretical clarity. Much of the literature assumes that democracies function like high-party states—that is, leaders care about and respond to domestic opinion. At the same time, many theories also center on, or are inspired by, states that have a small number of parties, most commonly the U.S. and UK. Assuming these states will behave like ‘true’ (i.e., high-party) democracies overlooks several important aspects of low-party systems, most notably the relative lack of influence held by the pubic, the paucity of the check provided by the opposition, and the diminished ability of the press to provide meaningful oversight. This work has major implications for scholarship that examines the legislature’s role in checking the executive in the realm of foreign policy.[1]

This highlights Baum and Potter’s second contribution, clarifying what we really mean when say things like ‘opposition’ and ‘free press.’ All too often these terms are oversimplified, and defined vis-à-vis nondemocratic states. With autocratic states as a presumed baseline, so long as another party has some say in governance and the press is not controlled outright by the state, the opposition and free press criteria are assumed to be met. Baum and Potter demonstrate that the mere existence of these institutions does not guarantee that their effect will be the same across all democracies. For example, a small number of parties will encourage convergence on the median position, while a larger number of parties will lead to more diversity. As noted above, this has cascading effects for the tone of the domestic press. The nature of the press coverage, in turn, reinforces the ability for the public to constrain the leader in high-party systems and the leader’s ability to shape the agenda in low-party states.

Areas for Future Research

My sole criticism of the book is that it does not go far enough in exploring the implications of the theory. The authors focus on what I will term ‘front end constraint’ instead of examining how the number of parties and media access affect the public’s ability to constrain a leader once a war has already started. While there is only so much one can do in a single book, not addressing constraint on the ‘back end’ has at least two drawbacks. First, when many scholars discuss democratic constraint, they are referring to the public’s ability to stop a war already in progress.[2] Baum and Potter give some attention to this in their case discussions, but do not do a systematic exploration of the effect of their variables on war duration or outcome, which struck me as a missed opportunity. While we know democratic leaders are selective in the wars they choose to fight,[3] they often get it wrong and wind up in quagmires.

Second, looking at constraint on a war in progress would provide a much harder test of their theory because of the dynamics that emerge once a war is underway. We know from earlier work that leaders strategically pick conflicts they know will receive broad legislative support.[4] This logic should hold for both low and high-party states; the leader may have to cobble together a more diverse coalition in the latter to achieve a majority, but will do so nonetheless. The existence of this coalition of support is important because it means that there will be a large number of legislators who will share culpability for the conflict with the leader. As I have argued in my own work,[5] being culpable for a foreign policy decision imposes its own type of constraint on executives and legislators. Executives will continue fighting unpopular wars, even if they know they are unlikely to win. Likewise, legislators will continue to support the war, even if their constituents are calling for its end. This intransigence emerges since admitting a mistake on an earlier vote of support has negative political consequences. It makes the leader look incompetent and provides ammunition for challengers. Thus, perhaps somewhat counter-intuitively, the best course of action is to continue supporting the war. Critically, this logic holds for all legislators who voted for the war, even those outside the leader’s party. This effectively reduces the number of effective parties by pooling all of the legislators who voted for the war.

It also stymies the ability of any legislators outside of the culpable majority to exercise any constraint. So long as the culpable leader remains in power, it is in her interest to continue fighting the war until she wins or total defeat is forced upon her. It will also take a considerable amount of time for the culpable majority to ‘erode’ out of the legislature. Even if elections are scheduled soon after the conflict becomes unpopular, incumbency advantage will make it hard for nonculpable legislators to get elected. In parliamentary systems, of course, elections can, theoretically, be called at any time, but the culpable majority will want to prevent this from happening. Calling for a vote of no confidence could bring the prime minister down but also introduces a great deal of risk for legislators who supported the war.

Put differently, regardless of the number of parties and the availability of media, the political ‘transmission belt’ in democracies turns very slowly. There is a great deal of inertia in the system, which is reinforced by the pressures of culpability that arise when a state is participating in an international conflict. Cases like Spain, where political change was swift in the wake of mass public outcry, are noteworthy, but I am not sure they represent the norm. In many cases, it takes years for meaningful change to happen. I am happy to be corrected on this point, but, from a theoretical standpoint, it is not clear to me how the mechanisms proposed by the authors could overcome the issues created by culpability. It may be easier for countries with a large number of parties to gain electoral traction and wear down the culpable bloc more quickly. But, if Kenneth Schultz is correct about leaders behaving strategically and amassing broad support in the conflict selection stage (and I think he is), this should be a tall order once the war is underway. This is in addition to the political pressures that arise once resources have been deployed, such as need to appear supportive of the troops and confident in the nation’s chances of victory, all of which encourage unity rather than dissent. Applying this theory to the latter stages of conflict strikes me as both a useful and a fruitful endeavor. Given the light it shed on three important ‘front end’ outcomes, I have no doubt that it will help us understand behavior within democracies further into a war.


Review by Christopher Gelpi, Ohio State University

Perhaps the clearest way to express my overall evaluation of War and Democratic Constraint is to say that it is one of those books that I wish I had written. In this important work, Matthew Baum and Philip Potter are ambitious on a variety of fronts. First, they are theoretically ambitious in their attempt to provide a coherent conceptual foundation for the ‘democratic peace,’ which has often been characterized as an observed effect in search of a cause. Second, they double down on their theoretical ambition by providing a framework that is both elegant, and parsimonious. And finally, Baum and Potter are empirically ambitious in their claim that a single interaction effect can explain state decisions to initiate, reciprocate, and collaborate with regard to the use of force.

Ambition is admirable, but perhaps even more impressive is the fact that Baum and Potter deliver on most–if not quite all–of what they promise. This book emerges from a recent trend toward disaggregating our understanding of ‘democracy’ and ‘autocracy’ as they relate to foreign policy. Recent work on the varieties of autocracy from scholars such as Barbara Geddes and Jessica Weeks has been very successful in revealing nuanced patterns of conflict behavior among autocratic states.[6] The literature on the varieties of democracy, however, has lagged somewhat behind on this count. While democracies clearly vary on a number of potentially significant dimensions – including institutional structure and levels of public engagement–scholars have not synthesized these variations into a coherent unifying framework that can explain variations in conflict behavior.

Baum and Potter provide this theoretical framework by building on the growing literature on public opinion and foreign policy. Much of this work begins from the incontrovertible observation that members of the public obtain their information about foreign policy issues through the mass media because they have little or no direct exposure to foreign events. At the same time, a substantial literature in Communication also suggests that media organizations have strong structural incentives to echo elite messages rather than providing a truly independent source of information.[7] Bringing these two well-established patterns together leads Baum and Potter to the intuitive and yet surprisingly novel expectation that the public will only constrain the foreign policy behavior of democratic states when media penetration of the society is high and when elites present the media with contrasting narratives about foreign policy.

Much of my own work over the past decade or so has emphasized the ability of the public to collect foreign policy information from the media even in the absence of elite partisan conflict.[8] (Thus while I do not question Baum and Potter’s assertion that partisan disagreement will increase public constraint on foreign policy, I might suggest a corollary argument that provides–in my view–some added nuance to their theory.

Specifically, I would argue that–conditional on a free press-organizations within each state will always consider military casualties suffered by that state to be newsworthy. As a result, individuals will be able to gather information about military engagements that involve casualties even in the absence of elite partisan disagreement over the conflict. Consequently, the public will continue to update its support for wars involving casualties even if the effective number of parties is low. Once again, this is not to say that elite disagreement does not shape public attitudes at all, but rather to say that the public can update its views even in the absence of partisan conflict. However, democratically elected leaders will not feel constrained to alter their policies regarding these costly conflicts unless they face opposition from competing elites who seek to exploit public opposition to the conflict. Thus I would argue that the effective number of parties may be important partly because it shapes the public’s ability to gain access to information, but also – and perhaps even more fundamentally – because it pressures elected leaders to change their policies when the public disapproves.

The evolution of American public attitudes toward the war in Afghanistan provides a perfect illustration of this distinction. At the outset, American public support for the war in Afghanistan was astronomically high in the wake of the 9/11 attacks in 2001. As success remained elusive–especially following the failure of President Barack Obama’s 2009 to 2012 “surge”–and as casualties began to mount, public support for the war eroded. Thus in the March of 2009 an ABC News/Washington Post poll found that 56% of Americans believed the war in Afghanistan was worth the costs, but by April 2012 that number had dropped to 30%. This dramatic shift in public support for the war occurred despite a virtually complete lack of elite partisan discord over the issue. Nonetheless, unlike Iraq in 2004 and 2008, Afghanistan played little role in the 2012 election because Republican candidate Mitt Romney chose not to challenge Obama on the issue. As a result, American policy was unaffected by the widespread disaffection with the war. Thus I would argue that partisan discord does not shape public attitudes toward war as much as it conditions the impact of those attitudes on conflict behavior.

Regardless of whether one accepts the unalloyed Baum and Potter framework or the ‘Gelpi corollary,’ however, one should expect the interaction of public access to foreign policy information through a free press to work in combination with elite partisan discord to constrain the foreign conflict behavior of democratic states. This represents the core hypothesis that Baum and Potter test through and analysis of militarized interstate dispute (MID) initiation, and reciprocation, as well as state decisions to join the coalitions of the willing in Iraq and Afghanistan. The authors generate impressively consistent support for this expectation across these three very distinct conflict behaviors. And as a result, I think they set a new standard for theoretical explanations of the democratic peace. While their work will surely not be the last attempt to explain the impact of democratic constraint on foreign policy, any subsequent work on the subject will need to grapple with the parsimonious framework and persuasive empirical evidence that Baum and Potter set forth.

As with any work of substance and ambition, I inevitably have quibbles and concerns with some of the conclusions that they draw from their data. I shall focus here on three reservations that I think merit further attention. First, Baum and Potter’s attempt to link their analysis of MID reciprocation to the debate on audience costs struck me as somewhat strained and less than entirely persuasive. The authors acknowledge that their analysis of reciprocation is a step removed from audience costs, since they do not actually measure the coercive impact of audience costs on crisis bargaining behavior. But I think that their analysis is at least two steps removed from the audience-costs debate because Baum and Potter do not actually demonstrate the existence of any costs to MID reciprocation. The core claim of the audience-costs argument is that coercive threats will be more effective when opponents know that leaders will be punished for failing to make good on them. But Baum and Potter’s analysis of dispute reciprocation does not demonstrate either the effectiveness of threats made by constrained leaders or the punishment of such leaders for not making good on their threats. Instead, Baum and Potter show that democratically constrained leaders choose to avoid military conflict in situations where they might expect to incur audience costs if they were to escalate and fail. The existence of audience costs is one possible explanation for this pattern of conflict avoidance, but there are surely others as well. Thus the MID reciprocation analysis persuaded me of the existence of democratic constraint, but not of the existence of audience costs.

Second, a quick review of the estimated empirical effects in chapters three, four, and five reveals that much of the difference between the impact of media access on conflict behavior in high and low-party systems has to do with the bellicose impact of media access in low-party systems. That is, while Baum and Potter’s theoretical framework and much of the book are focused on the impact of the media on democratic constraint depending on the effective number of parties, some of the data seem to point to the media as a spur rather than a constraint in low-party systems. The authors acknowledge the possibility of conflict amplification in passing in their theory chapter, but their argument is really about levels of constraint. I found myself wishing that they had spent a bit more time talking about the ‘rally round the flag’ effect and the media ‘spiral of silence.’

Finally, I was a bit dissatisfied with Baum and Potter’s discussion of the recent changes in the media environment and their potential impact on the findings in the book. The authors rightly dismiss concerns that the rise of internet media will undermine attention to national partisan elites both because TV remains an overwhelmingly dominant media source and because even those individuals who do rely primarily on the internet still generally focus on national rather than international sources. But the authors are rebutting the wrong critique. The most important change in the media environment is not the rise of the Internet, but the rise of partisan media across various platforms including print, radio, television, and the Internet.

The core of Baum and Potter’s argument is that the combination of popular media access and partisan conflict will lead to the transmission of diverse information about foreign policy conflicts to members of the public. This argument assumes that the media will report the views of all elites, and that members of the public are exposed to these views. But the rise of global partisan media networks such as Rupert Murdoch’s News Corp appears to problematize this assumption. The availability of partisan media sources opens up the possibility that citizens will selectively expose themselves to news that reinforces their existing beliefs. In such a partisan media environment increased partisan political conflict and high levels of media access may not result in the public having access to more diverse information. Thus it may not be an accident that the home of Murdoch’s empire–Australia–stood out as one of the stronger supporters of the Iraq War despite its widespread media access and its relatively higher number of effective political parties than the U.S. or the UK.

The rise (or perhaps resurrection) of partisan media leads me to a second corollary hypothesis to the Baum and Potter model. The ideological diversity of political parties may increase democratic constraint on foreign policy, but the ideological polarization of media outlets will undermine this constraint. Thus while the combination of a large effective number of parties and widespread access to a free press will constrain democratic leaders, this constraint will be reduced as the ideological polarization of the mass media increases.

I recognize that testing this hypothesis may be challenging, given that the resurgence of partisan media is a relatively recent phenomenon, but examining the presence of News Corp in each of the media markets that Baum and Potter analyze might be a start. More broadly, the partisan nature of the current media environment harkens back to the days before television when newspapers were the primary sources of information. At that time partisan media were the rule rather than the exception. Perhaps Baum and Potter could in future works extend their analyses backward to include access to newspapers and gain a broader perspective on the impact of media self-selection on democratic constraint.

The potential implications my second corollary are substantial, since it implies that the democratic peace itself may be a function of a temporary and unusual structure of the media environment imposed by broadcast television (as opposed to newspapers, radio, cable TV, or the Internet which all provide relatively easier access for diverse partisan media). Regardless of whether my speculations are substantiated, however, Baum and Potter have provided us with a sweeping, insightful, and exciting new perspective on the role of democracy in foreign policy. Their work is certain to become the new standard for judging and evaluating future research in this area.


Review by Thomas J. Scotto, University of Strathclyde

The study of public opinion on matters of foreign policy and international affairs is at an important juncture. Students of the subfields of public opinion and international relations still too frequently work at arm’s length to one another. To oversimplify: the field of public opinion spent much of the last quarter of the twentieth century convincing both itself and its audience that average voters (often American only) were, in fact, capable of thinking coherently about the often abstract events that define a nation’s conduct with its allies and adversaries.[9]

Scholars of international relations focused mostly on state or state-to-state interactions as the appropriate unit of analysis, and they developed often elaborate formal and empirical tests as to why representative democracies behave differently when it comes to, for example, conflict initiation or termination.[10] The specific mechanisms via which electorates have the capacity to constrain leaders from going to war, however, remains a topic very much in need of intellectual scrutiny.

Matthew Baum and Philip Potter’s work is an impressive and eye-opening attempt to examine how cross-national differences in institutions shape the relevance of public attitudes towards war and conflict when it comes to constraining elite behaviour. What is so exciting about this work to a scholar of public opinion is that it moves the field beyond studies purporting to show that coherent foreign policy attitudes are present in single or a few electorates[11] and correlate with vote choice or affect towards parties and leaders. Instead, the book looks at when leaders are most likely to exercise constraint in response to public attitudes concerning conflict.

At its heart, the book offers a simple conjecture—there are two conditions that must be met in order for public opinion to feature heavily in the decisions of elected officials. First, there must be robust opposition to an elected government, operationalised by the number of effective parliamentary or elected parties. The presence of multiple parties in opposition or coalition raises the probability that one or many will act as ‘whistle blowers,’ essentially calling time on leaders who have an unhealthy yearning to take their nations into war. A large number of parties, however, is a necessary but not sufficient condition for constraint—to be effective, opposition parties need to articulate their case to the public and this requires citizen access to a free press. The authors mostly operationalise access via television and radio saturation, but, as is the case throughout the text, extraordinary care is made to undertake robustness checks that employ alternative operationalisations of variables key to testing the authors’ own and rival hypotheses.

After outlining in the second chapter coherent theoretical arguments, the heart of the book presents empirical models interrogating the role the party number-media nexus has on the decisions of democratic states to initiate conflicts, reciprocate in response to hostile actions and join ‘coalitions of the willing.’ The second half of the book takes on a more qualitative content analysis and process-tracing tone, examining variations in media coverage of specific conflicts by the type of party system and the process via which four nations (United Kingdom, Germany, Poland, and Spain) decided for or against joining the 2003 U.S. led intervention in Iraq. The latter two chapters add rich support to the book’s main conjecture and make the book an exemplar of mixed method research.

The empirical findings that stand out are as follows: in the presence of multipartyism and widespread media saturation, conflict initiation is much less likely but, to underscore the necessary but not sufficient nature of the book’s main hypothesis, media saturation in the presence of a two party system makes conflict initiation more likely. Faced with only a single opposition party, governments become confident in their ability to control the message and media saturation reinforces the drumbeat to war.

In terms of conflict reciprocation, the authors add insights into the audience-costs debate by demonstrating empirically that the propensity of established representative democracies to ‘return fire,’ via (mostly) violent reciprocation also is a function of party systems and media access. Governments not facing a diverse set of opposition parties who have the ability to convey their critiques of military action to publics via the press apparently do not have credibility when it comes to issuing threats. This is why we observe more failures of bargaining and compromise (wars) when media access and the number of parties are low. Elected officials who have to face strong opposition in states with a robust free media presence likely issue threats sparingly and this makes conflict reciprocation low. (Although the empirics of this argument are sound, future work on the topic may want to push the mechanisms behind this argument further). Critics of the audience-costs argument sometimes question the credibility behind the mechanisms, citing evidence that authoritarians who initiate conflicts are completely unaware of or consider irrelevant the attitudes of domestic publics in target states.[12] The authors now ask elite authoritarians to go a step further—not just consider electorate reactions to state-on-state conflict in representative democracies but to understand that institutional contexts make the process work differently. A logical extension of this work might be an attempt to explain the mechanisms via which leaders react to democracies according to variation on the two key constructs of interest.

Chapters five and seven present to the reader an in depth examination of why states do or do not acquiesce to what essentially is a request from the United States to join in foreign policy adventures abroad. Interesting empirical results emerge from Chapter 5 where the number of troops per capita states committed to the 2003 invasion in Iraq is run against the key conjecture—again, we see that nations with high media saturation but a small number of opposition parties were likely to commit the most troops, and the interaction between multipartyism and media access is what made support for the coalition tepid. Afghanistan is perhaps an outlier of a conflict, due to (initially) high levels of public support and NATO’s involvement—therefore, we witness multi-party states contributing more troops per capita than nations with a single opposition parties. This raises an interesting question that future research might wish to explore—an implicit frame in the text is that the public exists to constrain the impulse of elites to rush into military action. In many nations, support for ‘doing something’ in response to the 11 September 2001 attacks against the United States was high, particularly in Western Europe. How might this model work in an environment where the public and press were partial to favour conflict with elites attempting to curtail a bloodthirsty public—or does the emergence of new media covered in many models in the robustness checks make such an environment less and less likely?

This is an exceptional work that lends itself to numerous extensions—for example, how might the models discussed perform when other foreign policy actions of leaders are put under the microscope? Questions of analysing public support for foreign aid and assistance to refugees are topics that cry out for additional analysis. Might the book’s key conjecture be tested with a focus on emerging democratic states in more conflict-prone areas of the world—a deeper analysis of, say, conflicts in sub-Saharan Africa or conflicts where non-state actors play the key role—be a point of departure?

The robustness of the findings in support of the party system-media saturation interconnection raise questions for we who attempt to build on more conventional areas of research into the coherence of public attitudes on matters of foreign policy. As discussed in the opening, many of us have spent time examining the structure of citizens’ attitudes on a set of foreign policy issues—the propensity to strengthen and utilize a nation’s military being one of the most prominent. Given the findings of this work, is there any room to consider these latent attitudes that form the middle of a hierarchy of foreign policy beliefs that Jon Hurwitz and Mark Peffley label postures?[13] In short, is there any role left for the starting beliefs people have about the desirability of military readiness and conflict initiation/reciprocation in determining how publics may respond to conflicts even before leaders, opposition officials and the media stake their positions? Those of us who study public opinion on matters of international relations and who have not given much thought to institutions now have to take even more seriously these questions because of this outstanding book.


Author’s Response by Matthew A. Baum, Harvard University, and Philip B.K. Potter, University of Virginia

We thank the editors and reviewers for their careful attention to our work. Writing a book can seem thankless in the moment, but receiving detailed feedback from such an impressive group of commenters helps to make the effort worthwhile.

Our goal for this project was to reset the agenda when it comes to the study of democracy and conflict. We believed that, despite the huge volume of work in this area, we as a community of scholars had failed to take institutional variation among democracies seriously enough in our theories of war and peace. In our view, assuming that all democracies are alike masked the mechanisms that underpin important empirical observations like the democratic peace.

The question is, what dimension of difference would influence the mechanisms that we think drive the ‘democratic difference’ in conflict behavior? We believe that the answer lies in the role of information, and specifically the ways in which its presence or absence either contributes to or undermines constraints on leaders. We point to the ways in which democratic institutions–particularly electoral and media institutions–influence this flow of information within democracies. The core insight is that those with extensive penetration by a free media and robust political opposition (empirically, having many parties) tend to constrain their leaders more than those without these attributes. The effects are felt in conflict initiation, reciprocation, and coalition formation.

The participants in the forum have identified a number of interesting wrinkles in that argument, exceptions to the rule, and opportunities for extension. We greatly appreciate the insights and the opportunity to engage with them here.

Sarah Croco makes an excellent point about the relationship between the constraints on leaders prior to the initiation of a conflict and those that arise once a conflict is underway. Drawing out the implications of that logic, she points out that we could also think about the extent to which the mechanisms of constraint that we identify could influence the likelihood that leaders back out of ongoing unpopular conflicts and the outcomes of conflicts more generally.

The point is well taken. Casual references to public limits on foreign policy indeed often refer to stopping a war already underway—the most prominent instance being the impact of anti-Vietnam protests. While we do not seriously address this phase of conflict in our book–we focused on the ‘front end’–our instinct is that Croco is probably right that the mechanisms we identify should also play an important role in the dynamics of ongoing conflict.

While a conclusive answer must be reserved for future work, we anticipate that the effect of constraint on ongoing conflict would manifest in complicated ways. This is because there are two distinct but conditional factors to consider: 1) the way information asymmetries change over the course of a conflict, and 2) the selection mechanisms that shape which conflicts these states engage in to begin with and, by extension, how committed the leaders are to them.

Starting with the first of these factors, in our view, the information gap between leaders and the public naturally closes as conflicts drag on (for an illustration see Chapter 2, page 38). Leaders have a natural informational advantage at the outset of hostilities, but this erodes as conflicts unfold. Independent voices emerge, casualties mount, elites defect, and rallies fade. Through this process constraints eventually emerge for even relatively unencumbered democratic leaders. All else being equal, these leaders might then be more likely to withdraw from unpopular conflicts.

Returning to the prior example, the Vietnam War plausibly illustrates this process. Over time, public disapproval led to mounting pressures on a previously less constrained executive branch.

It seems reasonable to anticipate that this convergence happens more quickly in countries with the intuitional arrangements that we argue are conducive to an informed electorate—many parties and high access to free media. Again, Vietnam may illustrate this trend. Constraint emerged, but it took quite a while to do so.

The second factor, however, could plausibly cut the other way. The attributes associated with high constraint may cause a leader’s information advantage to erode more quickly, but this does not necessarily mean that the leader will quickly back down. First, publics in these systems may not turn against a conflict even if they know more because leaders carefully selected these conflicts, aware that they would quickly face scrutiny. Publics in these systems are therefore likely to be more patient with leaders they trust to have selected well. Even if the public does turn against a conflict, then the culpability point that Croco brings up comes into play. It is likely that the diverse coalition of leaders in a high party state who are culpable for a conflict will stick it out even in the face of a downturn in public support in the hopes that winning will lead to a rebound in public opinion.

Croco’s insight is important because, if true, it helps to tie our mechanism both to her work and to the Dan Reiter and Allan Stam logic that democracies may be slow to anger but are tenacious once engaged.[14] Croco points out that leaders often end up in quagmires, but we also know from Reiter and Stam, among others, that they tend to win more. Our expectation is that constrained leaders are actually much less likely than their unconstrained counterparts to end up in quagmires because they are careful on the front end,[15] and because when they do fight they have sufficient support behind the conflict that they will dedicate the necessary resources to see it through even in the face of casualties and other setbacks. Our suspicion is that, as for the democratic peace and audience cost arguments, the constrained types are driving the overall findings for democracies. If this is right, and the institutional mechanisms we identify can drive this by increasing both selectivity and tenacity, this adds yet another consistent piece to our story about democratic differences in foreign policy.

The insight would also generate some specific and testable empirical implications that would be worth exploring. One, in line with the Reiter and Stam model, is that constraint once conflict is underway might not show up so much in the ability of citizens to pull leaders out of unpopular wars as in the prospects for ultimate victory. Because the conflicts that these leaders enter are going to be more carefully selected, they will be those that the public initially approved of and, in many cases, as Croco points out, will involve large and carefully constructed coalitions of parties on the record in support of them. High party/high access states are therefore going to be more tenacious (meaning they might fight longer conflicts, or persist longer in the face of casualties, and win more. Low party/low access states are going to back out of conflicts more often, be more casualty averse, and lose more often. Interestingly, the story that Croco tells leads to off diagonal predictions that are less clearly examined in the book. Specifically, her coalition story suggests that high party/low access states may exhibit more tenacity than low party/high access states.

Elements of Christopher Gelpi’s comments dovetail nicely with those made by Croco. To begin with, he raises the interesting conjecture, based on his prior research, that the press, and by extension citizens, will always be able to gather sufficient information about military engagements if there are casualties even absent elite partisan disagreement over the conflict.[16] Given the centrality of casualties in the longstanding literature on public opinion and conflict, this is a possibility with which scholars should grapple. We did not do so in the book in part because any story involving casualties is necessarily about constraint on conflict already underway. On other words, this is primarily about “back–end constraint” (to use Croco’s useful turn of phrase), which fell outside the original scope of our inquiry.

Thinking this possibility through, it seems likely that the impact of casualties will be multifaceted. Gelpi posits that that high (free) media access/low party systems are likely to exhibit constraint on conflict already underway because the casualty counts creep through to the voters through media access alone. This is of similar spirit to Croco’s off diagonal prediction, but the mechanism is different. Where Croco sees low tenacity in these states due to the smaller size of the governing coalition (meaning less parties own the policy), Gelpi sees more pressure from the public due to their relatively independent access to information on casualties. However, in his view, democratically elected leaders will not actually alter their policies regarding these costly conflicts unless they face opposition from competing elites who seek to electorally exploit public opposition to the conflict– something that is more likely to happen in higher party systems. Thus, in Gelpi’s telling, we end up in the same place as our informational story, but by a slightly different route. These mechanisms, however, should be distinguished in future work.

The upshot is that where Croco implies the potential for some responsiveness in low party countries because of the relatively small governing coalitions that would allow for change, Gelpi, drawing on the example of Afghanistan, suggests that there would be backing down among low party/high access states. This is an interesting set of theoretically reasonable hypotheses and possibilities that we intend to explore in follow-up work on constraint on conflicts.

Gelpi also voices concern about the distance between the audience cost argument and reciprocation – a concern that we share. Testing any phenomenon driven by strategic selection with observable data tends to push toward secondary implications. Our study is no exception. The use of reciprocation for this purpose is not our innovation, but it does seem to be the best available option at this point. That said, as Gelpi notes, the reciprocation test can establish the presence of democratic constraint even if it falls short as a test on some the other audience cost arguments that are so prevalent in the literature.

Gelpi notes that some of the effects we report are driven in part by the seemingly bellicose impact of media access in low-party systems. As he notes, we only touch on this issue briefly in the book, but we share Gelpi’s suspicion that the presence of more significant rally effects and ‘spirals of silence’ in low opposition states are driving this effect. Tying this again to the prior points about tenacity in conflicts underway, it would be interesting to see if this translates into a lack of staying power in these states. In particular, this bellicose tendency among low party/high access states would link neatly to the widely held idea that the U.S., at least in recent decades, is quick to get into fights but weak in the face of casualties.

Finally, Gelpi raises the valid issue that scholars need to devote greater attention to the implications of recent changes in the media environment and their potential impact on democratic constraint. In particular he raises the issue of the rise of partisan media across various platforms – traditional and new. This is an interesting point. It introduces the broader question of polarization and the extent to which it might influence the story that we are telling. In particular, changing media technology may allow for smaller audience sizes that can produce ‘echo chambers’ and thereby exacerbate polarization. Our instinct is that as this trend continues it will only increase the salience of our argument and the magnitudes of the effects that we would uncover. It is well-established that multiparty democracies have less polarization than two-party systems, and it seems plausible that the changes in media that we are considering would drive this divide still wider. This means that in two-party systems leaders are increasingly likely to be in a position where they can count on the support of co-partisans and disproval of the opposition almost regardless of the policy. The mechanism of constraint that we describe therefore relies on a shrinking middle and a media that is increasingly poorly positioned to inform that middle in ways that are likely to lead to meaningful constraint on executive action. This would indeed lead to Gelpi’s corollary that while the ideological diversity of political parties may increase democratic constraint on foreign policy, the ideological polarization of media outlets can undermine this constraint.

Thomas Scotto’s critique focuses on conflict reciprocation, where he identifies a series of interesting concerns about our arguments. Notably, he raises the possibility that our argument might ask too much of authoritarian leaders. The audience cost model already implies that the target of a threat has some capacity to ‘look inside’ the issuer’s polity to understand how regime type could lead to credibility through costly signals. Our argument takes this one step further by requiring that these states not only assess the broad distinction between democracies and autocracies, but actually recognize the difference between high information systems with robust parties and media and those without these attributes.

This is a fair point, to which there are two plausible responses. First, in many instances we believe that these informational attributes that derive from fine-grained institutional mechanisms gel into reputations that other states can rely on as heuristics. Threats and the reactions that states have to them do not arise in a vacuum. Rather, states can look back on the past behavior of the state issuing a threat, and the reputation its behavior engenders, to estimate how the threatening state is likely to proceed in the present instance. Thus while institutions may be the ‘deep cause,’ adversaries need not assess them in detail in real time. Second, selection is an important element of the credibility mechanism that we identify. Constrained democrats contribute to the observed effect by limiting their threats to cases where they are the stronger party and therefore anticipate that the other side will back down out of weakness rather than an ability to forecast intent. The target does not back down because they know that the threatened state is credible due to its institutions. Rather, they back down because they are the weaker party.

That said, the core point that more should be done to explore the mechanism is well taken. It is likely the case that there is some combination of selection (by the initiator), reputation (on both sides), and perception (by the target) that drives our empirical findings on reciprocation. It would be very helpful both to our argument and to the audience cost proposition in general to parse out the relative influence of each of these factors.

Scotto also points out that we implicitly model the public as being less bellicose than leadership. He rightly notes that this is not always the case; sometimes the public may be bloodthirsty while leadership sees strategic value in restraint. While we are unaware of available data to test this proposition, our expectation is that the theory would still work, but in such cases high party/high access states would initiate more conflicts. Our deeper point is about the extent to which leaders’ actions are congruent with the public’s preferences rather than specifically about the probability of conflict.

Similarly, Scotto asks whether we might fruitfully ask how different states might navigate gaps between elite and public preferences over other foreign policy matters such as trade or foreign aid. We have not done these analyses, but our theory does provide clear and testable hypotheses that could guide such and analysis. We expect that high party/high access states will tend toward the public’s preferences while those without both of these attributes will tilt more toward the preferences of leaders.

This brings us to the question of how we should be thinking about emerging or poorly institutionalized democracies. We largely skirt this issue in the book, but it is firmly established that transitional democracies, anocracies, or whatever else we may call them, are the problem children of the international system.[17] The numeric thresholds at which we cut off democracy in the book generally exclude these states. But future analysis would do well to consider them. Our expectation is that since institutionalization and reliable information transmission are our basic ingredients for constraint, the leaders of these states should be relatively unconstrained, all else being equal. Chaotic institutional environments are not conducive to establishing the link between media and elites that will reliably transmit information. The mechanisms of elite opposition and contestation themselves are likely to be in flux in such states.

Finally, Scotto asks whether, given our work, there remains a role for the starting beliefs that individuals have about foreign policy. To answer, we do not see ourselves as displacing the extant literature on the structure of citizens’ foreign policy attitudes regarding the desirability of military readiness or appetite for conflict. We do think we have something to say about the deeper origins of these attitudes and when they are likely to influence actual foreign policy. Specifically, we would argue that to a certain extent these attitudes are endogenous to institutions. To the extent that they are not, institutions then shape the degree to which these public preferences and beliefs are subsequently manifested in actual policy, or are ignored.

In closing, we appreciate the thought-provoking engagement with our arguments, and H-Diplo’s role in convening this conversation.



[1] E.g., Douglas L. Kriner, After the Rubicon: Congress, the President and the Politics of Waging War (Chicago: University of Chicago Press, 2010) and William G. Howell, and Jon C. Pevehouse, While Dangers Gather: Congressional Checks on Presidential War Powers (Princeton: Princeton University Press, 2007).

[2] D. Scott Bennett and Allan C. Stam III, “The Declining Advantage of Democracy: A Combined Model of War Duration and Outcomes,” Journal of Conflict Resolution 42.3: 344-366.

[3] Dani Reiter and Allan C. Stam III, Democracies at War (Princeton: Princeton University Press, 2002).

[4] Kenneth A. Schultz, Democracy and Coercive Diplomacy (Cambridge: Cambridge University Press. 2001).

[5] Sarah E. Croco, Peace at What Price? Leaders and the Domestic Politics of War Termination (Cambridge: Cambridge University Press. 2015), especially Chapter 5.

[6] See B. Geddes, What Do We Know about Democratization after Twenty Years?. Annual Review of Political Science 2:1 (1999): 115-144; B. Geddes, J. Wright, and E. Frantz, “Autocratic Breakdown and Regime Transitions: A New Data set,” Perspectives on Politics 12:2 (2014): 313-331; J.L. Weeks, “Autocratic Audience Costs: Regime Type and Signaling Resolve,” International Organization, 62:1 (2008): 35-64; and J.L. Weeks, “Strongmen and Straw Men: Authoritarian Regimes and the Initiation of International Conflict,” American Political Science Review 106:2 (2012): 326-347.

[7] See W. L. Bennett, “Toward A Theory of Press-State Relations in the United States,” Journal of Communication, 40:2 (1999): 103-127; W.L. Bennett, W.L., R.G. Lawrence, and S. Livingston. “None Dare Call it Torture: Indexing and the Limits of Press Independence in the Abu Ghraib Scandal.” Journal of Communication 56:3 (2006): 467-485; Bennett, W.L., Lawrence, R.G. and and S. Livingston, When the Press Fails: Political Power and the News Media from Iraq to Katrina (Chicago: University of Chicago Press, 2007); and D. C. Hallin, The Uncensored War: The media and Vietnam (New York: Oxford University Press, 1989).

[8] See C. Gelpi, P.D. Feaver, and J. Reifler, “Success Matters: Casualty Sensitivity and the War in Iraq,” International Security 30:3 (Winter 2005/2006): 7-46; C. Gelpi, “Performing on Cue? The Formation of Public Opinion toward War,” Journal of Conflict Resolution 54:1 (2010): 88-116; and S.D. Gartner, and C.F. Gelpi, “The Affect and Effect of Images of War on Individual Opinion and Emotions,” International Interactions 42:1 (2016): 172-188.

[9] Michael A. Maggiotto and Eugene R. Wittkopf, “American Public Attitudes Toward Foreign Policy”, International Studies Quarterly 25:4 (1981): 601-631, DOI:; Ole R. Holsti, Public Opinion and American Foreign Policy (Ann Arbor: University of Michigan Press, 2004).

[10] Paul R. Hensel, Gary Goertz, and Paul F. Diehl, “The Democratic Peace and Rivalries,” The Journal of Politics 62:4 (2000): 1173-1188, DOI:; Henk E. Goemans, War and Punishment (Princeton: Princeton University Press, 2000).

[11] Jason Reifler, Thomas J. Scotto, and Harold D. Clarke, “Foreign Policy Beliefs In Contemporary Britain: Structure And Relevance,” International Studies Quarterly 55:1 (2011): 245-266. DOI:; Timothy B. Gravelle, Thomas J. Scotto, Jason Reifler, and Harold D. Clarke, “Foreign Policy Beliefs And Support For Stephen Harper And The Conservative Party,” Canadian Foreign Policy Journal 20:2 (2014): 111-130, DOI:

[12] Jack Snyder and Erica D. Borghard, “The Cost Of Empty Threats: A Penny, Not A Pound,” The American Political Science Review 105:3 (2011): 427-456; DOI:; Marc Trachtenberg, “Audience Costs: An Historical Analysis,” Security Studies 21:1 (2012): 3-42, DOI:

[13] Jon Hurwitz and Mark Peffley, “How Are Foreign Policy Attitudes Structured? A Hierarchical Model,” The American Political Science Review 81:4 (1987): 1099-1120, DOI:

[14] Dan Reiter and Allan C. Stam, Democracies at War. (Princeton: Princeton University Press, 2002).

[15] There is some evidence in the reciprocation portion of our argument that constrained leaders are, in fact, selecting better (Chapter 4).

[16] See C. Gelpi, P.D. Feaver, and J. Reifler, “Success Matters: Casualty Sensitivity and the War in Iraq,” International Security 30:3 (Winter 2005/2006): 7-46; C. Gelpi, C., “Performing on Cue? The Formation of Public Opinion toward War,” Journal of Conflict Resolution 54:1 (2010): 88-116; and S.D. Gartner, and C.F. Gelpi, “The Affect and Effect of Images of War on Individual Opinion and Emotions,” International Interactions 42:1 (2016): 172-188.

[17] Edward D. Mansfield and Jack Snyder. Electing to Fight: Why Emerging Democracies go to War (Cambridge: MIT Press, 2007).