Roundtable 9-9 on Endurance and War: The National Sources of Military Cohesion

Endurance and War coverThe study of military effectiveness in political science has come a long way in a short period of time. When I started graduate school in the mid-1990s, most of the key works on the subject were written by historians and sociologists rather than political scientists.[1] Beginning in the late 1990s, however, military effectiveness began to enter the mainstream of international security studies in political science. Scholars began to produce a series of works that detailed, inter alia, the martial shortcomings of dictatorships and Arab states,[2] the battlefield virtues of democracies,[3] the critical importance of the ‘modern system’ of force employment,[4] and the link between civil-military relations and effective preparation for and conduct of hostilities.[5] Lively debate continues on many of these subjects, particularly the relative effectiveness of different regime types and how civil-military relations influence adoption of the modern system.[6] This debate has unfolded primarily in the context of conventional (interstate) war, but a related literature on effectiveness in counterinsurgency has been reinvigorated in the wake of U.S. occupations in Afghanistan and Iraq.[7]

H-Diplo | ISSF Roundtable, Volume IX, No. 9 (2017)

A production of H-Diplo with the journals Security Studies, International Security, Journal of Strategic Studies, and the International Studies Association’s Security Studies Section (ISSS).

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

Jasen J. Castillo.  Endurance and War:  The National Sources of Military Cohesion.  Stanford:  Stanford University Press, 2014.  ISBN:  9780804789103 (cloth, $55.00).

Published on 3 January 2017



© Copyright 2017 The Authors

Introduction by Alexander B. Downes, The George Washington University

The study of military effectiveness in political science has come a long way in a short period of time. When I started graduate school in the mid-1990s, most of the key works on the subject were written by historians and sociologists rather than political scientists.[1] Beginning in the late 1990s, however, military effectiveness began to enter the mainstream of international security studies in political science. Scholars began to produce a series of works that detailed, inter alia, the martial shortcomings of dictatorships and Arab states,[2] the battlefield virtues of democracies,[3] the critical importance of the ‘modern system’ of force employment,[4] and the link between civil-military relations and effective preparation for and conduct of hostilities.[5] Lively debate continues on many of these subjects, particularly the relative effectiveness of different regime types and how civil-military relations influence adoption of the modern system.[6] This debate has unfolded primarily in the context of conventional (interstate) war, but a related literature on effectiveness in counterinsurgency has been reinvigorated in the wake of U.S. occupations in Afghanistan and Iraq.[7]

As some of the contributors to this roundtable observe, however, the burgeoning military effectiveness literature has focused almost entirely on those factors that promote or inhibit states from fielding armies characterized by a high level of skill. Do their soldiers have the ability to execute basic tactics, such as exploiting cover and concealment or coordinating suppressive fire and movement? Can larger units combine different combat arms to achieve breakthrough and exploitation? Explanations for will—the grit and determination with which armies fight, especially when facing adversity—have lagged far behind. Some arguments—such as those emphasizing small unit cohesion—are more than half a century old, whereas more recent ones—like those highlighting the motivational advantages of democracies—have received tepid empirical support.[8]

Enter Jasen Castillo’s new book, Endurance and War: The National Sources of Military Cohesion. Castillo aims to explain why some armies fight hard in the face of bleak battlefield prospects, launching counterattacks and resisting the temptation to surrender, whereas others collapse in disorder. Why, for example, did the French Army in 1940, fighting in defense of its homeland just as its predecessor had in 1914, crumble in six weeks, suffering fewer than 100,000 combat deaths but losing over 2 million prisoners, whereas the army of 1914 suffered more fatalities in the first weeks of the war yet recovered and fought on for years? Why did the North Vietnamese Army outlast the technologically superior U.S. military despite suffering combat deaths an order of magnitude greater than what it inflicted on the Americans? Why, in other words, can some armies but not others maintain their cohesion in the face of adversity?

Castillo conceptualizes cohesion as consisting of two aspects: staying power and battlefield performance. Staying power refers to an army’s “ability to remain disciplined and capable of conducting operations as the probability of victory decreases” (19). When faced with battlefield reverses, do officers and men continue to fight and carry out orders, or are there mutinies, extensive desertions, and combat refusals? Can the army, in other words, hold itself together as an effective combat force? Battlefield performance refers to the “willingness of combat units to fight with determination and flexibility” (21). Determined units, according to Castillo, respond to setbacks, high losses, prolonged bombardment, or encirclement by continuing to resist, maintaining high morale, and launching counterattacks (21). Flexible units are able to resist panic, form new units out of shattered ones, and improvise when plans fail or they are cut off from communication with higher headquarters (21-22).

For Castillo, the sources of military cohesion lie in two factors: the persuasiveness of the regime’s ideology and coercive powers, on the one hand, and the military’s ability to train its troops free from political interference on the other. First, regimes that exert high levels of control over society—through a combination of an ideology that demands “unconditional loyalty” (such as nationalism, communism, or fascism) and the power to compel recalcitrant individuals to conform—will field militaries with greater staying power than states with low levels of societal control (28). This is important—and contrary to prevailing wisdom—because it means that certain kinds of autocracies have a fundamental advantage over democracies in producing cohesive militaries. Second, states that provide their military establishments with the autonomy necessary to engage in rigorous and realistic training will generate armies that fight in a determined yet flexible fashion.

Together, regime control and military autonomy yield four ideal types of militaries. Messianic militaries are the most fearsome of the lot. Produced by countries with high levels of regime control that give their militaries the autonomy to train, such as Nazi Germany, messianic militaries possess great staying power and superior battlefield performance. Authoritarian militaries are also generated by nations with strong regime control over society, but are a notch below their messianic cousins because the regime systematically interferes in the military’s affairs. These militaries have strong staying power but are less nimble on the battlefield. The Red Army under Joseph Stalin is a good example. Countries with low regime control but high military autonomy produce professional militaries. These militaries—such as the U.S. military in Vietnam—perform well in battle but gradually lose the will to fight as victory recedes into the distance. Apathetic militaries, finally, are characteristic of states with both low regime control and low military autonomy, like France in 1940. These militaries fall apart quickly when faced with adversity.

The participants in this roundtable find a great deal to admire in Endurance and War. Most effusive is Brendan Green, who lauds the book as a “tour de force,” extols the book’s theory as a “triumph,” and concludes that Endurance and War “may well be the last word on military cohesion.” Similarly, Caitlin Talmadge contends that Endurance and War is “[b]y far the best work on cohesion in many years” and praises the book as a “compelling, innovative, and thought-provoking account of a critical aspect of military performance that provides both scholars and policymakers with essential insights into the nature of war.” Phil Haun praises the book as “well written and accessible,” while Austin Long finds the book’s marriage of “parsimonious theory…with extensive empirical research…illuminating and engaging.”

The reviewers also note that cohesion theory travels beyond the cases that Castillo addresses in the book. Talmadge, for example, points out that Castillo’s argument helps explain the poor performance and eventual collapse of the South Vietnamese Army. Long contends that the theory illuminates the reasons for the disintegration of the Iraqi Army in 2014 and the strong cohesion of non-state actors like the Islamic State and Hezbollah. Talmadge also notes that cohesion theory calls into question the use of some indicators of military skill—such as loss-exchange ratios (one side’s casualties divided by its opponent’s casualties)—since “some militaries have proven effective precisely because they were willing to suffer enormous casualties.” Relative casualty counts therefore may not provide a straightforward metric of battlefield performance.

Although the reviewers generally agree that Endurance and War represents a major advance in the study of military cohesion, they are not without criticisms. Green, for example, argues that Castillo devotes too much attention to coding the value of the dependent variable in his cases at the expense of the independent variables. In particular, Green questions whether the French Army in 1940 was noticeably different from its 1914 predecessor. The two armies obviously performed very differently on the battlefield, but Green argues that both had tense relations with their political masters and that the 1940 edition was at least not worse off in terms of training than the 1914 version. Green suggests that the key difference between the two forces was that by 1940 the French officer corps “had become utterly corrupt” and had lost the will to fight.

Haun observes that armies in Endurance and War sometimes shift from one theoretical category to another between different wars (e.g., the German Army’s change from professional in World War I to messianic in World War II), but points out that this is also possible within the same war. Haun contends that the Red Army in 1941 was an authoritarian military, ravaged by Stalin’s extensive and murderous purges. As the war progressed, however, Stalin granted greater autonomy to the military, perhaps causing it to change type.[9] Future research could investigate the question of intrawar variation on Castillo’s independent variables.

Long devotes much of his essay to the question of whether Castillo’s cohesion theory applies to contemporary all-volunteer forces (AVFs). All of the armies Castillo examines in Endurance and War are mass mobilization forces composed of large numbers of conscripts. Because today’s AVFs are composed of individuals who chose to enlist rather than draftees, these armies may display higher levels of cohesion than the conscript services of previous eras. Long suggests that NATO operations in Afghanistan since 2001 provide a fertile testing ground for this idea.

Talmadge expresses three reservations about the book. First, she argues that Castillo’s theory sometimes under-predicts the cohesion of democratic militaries, which for Castillo are at a fundamental disadvantage because they lack high levels of regime control. Both the Israeli and U.S. militaries have displayed high levels of cohesion in several conflicts, perhaps suggesting that some additional factor is at work in these cases. Second, Talmadge questions the link between military autonomy, realistic training, and cohesion, noting that many militaries that have been left to their own devices by political leaders have nevertheless failed to train effectively for war. Moreover, the highly cohesive North Vietnamese military was not autonomous from political intervention and many of its officers were also political leaders. Finally, although she agrees that Castillo’s “emphasis on ‘will’ over ‘skill’ does offer a much-needed corrective to the notion that training is just about learning tactics,” Talmadge argues that the two concepts are not neatly separable. As she puts it, “Training builds the will to fight, but it likely does so in part by convincing soldiers that they and their leaders possess the skills necessary to prevail.” Talmadge highlights the interaction of will and skill as a fruitful avenue for further research.

These criticisms notwithstanding, I think I can safely speak for the reviewers when I say that Jasen Castillo has written—if not the last word on military cohesion—then an excellent book on a critically important but much neglected subject. The military effectiveness literature has been remarkably unbalanced in its focus on skill over will. Hopefully Endurance and War will begin the re-balancing of this literature. There is no doubt that every work on cohesion that follows will have to reckon with its powerful and persuasive arguments.


Jasen J. Castillo is an Associate Professor at Texas A&M University’s George H.W. Bush School of Government and Public Service. He came to the Bush School after serving on the staff of the Policy Planning Office at the U.S. Department of Defense. Before working at the Pentagon, he was a defense analyst at the RAND Corporation and a consultant for the Institute for Defense Analysis. He holds a Ph.D. in political science from the University of Chicago. His publications include: Endurance and War: The National Sources of Military Cohesion (Palo Alto: Stanford University Press, 2014); Flexible Response Revisited: Assessing Pakistan’s Potential Nuclear Strategies, PM-2383 (Santa Monica: The Rand Corporation, 2007); Striking First: Preemptive and Preventive Attack in U.S. National Security Policy (Santa Monica: The Rand Corporation, 2004); “Nuclear Terrorism: Why Deterrence Still Matters,” Current History 2:668 (2003), Economic Growth and Military Expenditures, MR-112-A, (Santa Monica: The Rand Corporation, 2002). His current work focuses on nuclear deterrence, especially the future of the nuclear revolution.

Alexander B. Downes (Ph.D., University of Chicago) is Associate Professor of Political Science and International Affairs at The George Washington University. Downes is the author of Targeting Civilians in War (Ithaca: Cornell University Press, 2008), winner of the Joseph Lepgold Prize for best book published in international relations in 2008, as well as numerous articles and chapters on civilian victimization, foreign-imposed regime change, military effectiveness, democracy, coercion, and solutions to civil wars. His current work on the consequences of foreign-imposed regime change for intervener-target trade and conflict is forthcoming in the British Journal of Political Science and International Security.

Brendan Rittenhouse Green is Assistant Professor of Political Science at the University of Cincinnati. For the 2015-2016 academic year, he is the Stanton Junior Faculty Fellow in Nuclear Security at MIT. He writes on issues of international security, nuclear strategy, and foreign policy; his work has appeared in venues like International Security, the Journal of Strategic Studies, and The National Interest online edition. He is fanatically dedicated to terrible sports teams like the Cleveland Indians and Cincinnati Bengals, despite the crushing material superiority of the competition.

Phil M. Haun is Professor of Aerospace Studies at Yale University.  His research and teaching is on military strategy, national security affairs, coercion, deterrence, and air power theory.  He is the author of Coercion, Survival and War: Why Weak States Resist the United States (Stanford University Press, 2015).

Austin Long is an Associate Professor at the School of International and Public Affairs and a Member of the Arnold A. Saltzman Institute of War and Peace Studies and the Harriman Institute for Russian, Eurasian, and East European Studies at Columbia University. He is also a non-resident Senior Fellow at the Foreign Policy Research Institute. Long was previously an Associate Political Scientist at the RAND Corporation.  He was an analyst and adviser to the U.S. military in Iraq (2007-2008) and Afghanistan (2011 and 2013). In 2014-2015, he was a Council on Foreign Relations International Affairs Fellow in Nuclear Security, serving in the Joint Staff J5 (Strategic Plans and Policy) Strategic Deterrence and Nuclear Policy Division. Long’s research has appeared in International Security, Security Studies, the Journal of Strategic Studies, Orbis, and Survival. He is also the author of The Soul of Armies: Counterinsurgency Doctrine and Military Culture in the United States and United Kingdom (Ithaca: Cornell University Press, 2016). He received his B.S. from the Sam Nunn School of International Affairs at the Georgia Institute of Technology and his Ph.D. in political science from the Massachusetts Institute of Technology.

Caitlin Talmadge is Assistant Professor of Political Science and International Affairs at the George Washington University. She is author of The Dictator’s Army: Battlefield Effectiveness in Authoritarian Regimes (Ithaca: Cornell University Press, 2015) and co-author of U.S. Defense Politics: the Origins of Security Policy (New York: Routledge 2014). Dr. Talmadge was a Stanton Nuclear Security Fellow at the Council on Foreign Relations in 2014-2015 and is currently writing a book on the problem of nuclear escalation in conventional war. She holds a Ph.D. in Political Science from the Massachusetts Institute of Technology and an A.B. in Government from Harvard.


Review by Brendan Rittenhouse Green, University of Cincinnati

The field of academic security studies has been much shaped by an intellectual tradition one might term the ‘Chicago school.’ Members of the Chicago school are characterized by sophisticated, high-style social science theorizing of the type that would make sociologist Max Weber proud. Though usually influenced in some way by John Mearsheimer—either in his fascination with murder and mayhem, or his advocacy for political realism—these denizens of Hyde Park are conceptually eclectic in their approach to understanding the problems of large-scale violence. One thinks here, inter alia, of Dale Copeland’s careful theorizing about future expectations for trade and war; Sebastian Rosato’s provocative re-conceptualization of European integration; Roger Petersen’s turn towards the micro-level in the study of ethnic violence; Alexander Downes’ examination of civilian targeting during war; or Keir Lieber’s dissection of the offense-defense balance.[10] Together the Chicago school has cast much needed light onto obscure terrain in international politics; if these subjects remain dark, it is at least no longer for lack of our knowledge.

This history bears repeating because, with the publication of Endurance and War: the National Sources of Military Cohesion, Jasen Castillo takes his place among the ranks of the Chicago school.[11] The book is a tour-de-force, and bears all the hallmarks of the great tradition in which it stands. It tackles an incredibly important but understudied question: how are we to understand the ‘moral’—as opposed to the ‘material’—aspects of military power? In particular, why do some great armies cohere in the face of impossible difficulties, while other similarly situated militaries collapse under pressure? Castillo draws up a conceptual and theoretical vocabulary with which to understand the problem, synthesizing the best of existing work en route to his own original model of military cohesion. The model is put through its paces in a series of well-chosen cases, and it performs quite well. In the end, Endurance and War is a difficult book to criticize: one of only a few systematic studies on the topic, it may well be the last word on military cohesion.

I shall nevertheless try to criticize it, or at least make some suggestions for further research. I begin by describing the most exciting aspects of the book and the most convincing portions of the analysis. Next, I offer some reservations.  Finally, I make some suggestions for how the field might pursue inquiry into ‘cohesion after Castillo.’

Cohesive Arguments

Castillo deserves credit first and foremost for tackling an essential and difficult question. Realist analysis sometimes sounds a bit as though international politics were a game of chess: cool calculations made in automatic reaction to the balance of material forces in a given geography. But in chess, knights do not turn and run away because pawns look at them fiercely. Those with a special genius for violence have long understood that battlefield puissance rests as much in the minds of men as it does in their number or equipment, and Castillo quotes them liberally. Napoleon Bonaparte’s well known remark that the “moral is to the physical as three to one” (17) and Karl Von Clausewitz’s admonition that “the physical seems little more than the wooden hilt, while the moral factors are the precious metal, the real weapon, the finely-honed blade” (1) underscore the key point: we cannot understand the dynamics of the ‘balance of power’ without grasping the will of the men who will be its measure.

But the topic has proved elusive. A graduate school colleague once remarked to me how little we knew on the topic of what motivates men to endure the rigors of combat, as the two most famous studies at the time had undergone major revision or refutation. S.L.A. Marshall’s claim that fear prevented many soldiers from firing their weapons had been brought into empirical disrepute, and Edward A. Shils and Morris Janowitz’s argument that small unit bonds drove fighting cohesion in the Wehrmacht had been undermined by Omer Bartov’s revelations about the importance of ideological motivations for Nazi soldiers.[12] Castillo expands on these intuitions by pointing out that even where powerful, small unit bonds cannot scale up to explain large unit collapse and cohesion. Similarly, while lots of ideologies can motivate men to fight and die, they do not appear to do so equally or evenly: for instance, the powerful nationalisms of the type that gripped Europe after the French revolution appear to have motivated armies with different intensities and results (27).

In short, while we might possess some fair propositions about why men endure war, we have little systematic understanding of variation in that endurance—especially under extremely difficult conditions. No matter what the reasons, why continue to fight when the struggle is clearly hopeless? The fate of nations, and the success of their foreign policies, can hang on this central question of cohesion. As Secretary of State John Kerry famously put it, “How do you ask a man to be the last man to die for a mistake?”[13] A glib version of Castillo’s two-fold answer would be: “don’t ask them” and “don’t let them believe it’s a mistake.” Less glibly, Castillo’s model is a sophisticated blend of constructivist and rationalist mechanisms, with all the elegance and rigor one expects from the Chicago school.

The engine of the theory is ideational. Castillo begins with a standard story about nationalism: that men fight for “imagined communities” which extend beyond the real relationships they can see and touch. Men who continue to struggle against long odds do so in large part because they assimilate norms of unconditional loyalty and trust; even an honorable surrender seems a betrayal of sacred values, while even the slightest hope of victory seems worth the risk. To develop such an ethos requires profound ideological indoctrination (25-27).

Castillo then adds three important insights to the common story. First, both military organizations and political units can function as imagined communities; both have the resources to supply the necessary ideological indoctrination. Second, some political units and organizations will be more effective than others at supplying the necessary motivation. Political communities with an ideology that demands unconditional fealty to group goals over individual self-interest are much more likely to produce cohesion on the battlefield. Men will wage a hopeless struggle to defend the spirit of their people, their Gods, or the revolution; contrary to recent wisdom, liberal ideologies are unlikely to so inspire.[14] Meanwhile, military organizations can produce imagined communities only if they are given the space for intensive training, where they can inculcate a team ethic, trust between officers and men, and a set of values that exists separate from society at large (28-31).

Third, Castillo describes the important interplay between the two communities. A truly cohesive political ideology requires control over civil-society, or social divisions will disrupt national unity and home-front dissent will make its way into the military. At the same time, a totalitarian regime that suppresses ideological dissent may feel the need to restrict the military’s autonomy to train and indoctrinate its force to endure battlefield rigors. A free-wheeling civil society, though, does not guarantee that the military will be able train effectively: the military may end up reflecting social divisions rather than transcending them, making the prospect of dying for either the army or the regime of dubious value. And even a military that trains its men effectively into a separate caste will find limits to its cohesion: as the probability of victory declines, the unconditional loyalty to the organization can trump more conditional loyalties to the failing political unit.

In sum, the cohesion of men against fire depends on the interaction and strength of their two main sources of loyalty: the political and organizational communities. Castillo finishes his theory by describing two rationalist mechanisms for how these norms are transmitted and enforced. First, states with high regime control over civil society can produce hard-core supporters on the home front and within the military who use social pressure and outright coercion to suppress dissent and enforce loyalty. Second, military organizations with sufficient autonomy can use rigorous and realistic training to prepare their personnel for the shock of combat, especially for its surprises and reversals.

The end result is that paragon of social scientific excellence, the two-by-two matrix of ideal types. ‘Messianic armies’ are the homicidally dedicated product of tightly controlled totalitarian societies that have given their militaries sufficient autonomy to train. Their soldiers are unflinchingly loyal to the regime and to their organization, with groups of hard-core ideologues prepared to coerce anyone who might have doubts. They will collapse only under the weight of crushing material superiority, and can adapt well to battle’s unexpected pitfalls. ‘Authoritarian armies’ are created by totalitarian regimes that do not exempt their militaries from their paranoia, meaning that training is hamstrung by purges or other political considerations. The hard-core regime supporters in these militaries keep them cohesive in the face of disastrous circumstances, but they do not adapt especially well to surprising battlefield events. These armies do not surrender, they simply die (32-34).

‘Professional armies’ come from liberal or divided societies but have the autonomy for the rigorous training that creates a separate caste of men from normal society. They adapt well on the battlefield and yield only slowly in the face of adverse conditions, but eventually breakdown as defeat appears more certain for a regime to which they are only conditionally loyal. Finally, ‘apathetic armies’ reflect organizationally the social divisions of the political units they defend. They lack the training to adapt to battlefield shocks or the will to face tough odds—a sharp rap can often be enough to crack them (35-36).

Before examining the empirical validity of Castillo’s model, it is worth pausing to reflect on its…beauty. Chicago school theories have an undeniable aesthetic attraction about them, and this one is no exception. I have engaged in a somewhat ‘thick’ description here to emphasize how Castillo builds and links together various concepts, but the structure as a whole is quite parsimonious. Two variables (regime control and military autonomy) are sufficient to explain a range of important cohesion outcomes. The mix of rationalist and constructivist mechanisms is logical, compelling, and avoids the still-too-common fallacy that humans cannot act instrumentally upon normative values (or the analogous claim that a failure to act by a “logic of appropriateness” somehow vitiates ideational explanations). The model presents a host of observable implications. Theoretically, it is a triumph.

The empirical assessment of Castillo’s theory is well designed. The heart of his case selection is the comparison of the French, German, and Soviet armies across the first and second World Wars.[15] Exploring variation in military cohesion during these especially brutal conflicts provides a clear test of the model’s mechanisms and predictions. Two chapters on the North Vietnamese Army and the American Army during the Vietnam War provide an examination of military cohesion in the context of counterinsurgency.

The comparison between the French and German cases is especially instructive. Post-Napoleonic Europe saw two very different paths to modernization in France and Germany. France declined in power, stagnated in population, grew more slowly economically, and finally democratized after the Franco-Prussian war. Germany, meanwhile, grew from a confederation of linguistically united provinces to an empire united under Prussian military might; had a high birth-rate; saw a booming economy driven by the industrial and technological might of the Ruhr valley; and resisted far-reaching democratization. Adapting to these changes proved a highly combustible process in both countries, each of which saw widespread social divisions along class, religious, and social lines. At the same time, each country had different traditional patterns of civil-military relations. In short, Castillo’s key ‘imagined communities’ each experience potential for considerable variation in France and Germany.

The cases are very well executed, especially the World War II examples. In coding his independent variables, Castillo effectively contrasts the horrifying unity of Nazi Germany with the comic opera of French politics, and the single-minded efficiency of the Wehrmacht with the hydra-headed disaster of the French Army. The process evidence from the battlefield is poignantly chosen to reflect these differences.  In 1944, a collection of few thousand wounded German “Christmas tree soldiers” defended the doomed city of Aachen, holding off several multi-division assaults over the course of more than a month, and ultimately forcing the allies to take the city in house-to-house fighting (64-67). In 1940, surprised French tank formations retreated from unsupported German infantry sloshing across the Meuse, who carried only small arms (111-112).

The mechanisms in Castillo’s model are also clearly evident. German letters, even from the Western front, emphasize that soldiers felt that surrender would be an apocalyptic outcome. German commanders executed Hitler’s suicidal counter-thrust at the Bulge in 1944, despite knowing it was foredoomed; their men carried out the operation in high spirits, with an obvious trust in their commanders (67-68). And in case anyone got cold feet, the hard-core SS executed suspected deserters (71-72). By June of 1940, in contrast, the French high command preferred to save the army than save the country (123-124). Had hard-core supporters of the French government existed, they probably would not have known which way to aim their rifles.

Arguments More Difficult to Endure

I have two principal reservations about Endurance and War. First, I thought the empirical material was too lopsided in its emphasis on coding military cohesion. Few enjoy the intimate detail of military history more than I, so I greatly enjoyed these sections. But they did not seem all that necessary for establishing the value of the dependent variable in most cases. Readers can probably be asked to accept, on the basis of citation, all the critical features of a campaign they need to know in order to set the stage for how the mass of units were performing (or not) on the battlefield. Order of battle framing and blow-by-blow accounts of conflict are always appreciated by students of a given war, but these descriptions belabor the point when viewed from a theoretical perspective.

This kind of emphasis comes with significant costs. Castillo’s theory is most powerfully confirmed when he presents process evidence of the mechanisms his model posits: demonstrations of trust and loyalty (or fear) within the military; the presence (or absence) of hard-core regime supporters; the two kinds of communal identification motivating officers and men; the positive (and coercive) incentives used to enforce norms. The reader gets a taste of this kind of evidence in the book, and it is piquant, but it is also scattered and somewhat unsystematic. Powerful examples of German letters before/after key battles in 1944 might have been paired with after-action reports from French soldiers in 1940, for instance—one wonders how they described their own motivations or lack thereof?  With more brevity in coding the value of the primary dependent variable, there might have been more space to demonstrate evidence on the theory’s other observable implications.

More importantly, there might have been increased space for a more concentrated emphasis on the Castillo’s independent variables. His model, after all, depends on being able to distinguish between different sorts of ‘imagined communities’ in politics and organization. Alas, even in cases like Germany, where I am largely persuaded, Endurance and War fails to deliver as well as one would like.

For instance, Castillo’s regime control variable consists of two elements: ideological content and the status of civil society. Castillo does an admirable job demonstrating the differences between Adolf Hitler’s special brand of nationalist horror (48-51) and the less unifying version propagated under the Kaiserreich (81-82). But the discussion of civil society— which is critical in his account to the maintenance of unity and the crushing of dissent—gets short shrift. Various civil society groups existed and caused trouble for the Imperial German government before 1914, and Weimar was lousy with them; Hitler apparently… willed them out of existence? Divided and conquered them? Though gesturing at a number of explanations, Castillo’s short treatment of the issue mostly recapitulates the power of Hitlerian ideology. I accept the basic difference he asserts between the two regimes, but Endurance and War persuades best where it clearly and cogently demonstrates such differences.

The focus on demonstrating the differences between observations on the independent variables leads me to my second reservation. In a word: France. Seventy-five years after Marc Bloch gave us our first account of the Third Republic’s fall, France’s defeat remains ‘strange.’[16] Castillo’s rendering of the case only raises new mysteries for me. Central to Castillo’s interpretation is his assertion that the French army of 1914 had more autonomy for training than that of 1940, producing only gradually diminishing resilience as compared to the latter’s sharp collapse. Perhaps this is so, but the evidence for this claim is not to be found in Endurance and War, which contains less than a page on the pre-WWI French military.

Castillo’s story is at odds with the canonical account of French civil-military relations, which emphasizes the violent struggle between French politicians and the military following the Dreyfus affair. Jack Snyder famously argues that the utter insanity of the French offensive à l’outrance in July 1914 was in large part a measure of self-protection against the interference of civilians, who had imposed short service terms and other measures designed to restrict the French military autonomy that is so important to Castillo’s theory.[17] While it is correct, as Castillo points out, that the pendulum had swung back towards a longer three year service term by 1914, his judgment that a wholesale change in training and indoctrination had taken place seems dubious (130). Compare his take to David Hermann’s assessment:

In France…it is doubtful whether the expansions after 1912 brought an overall improvement in fighting power…. the three-year law was still causing more dislocation than anything else by the time the July crisis arrived…. At first the result was nothing but overcrowded barracks and insufficient uniforms and equipment. Once the outgoing class went home, the army was left in the absurd situation of having only one trained class under arms alongside two larger untrained classes during the winter of 1913-1914. The law produced recruits so much faster than officers and NCOs to instruct them that training sunk to perilously low levels. Although the problem of weak unit establishments receded, the hoped-for consequence of quicker mobilization and greater unit cohesion due to peacetime strengths could not be expected for about three years….The French army was in any case less well equipped than the German to absorb such training demands, since it has a somewhat smaller proportion of officers to men, and far fewer of the all important NCOs….Another important obstacle lay in the shortage of large training camps in France….Despite frequent efforts, the war ministry did not secure any new ones by 1914.[18]

Is this the kind of training environment that produces unconditional trust and loyalty to the military institution? Even supposing that Hermann’s picture is over-drawn, are we to believe that a highly autonomous military rapidly emerged from a series of stringent restrictions after a decade-plus blood feud with its civilian masters? Castillo is not wrong about the French army’s cohesion. Whatever the outlandish features of French doctrine, the organization did not lack for élan. But something else is going on here.

Similarly, there is the problem of the 1940 French army. Did it really lack autonomy for training and indoctrination of personnel into an imagined community? Paris had reinstituted the two-year term of service by 1936, meaning that France’s World War II force was being trained at approximately the same level as its fin-de-siècle military. Castillo’s robust treatment of this issue highlights a rather different problem to my eye. He paints a picture of an officer corps that had become utterly corrupt by 1940. The French military leadership was contemptuous of civilians, sympathetic to the anti-democratic right, and ruled by an aristocratic class that valued status and stability over merit and innovation. It appears that all of the divides and disputes within the French military during the Dreyfus era were decisively settled for the worse following France’s Great-War victory. And as Castillo’s account makes clear, it was not really the courage of the average French soldier that failed on the Meuse, but rather that of a series of frail, timorous, incompetent, and ultimately, treasonous commanders. So perhaps this alternative take on the causes of French collapse makes sense—but it is more than passing strange, and departs from the theory presented in Endurance and War.

Cohesion After Castillo

The mysteries of the French notwithstanding, the model’s success in explaining cohesion across its other cases is impressive. So impressive, in fact, that one must wonder where the study of military cohesion is to go from here. Is there anything left to say? I would suggest that by reconceptualizing the dependent variable, security analysts can build on Castillo’s work to study conflicts other than the brutal struggles for national survival that are depicted in Endurance and War.

Castillo has a two-part dependent variable, the second portion of which also has two parts. The first dimension of cohesion is “staying power,” which describes the threshold of material inferiority at which an army will collapse. Empirical indicators include whether top officers obey the regime, whether there are large-scale mutinies or surrenders, and the rates of desertion, combat refusal, and crime. The second dimension is “battlefield performance,” a measure of whether individual units fight with “determination” and “flexibility.” Determination is indicated by whether units will obey orders to attack and defend, fight in the face of high losses, and form new units from damaged ones. Flexibility is evidenced by the avoidance of panic, tactical flexibility in the face of new circumstances, and fighting even when command and control is disrupted (charts: 20, 22).

I found this schematic overly complex. It is not clear why “staying power” is conceptually different from “determination,” except that perhaps the former uses the entire army as the unit of analysis. “Determination” and “flexibility” seemed distinct enough to merit their own categories, though, rather than being grouped together. Meanwhile, the empirical indicators for staying power highlight what appears to be a third set of behavior: that of individual officers (towards the high command) and men (desertion, crime, and combat refusal rates), rather than units. As I read through the cases, I thought the most natural understanding of the dependent variable was much simpler, reflecting the two dimensions Castillo groups under “battlefield performance.”  That is, cohesion reflects the willingness of units to obey difficult orders in the face of materially poor odds, and it also reflects an ability to avoid panic and keep fighting in the face of surprise. Cohesive behaviors can then be aggregated at whatever unit of analysis one likes, from platoons to the forces of nations.

A further simplification might be to think of all these behaviors as part of a single continuum. On one end of the continuum are actions that reflect organized disobedience, like strikes, surrenders, mutinies, coups, etc. Less dissolute but still not especially cohesive actions might be termed unorganized non-obedience, including combat refusal, showing up late, or incomplete obedience. More cohesive actions are characterized by obedience to orders, while the most cohesive actions go further towards obedience to intentions, which would capture the difficulty of adapting to surprise and avoiding panic. Something like this scheme might accommodate much of Endurance and War’s considerable contribution, while also providing a suppler tool for looking at cohesion in conflicts other than total war.

For that is where the money is, if there is to be anything further learned in the academic study of cohesion. The complexity of Castillo’s dependent variable, to my mind, looks like it is designed in part to accommodate the case of the United States in Vietnam, where the appalling state of the draft army and its links to popular unhappiness with the war at home can be used as evidence for the proposition that the cohesion of the military is on the decline. But this argument does not seem correct to me; as the evidence in Endurance and War makes clear, the performance of the American military in Vietnam is really not comparable to the French and German militaries from 1917-1918, despite Castillo’s similar coding. Fragging officers and smoking dope are not the same as mutinying or surrendering en masse; worrying about disintegrating morale is distinct from watching fielded forces disintegrate. It is the will of the American body politic that was broken in Vietnam, not its army.

At the same time, it seems eminently fair to stress that the cohesion of the American military was not exactly brilliant. Despite the efforts of his latter day apologists, it is not clear to me that General Creighton Abrahams was ever able to get the Vietnam-era Army to fight a ‘better war’ in any location where he himself was not personally standing. Likewise, once one strips away popular mythologies, American performance in contemporary counterinsurgencies also reveals a mixed record. These kinds of operations often demand that units pursue unappealing missions that involve substantial risks to personnel; obedience is sometimes to the letter, rather than the spirit of such strategies. Thinking of the dependent variable as a single continuum might allow future scholars to explore variation in cohesion at the high end of the scale among counterinsurgents.

Similarly, the late excitement over the concept of ‘hybrid’ warfare might be usefully accommodated by thinking of cohesion along the lines just sketched. The conflict in Eastern Ukraine, with its mix of propaganda, terrorism, insurgency, paramilitaries, and conventional warfare, seems dictated as much by what each side can actually ask its military forces to do without revolt as by the strategic incentives they face. A careful examination of the lower end of the scale might confirm or refine Castillo’s predictions about cohesion for a type of conflict that departs from the great power clashes of Europe.

Still, whatever the future holds, we owe Castillo our thanks. Great power conflict of the type he illuminates for us may or may not return, but the theoretical elegance and empirical richness of Endurance and War is a sure sign of the book’s staying power. To bring about a book of this quality is itself an act of messianic dedication, an unflinching devotion to the imagined community of fellow scholars and the truth we collectively seek. Like the body politic protected by the cohesion of soldiers, it is not clear that we deserve this sacrifice, but we can, both as scholars and as citizens, express our gratitude.


Review by Phil Haun, Yale University

This book examines the puzzle of why some armies fight hard until the bitter end while others do not.  Evaluating the expected wartime performance and endurance of an enemy’s military remains a relevant task for policymakers and military planners assessing the likely outcome of war.  Jasen Castillo does both groups a service with this well-written and accessible book presenting a theory of military cohesion.  Rather than focusing on the material components of military power, he examines the domestic factors that affect an army’s willingness to endure the costs of victory or suffer the losses of defeat.  How well and for how long an opponent will fight are both critical factors in calculating one’s probability of victory, and the costs of fighting making military cohesion a key determinant as to whether to go to war and when to terminate it.

To explain variation in wartime performance and endurance, Castillo focuses on military cohesion, which he defines as an army’s ability to fight with determination and flexibility, along with an army’s willingness to endure losses even when the chances of victory are slim.  Castillo measures military cohesion by observing battlefield performance and military staying power.  Battlefield performance is determined by an army’s determination to fight, by its officers’ willingness to take risks, and by its soldiers’ flexibility to overcome difficult circumstances.  A flexible force neither panics nor shatters when pressured, and units continue to perform their mission even when command and control links are severed.  Military staying power refers to an army which continues to fight even as the probability of victory deteriorates.  When victory no longer appears possible, there emerges from within the military sources of disintegration as soldiers view further sacrifice as futile, or discontent surfaces at home as the public grows dissatisfied with a long and costly war.

While the determinants of battlefield performance appear straightforward, the logic behind a military’s staying power is more complex.  Using a Clausewitzian trinitarian framework, in which war is a phenomenon composed of passion, creativity, and reason, factors which are  normally associated with a nation’s population, military and government, respectively, it is the ruling regime that is primarily responsible for assessing the costs and benefits of war with regards to obtaining national objectives.[19]  If a regime fails to seek peace when facing certain defeat then the degradation of military staying power may be an effort by the military and/or population to reinsert rationality into national decision-making.  The challenge is determining ex ante whether a regime’s defiance is rational and to what degree it is capable of controlling the military and population under dire circumstances.  The Nazi regime faced Allied demands for unconditional surrender, which motivated the state to hold out so long as it had the means to do so and to hope for an end to war on better terms.  Only in the final stages of the war, with Allied and Red Armies breaching German defenses was it no longer reasonable to continue fighting.  But even in defeat, the hope of eking out some concessions from victors may make further resistance worthwhile.

Castillo does not examine in detail this question of the rationality of a regime in its decision to continue a losing war.  Instead, he focuses on the degree to which a regime can control its military without disabling it.  According to cohesion theory, military cohesion depends on two primary factors:  the degree of control a regime has over its military and population, and the autonomy its army has to train and conduct operations.  In combination, high and/or low valuations for these variables yield four ideal types of armies.  A messianic army fights with determination and flexibility under any circumstance as the regime has a high degree of control over its army and population, yet the army maintains the autonomy to train and form bonds of trust between officers and soldiers.  A professional army fights with determination and flexibility as it is given autonomy to train and form bonds of trust, but when facing defeat, the regime does not have the means to coerce its army or population to continue fighting for a lost cause.  An authoritarian army fights with determination as its regime can coerce both the military and the population to fight, but the regime, fearing the domestic threat posed by an autonomous army, prevents its soldiers from exercising the initiative required to respond effectively in combat.  An apathetic army has neither a regime’s firm control nor the autonomy to train.  As a result it is unwilling to endure heavy losses and is unable to make adjustments in combat (2×2 matrix on 34).

Fortunately, messianic armies are rare (225).  A regime capable of coercing the military also has the ability to restrict its autonomy.  Given the potential domestic threat an armed force poses to a ruling regime, there is an incentive to restrict the army’s autonomy.  Only if other measures are available to safeguard the regime from a military coup will it be willing to grant the army the autonomy required to fight well.  Castillo argues that a combination of ideology and nationalism convinced the population of Nazi Germany and Communist North Vietnam to suffer heavy losses in war.  Ideology and nationalism further allowed the regimes to grant the military the autonomy to form a determined and flexible force.  Within a messianic army, hard-core regime supporters are not only willing to fight to the end, but also pressure others to do the same.  Unfortunately, Castillo does not spend much time discussing the implications of varying distributions of regime support within the military or population.  What percentage of hard-core believers is sufficient to be able to convince others to comply?  To what degree must a regime have such internal support before it is willing to cede autonomy?  Answers to such questions might assist in assessing the rare conditions under which a regime can maintain military control for regime stability while granting autonomy to improve army performance.

Though Castillo does not explicitly discuss it, democracies presumably also offer safeguards which subordinate the military to civilian rule without the need for coercive measures.  Recruits are drawn from a population educated on democratic ideals under which the military executes the policies of a government that represents the people’s interests.  Further, civilian leaders vet senior military leaders to ensure loyalty.  As such, democracies are likely to grant the autonomy to enable professional armies.  Democratic governments, however, since they draw their power from the population, lack the coercive mechanisms to convince the military and population to continue fighting indefinitely, as the United States experienced during the Vietnam War.

Though messianic armies are rare, they are extremely dangerous as they punch above their weight and can cause significant damage as demonstrated by Germany and Japan in World War II and the North Vietnamese in the Vietnam War.  By contrast, apathetic armies are only a danger to themselves and to those who depend upon them, as the British discovered with the French in 1940.  A strength of this book lies in Castillo’s ability to examine in detail the endurance and performance of armies in key military engagements.  Castillo persuasively demonstrates that the German victory in May 1940 was delivered not by German strategy or leadership alone, but also by the determination and flexibility of German soldiers to execute their mission at critical moments on the battlefield such as at Dinant and Sedan.  By contrast, French soldiers lacked the willingness and capability to do the same, even when the fate of their nation lay in the balance.  A challenge in evaluating the 1940 French case, however, is that an apathetic army faced a messianic army.  It is difficult to disaggregate German superiority from French incompetence.  Elsewhere, Castillo avoids this conundrum by matching the Germans on the defense against the Allied army in 1944-1945 and the North Vietnamese Army against the Americans in 1965-1968.

Besides providing a typology for armies based on performance and endurance, this book also considers how an army can adapt.  The book examines how the nature of the German army changed from professional to messianic in the displacement of the Weimar Republic with Adolf Hitler’s Nazi regime.  Likewise, a professional army turned apathetic in France during the interwar period where rifts between the government and the military reduced both regime control and army autonomy.   Less thoroughly discussed, but equally interesting, is how an army can change during war as with the case of the Soviet Red Army.  The army had been weakened prior to the war by Joseph Stalin’s purges and molded into an authoritarian army, but, because of the threat to state survival posed by the Germans, Stalin allowed the Red Army the autonomy to develop into a professional force.  The Soviet case does not, however, fit as well with Castillo’s theory.  Further clarification on how such a transformation can take place would be a useful contribution to understanding how armies adapt during war.

Even with these untapped questions, cohesion theory is developed sufficiently to test its hypotheses through rich, qualitative cases.  It is in the detailed studies of military cohesion in World War I, World War II and Vietnam where Castillo is at his analytical best.  As a result, the reader emerges from this work with a better theoretical and practical understanding of how to assess an army’s resolve to fight and endure.


Review by Austin Long, Columbia University

Modern war is a deeply unnatural phenomenon.  While humans have fought one another in an organized fashion for millennia, there were constraints on the scale of violence; the principal source of killing power was human and animal muscle.  Over the past two centuries these constraints have been shattered, as chemical and even nuclear power replaced muscle power, and lethality has increased by many orders of magnitude.  Charging an enemy who is armed with a sword is brave; charging an enemy armed with a machine gun is, from a self-preservation perspective, foolhardy.  This is to say nothing of artillery and aerial bombardment, which kill without even providing the opportunity to see much less charge the adversary.

Yet men (and increasingly women) continue to wage war in the modern era and in some cases even do so when victory seems unlikely or impossible.  In other cases, however, the instinct to flee or shirk from modern war dominates and armies collapse when the odds of victory seem low.  Moreover, some armies that fight on against overwhelming odds fight bravely, but not well.

Jasen Castillo’s Endurance and War is a systematic effort to explore and explain this variation.  He does so by combining a relatively parsimonious theory (based on a variety of existing literatures) with extensive empirical research.  The result is illuminating and engaging.

Castillo begins by surveying the literature on this will to fight, termed military cohesion.  This literature, which Castillo notes has four major strands of research, is deeply unsatisfying or at least incomplete.  Some scholars point to macro variables in society (democracy, ideology/nationalism) while others point to micro variables (individual incentives, small group bonds).  Unfortunately these scholars seldom spend much time in dialogue with one another, so the development of unified theory (or theories) of cohesion has been stunted.

This underdevelopment is puzzling, as cohesion has long been viewed as central to modern war.  For example, during the Cold War one of the major questions about a potential conflict in Central Europe was the role of the Soviet Union’s allies in the Warsaw Pact.  As future Secretary of State Condoleezza Rice noted in 1982:

The armed forces of Eastern Europe are generally understudied, and very little has been written about intramilitary politics in the satellite Warsaw Pact states. Key issues of the ability of the elite to undertake concerted political action and the cohesion of the elite in times of crisis have been afforded virtually no attention. Nevertheless, there is no lack of speculation about how the armed forces and their elites would behave in the event of a NATO-Warsaw Pact confrontation. Additionally, the constant conflict between the Soviet Union and her allies, which produced ‘fraternal’ invasions of Hungary in 1956 and Czechoslovakia in 1968, has led concerned analysts to wonder about the loyalty of the East European forces. It is not too difficult to imagine, for example, circumstances under which the Soviet Union might be faced with armed resistance from elements of the armed forces of the Warsaw Pact nations.[20]

The concern Rice highlighted subsequently produced some significant empirical research but did not lead to a sustained research tradition, in large part because of the end of the Cold War.[21]

Fortunately, Castillo takes up this task and does an excellent job of synthesizing the strands in the literature into his own ‘cohesion theory.’  Cohesion theory posits two central variables. The first is control of a state over its society.  The second is the autonomy a military is granted to pursue professional training and development.

State control is important because it determines how well the state can both inculcate loyalty in troops (via intensive ideological indoctrination for example) and coerce loyalty (e.g. via threats from the state security apparatus).  Autonomy is important because military organizations need autonomy to effectively refine their warfighting skills without political interference but also to promote some of the vital micro level incentives such as professional pride. This latter element, which pops up again and again in historical accounts of warfare under such names as élan and esprit de corps, is particularly important for modern warfare, which requires seizing the initiative to develop and exploit breakthroughs in enemy defensive lines.

From these two variables the author generates a fourfold typology expressed in a standard two by two table which is ubiquitous to a certain research tradition.[22]  At one extreme are messianic armies, which combine high levels of state social control with high levels of autonomy.  These armies are incredibly cohesive and, if sufficiently well resourced, devastating on the battlefield.  At the other extreme slouch the apathetic armies, which lack both autonomy and state control.  These armies perform remarkably poorly, with the whole adding up in many instances to less than the sum of its parts.

In between these two extremes are the mixed cases.  High autonomy and low social control produce professional armies, which fight well but began to flag in conflicts where victory begins to look remote.  Conversely, low autonomy and high social control produce authoritarian armies, which fight hard even when losing, which they often are (at least initially) given the lack of professional development.

Castillo then applies this theory to five carefully chosen cases, which he explores through very careful application of the comparative method.  The Wehrmacht of World War II is the messianic army par excellence. Indeed, it is difficult to imagine any other army ever exceeding the Wehrmacht on both variables, for reasons which are discussed more below.  The Red Army of early World War II is a model authoritarian army, combining extensive totalitarian control with purges and commissars, which gutted professionalism.

The interaction of these two titans is illuminating, as the Red Army survived the 1941 onslaught and over time was able to recover, in large part because it was given more autonomy.  This suggests those who fight authoritarian armies must strive mightily for quick victory, as the state may be able to adjust the autonomy of the army relatively quickly.  The work of Caitlin Talmadge is instructive on this point as well as the existence of intra-army variation on autonomy in some cases (some elite units being given much more autonomy than regular units).[23]

The U.S. Army of Vietnam is Castillo’s main professional army case.  It fought well initially but as victory receded and casualties mounted it did not hold together.  The U.S. Army simply fell apart over time, with ‘coat refusals’ (a euphemism for a mutiny), ‘search and avoid’ missions (where units would agree to perform but would shirk by deliberately avoiding contact with the enemy), and the occasional ‘fragging’ (killing of an officer too eager to actually lead troops into combat).

The performance of the U.S. Army raises a question that Castillo does not directly answer in his case work but which deserves serious treatment in future work.  The professional armies he examines were nonetheless mass-mobilization armies.  Most democracies, and some non-democracies, are increasingly turning to an all-volunteer model rather than conscription.  This has been the case in the United States for more than forty years.

Does cohesion theory apply to these ultra-professional armies? There is a ready-made case in Afghanistan (2001 to present). Several all-volunteer armies from NATO fought extended campaigns with significant hard fighting.  In addition to the U.S. Army and Marine Corps, the British, Canadian, Dutch, and Danish armies all fought hard in southern Afghanistan.  Yet by a certain point (for the U.S. roughly 2012 when the troop surge ended, perhaps different for others) it was clear nothing like ‘victory’ was in reach and much of the territory seized at great cost would be lost to the Taliban.[24]

Castillo’s argument would suggest all of these armies should have started to fray at this point. It is difficult to tell how much this has happened in a granular and systematic way at this point.  Only when more operational records are available will it become clear.  But at least on the surface it seems that the U.S. Army experience did not match its earlier Vietnam experience.

As an example, the 101st Airborne Division deployed to Afghanistan in the surge period. In hard fighting in the east and south it lost 131 soldiers, the most since Vietnam.[25]  This was not a unit shirking the fight.

Yet the soldiers had no real illusions of victory.  One captain noted “It is very hard to see change… It was very hard to get that across to my soldiers.” One of the brigade commanders was only modestly more positive: “They have to accept that it was worth it… It was a fight that needed to happen and no soldier died in vain.”[26]

Why might cohesion theory not apply as well to all volunteer armies?  One aspect could be that the selection bias in all volunteer force recruitment provides a substitute for social control.  Rather than needing to indoctrinate a broad population and then to coerce them with state security to fight hard, those who select in to military service, and particularly into combat arms, may already have an ethos and self-motivation that provides a good substitute.  While there were undoubtedly many who joined for similar reasons in the conscription era of these armies, they may not have achieved critical mass outside certain elite units.

Another possibility is that the smaller size of all volunteer forces allows the individual level incentives and primary group bonds to work more effectively.  These armies are able to do more ‘retail’ work on soldiers rather than ‘wholesale’ work.  Certainly the training periods are longer and often more realistic.

The primary group bonds were also strengthened by the fact that the U.S. Army rarely took major casualties to a single tactical unit (platoon/company) and overall casualties were low by historic standards.  The 101st Division’s loss of 131 soldiers took place over a yearlong deployment.  In Vietnam the unit suffered casualties nearly as great in single battles.[27]  All volunteer armies might prove less resilient in warfare where entire platoons and companies, even battalions, are rendered combat ineffective in minutes or hours, as happened frequently on the Eastern Front in World War II.

A final hypothesis on why this argument might not apply to all volunteer armies is that the soldiers, on limited duration rotations, displace much of the frustration, anger, and depression to the post-deployment period.  In Vietnam these feelings led to combat refusal and drug abuse in theater.  In Afghanistan, these feelings are held in check until back in ‘CONUS’ (the current military jargon for the Continental United States, aka home) where they manifest as prescription drug problems, domestic discord, and in the extreme, suicide.

Yet Castillo can make a strong case these Western all-volunteer militaries are outliers even among all-volunteer militaries.  Both Iraq and Afghanistan have all-volunteer forces and even a casual reader of the news is well aware of cohesion problems in both.  In large battles at places like Mosul and Kunduz both armies have experienced major loss of cohesion when confronted against less numerous but more cohesive opponents.

Indeed, one of the great contributions of cohesion theory is that it very accurately demonstrates why building a cohesive Iraqi or Afghan army is such a challenge, particularly once both were democratized.  Social control is very low in both cases, while corruption and politicization cripple military autonomy.  Indeed, cohesion theory would probably lead one to conclude that the Iraqi Army is beyond repair, absent a major political effort to end corruption and politicization.

This appears to be the conclusion of many in Iraq (and Iran), as rather than fixing the army the government generated an alternative force.  Known in English as the Popular Mobilization Forces (PMF), these are primarily Shiite semi-regular militias.  They have evinced much higher cohesion than most of the Iraqi security forces (absent some elite units such as the Counter Terrorism Service).  They too are volunteers, but the force has much higher autonomy (certainly in terms of officer selection) and it has a religious imprimatur.  Combined with the more intimate nature of these militias, this probably provides a reasonable approximation of social control.[28]

Cohesion theory also accurately predicts that the Islamic State will be very cohesive, as it is.  It has many volunteers, but practices extensive social control in areas it controls.  Its military components are autonomous, in the sense of being able to choose leaders and develop tactics without overt political interference.[29] Given the eschatology running through much of the group’s ideology, it is quite literally a messianic army.  Western militaries should take heed of Castillo’s work in net assessments of conflict with the group.

This returns to the question of whether a more messianic army than the Wehrmacht will ever exist.  The Islamic State and probably Hezbollah demonstrate that messianic armies are possible for non-state actors on a relatively small scale (both number in the tens of thousands rather than hundreds of thousands). But can a modern state combine the level of social control with military autonomy on the scale of Nazi Germany, which brought millions and millions of men into the army over the course of World War II?

I think it would be difficult to do so for a few reasons.  First, effective execution of twenty-first century warfare requires significant investment in technology and training for personnel.  Generating the resources necessary for such investment has, in recent decades, required economic and intellectual openness in ways at odds with truly intense social control.[30] Second, this openness requires states to allow citizens to travel and communicate in ways that further limit social control.

China, for example, has maintained extensive political control of its population but has spent thirty five years evolving towards a more open, global economy.  It tries to control media, but much is able to slip through, especially in new forms such as social media.  China no doubt would still score high on social control, but probably not as high as Nazi Germany.  Ambivalence towards America, with many Chinese loving America the country (home of hip-hop and great higher education) and hating America the government (seeking to undermine China at every turn), underscores the inability to achieve extremely high social control.[31] The same is probably true of Russia, though some might argue the point, given Russia’s ability to survive on resource extraction rents.

In contrast, North Korea probably scores even higher than Nazi Germany on social control.  Yet its army, outside of a few long range artillery and rocket systems, has serious shortfalls in both technology and training.  It also has very low autonomy, so it may be a model authoritarian army but it is not messianic- and it still drives tanks that were old when the Cold War ended.

The Wehrmacht may thus be the high-water mark for cohesion, at least for truly mass armies capable of conducting effective modern operations.  Yet Castillo’s work is hopefully only the beginning of a renaissance in studies of why and how armies hold together when confronting the shock and terror of modern war.


Review by Caitlin Talmadge, The George Washington University

Jasen Castillo’s Endurance and War tackles one of the most important and neglected questions in the study of conflict. Why do militaries vary so dramatically in their willingness and ability to continue fighting on the battlefield, especially in situations of intense combat stress and unfavorable strategic odds? Castillo presents a new answer to this question, cohesion theory, and conducts seven historical case studies to test his argument against potential competitors.

The result is a compelling, innovative, and thought-provoking account of a critical aspect of military performance that provides both scholars and policymakers with essential insights into the nature of war. By far the best work on cohesion in many years, Endurance and War will without question be of deep interest and lasting relevance to anyone interested in civil-military relations, combat effectiveness, military strategy, defense policy, and security studies more generally.

It is common to note that a given work ‘fills a gap’ in the literature, but here the gap is more like a chasm, especially in political science.[32] Some of the most frequently cited and assigned works on military cohesion are now decades old.[33] Furthermore, despite great interest in the general subject of military effectiveness, many analyses still largely avoid the specific question of cohesion. Scholars have focused much more on questions related to strategic assessment, military doctrine and strategy, operational concepts, and tactical skill—which types contribute to victory and defeat, and which armies are likely to adopt and execute which ones successfully.[34] These are important issues, but anyone who witnessed last summer’s shocking collapse of the Iraqi army in the face of Islamic State attacks, or who is still puzzling over the stunningly quick “strange victory” of Nazi Germany over the Third Republic, knows that military cohesion (or lack thereof) is a vital driver of some of the most important events in world politics.[35]

Castillo’s first contribution is simply to help us understand how we should think about this concept. Eschewing some of the narrow definitions that have characterized past research, Castillo emphasizes two related but distinct aspects of cohesion. The first is what he calls staying power: whether a country’s armed forces hold together and continue to execute operations with discipline and resolve, even as the probability of overall victory appears to diminish (19). The second aspect he calls battlefield performance: whether combat units continue to fight with determination and flexibility even after experiencing operational and tactical setbacks (21). Castillo’s parsing here is useful because it enables us to see that military cohesion influences all levels of warfare—tactical, operational, and strategic—and can vary along several dimensions both within and between militaries.

Castillo then presents his theory about why some armies display both aspects of cohesion, some only one, and some neither. For Castillo, the answer lies in two key independent variables: the degree of a given regime’s control over its citizens, and the degree of autonomy that the regime allows the armed forces for training (3). First, regimes with a high level of control over the citizenry, usually forged through a unifying ideology robustly enforced with a strong state security apparatus, will be able to produce militaries with a high degree of staying power. These armies, manned by soldiers imbued with a sense of collective obligation and often backstopped by a high degree of internally directed coercion, will fight hard to the last man. By contrast, regimes lacking this control will be much more likely to have armies that disintegrate quickly when the war looks like it is lost (3-4).

Second, regimes that endow the armed forces with a high degree of autonomy for training allow their militaries to develop strong unit bonds of trust and loyalty that result in determined yet flexible combat performance. By contrast, armies that lack this autonomy will generate combat units lacking these bonds, and therefore will demonstrate less operational and tactical cohesion (30-31).

Combining Castillo’s two independent variables—the degree of regime control and degree of military autonomy—results in four ideal-typical militaries with varying degrees and types of cohesion (32-36). Messianic militaries emerge from a high level of regime control and high military autonomy, and therefore display both staying power and good battlefield performance. Authoritarian militaries result from a high level of regime control in the absence of military autonomy, and will display good staying power but only moderate battlefield effectiveness. Professional militaries enjoy autonomy but do not come from societies characterized by a high level of regime control, and therefore evince strong battlefield performance but only moderate staying power. Apathetic militaries arise in regimes with both poor regime control and low military autonomy, resulting in weak staying power and weak battlefield performance.

Castillo’s case studies explore these ideal-types in action, tracing cohesion theory’s posited connections from regime control and military autonomy to staying power and battlefield performance. The best and most detailed analysis covers World War II battles. Castillo attributes France’s rapid defeat in 1940 to its status as an apathetic military and Nazi Germany’s surprising endurance in 1944-1945 to its status as a messianic military. He also cleverly contrasts each country’s World War II performance with its World War I showing. Castillo argues that the higher degree of autonomy afforded to France’s pre-1914 military enabled it to be a professional military that evinced somewhat better staying power and significantly better battlefield performance than in 1940. In a similar vein, he argues that despite the absence of strong regime control in Wilhelmine Germany, the country generated a professional military with good battlefield performance and only moderate staying power in 1917-1918 as compared to 1944-1945. In addition, Castillo examines the Soviet military under Joseph Stalin, arguing that it was an authoritarian military with immense staying power but only moderate battlefield performance.

Castillo also conducts two shorter case studies drawn from the Vietnam War. He classifies North Vietnam as a messianic military, pointing to Hanoi’s high level of regime control and willingness to allow military autonomy as the sources of its excellent staying power and battlefield performance. By contrast, he characterizes the U.S. military during this era as a professional military exhibiting good battlefield performance but only moderate staying power, helping to explain why the United States eventually withdrew from Vietnam despite facing a materially weaker opponent.

In sum, Castillo’s thorough yet readable case studies provide at least one if not several examples of each variant of military cohesion that his theory predicts. The case selection strategy and the specific battles he chooses to examine also usefully enable the testing of competing arguments drawn from scholarship on the role of nationalism, regime type, and small-group ties. In general, the evidence provides solid support for his argument.

Several important implications and questions flow from these findings. First, Castillo’s theory clearly travels to cases beyond those he examines in the main chapters or even in the conclusion, where he discusses the implications for cohesion in non-state groups. Castillo’s description of an apathetic military, for example, tracks closely with key characteristics of the South Vietnamese army before its collapse in 1975, as well as the Iraqi army prior to its implosion in 2014.[36] Scholars interested in trying to explain puzzling cases of military collapse or endurance will undoubtedly benefit from the application of Castillo’s typology, as will policymakers or other analysts trying to conduct net assessments.

Second, Castillo’s work forces reconsideration of a commonly used measure of military effectiveness, the loss-exchange ratio (one side’s casualties divided by the other side’s casualties). The general assumption in prevailing work employing this metric is that better armies impose more casualties than they suffer, largely because of their superior battlefield strategy and skill.[37] Castillo potentially turns this notion on its head by showing that some militaries have proven effective precisely because they were willing to suffer enormous casualties. This was, after all, a critical difference between U.S. and North Vietnamese forces in the Vietnam War, one integral to any explanation of Hanoi’s ultimate triumph. It was also essential to the eventual Soviet victory over the Nazis. Castillo’s approach to thinking about cohesion adds much-needed strategic context to casualty counts and will force scholars to re-examine what these numbers truly mean.

Third, Castillo’s work goes beyond a recent wave of scholarship that has questioned the purported advantage of democratic regimes in war.[38] He argues that totalitarian regimes are inherently more likely to generate cohesive militaries than their liberal counterparts or even other types of dictatorships. In his view, liberal societies will not—and some dictatorships cannot—adopt the regime control strategies needed to produce the most highly cohesive militaries that he labels messianic (28-30). This is a provocative and somewhat disturbing thought, tempered only by Castillo’s observation that even among non-democratic regimes, the messianic variety will be rare (33).

This claim about regime type is not entirely persuasive, however. At times democracies seem to have performed better than Castillo’s theory would predict. Some democracies, notably Israel but also the United States, have clearly developed highly cohesive armies even in the absence of totalitarian systems of regime control. It is perhaps understandable that Castillo would not classify these democratic militaries as messianic in the same way he does Nazi Germany and North Vietnam. Yet it is also clear that democracies at times can fight much better than the two he examines, France in 1940 and the United States during the late Vietnam War.

Israel, for example, displayed remarkable military cohesion by both of Castillo’s measures even in the darkest days of the 1973 Yom Kippur War.[39] So, too, did U.S. forces in the bloodiest days of World War II when Allied victory was far from assured.[40] The U.S. Army may not have been the Wehrmacht, but it also did not display the behaviors Castillo identifies as undermining U.S. cohesion in Vietnam, such as widespread fragging, refusal to fight, and declining domestic political support. Perhaps some democratic militaries simply become so professional that they develop both battlefield performance and staying power even in the absence of the strong methods of coercive regime control needed to generate this output in authoritarian states. Or perhaps the cohesion of democratic forces partially depends on other variables, such as the nature of the opponent or the cause for which the war is being fought. Scholars should build on Castillo’s contribution to examine these possibilities in future work.

Castillo also raises questions about wartime civil-military relations that deserve further inquiry. He notes, albeit briefly, that the armies able to fight with determination and initiative—the messianic militaries and the professional militaries—arise only in states where the regime trusts the military, because only in situations of civil-military trust will the military be afforded the autonomy needed for the type of training that produces cohesion (31). This is, essentially, a Huntingtonian argument that military independence from civilian interference—what he called “objective control”—is critical to the development of professionalism and ultimately to effectiveness in battle.[41]

Castillo’s treatment of the French and Soviet cases, in which concern about the military’s potential domestic role led to excessive political intervention in the military (albeit in different ways), lends credence to his underlying assumption that militaries do better when left to their own devices. Yet the North Vietnam case is not quite as convincing. It is hard to characterize the North Vietnamese military as highly autonomous, at least not in a Huntingtonian sort of way. The army did train vigorously, but this was not because political leaders left the North Vietnamese military to its own devices but rather because political leaders pushed the military to be this way.[42]

Furthermore, the notion of military autonomy from politics is somewhat inappropriate in the North Vietnamese case, because the political and military leadership of the country were deliberately overlapping and interlinked (as is the case in China today, incidentally). Top generals were dual-hatted as members of the Politburo, for example.[43] This reality does not fit neatly into the typology’s emphasis on high military autonomy as the guarantor of cohesion. Nor does the experience of other militaries that have enjoyed high autonomy—Argentina under the junta and Pakistan during periods of military rule, for instance— but have not used it to implement the sort of training that Castillo theorizes they should. These anomalies do not undermine his broader argument, but they do suggest the need for further theorizing about the links between civil-military relations and military performance.

Last, Castillo’s explanation of how exactly training improves cohesion is not entirely satisfying. His emphasis on ‘will’ over ‘skill’ does offer a much-needed corrective to the notion that training is just about learning tactics; he persuasively demonstrates that training is also very much about creating and cementing bonds of trust within units so that they fight cohesively on the battlefield (30-1). Yet his cases make clear that both will and skill are in part a function of the training environment, and that they interact and mutually reinforce each other in ways that scholars need to explore. Training builds the will to fight, but it likely does so in part by convincing soldiers that they and their leaders possess the skills necessary to prevail. Indeed, Castillo notes briefly in introducing his argument that “the opportunity to train creates familiarity with tactics, techniques, and procedures” (31).

This blending of will and skill comes across in the empirics as well, where Castillo’s discussion of one component of the dependent variable, battlefield performance, seems very much to involve assessments of military skill. Nazi training, he notes, displayed a “rigor and realism” that “rivaled that of any military in Europe” (83). In other words, the Nazis did not just want their soldiers to bond in training; they wanted them to master very specific tactics. These Germans were the heirs to Prussia, after all. Similarly, North Vietnam’s own history of its armed forces shows a very heavy emphasis on skill development in training—not just on the teamwork and bonding that Castillo discusses (170).[44] This tactical proficiency was critical to North Vietnam’s cohesion, and may have been inseparable from it.

Ultimately, skill and will may not be as distinct as Castillo’s argument suggests. By carefully exploring the latter, however, Castillo has made a major contribution to the study of war and usefully paved the way for scholars to examine these and other connections in the future.


Author’s Response by Jasen J. Castillo, Texas A&M University, Bush School of Government

John Mearsheimer equates writing a book to wrestling a bear.[45] To me, lacking much experience with wrestling animals, it felt more like passing a kidney stone. While we might all describe the pain of writing differently, most of us can agree that the solitary nature of our work can also prove difficult. For this reason, I am grateful to the editors for commissioning this roundtable. The forum provides an outlet for researchers to come out from behind their computer screens to share our work and interact with colleagues in a fruitful way. I greatly appreciate the time and effort the participants put into their comments.

The reviewers offer many positive comments. However, as a survivor of the “Chicago school,” acknowledging praise remains difficult. My comments, therefore, will address some of the more critical issues and questions raised by the reviewers.  Rather than pursue every issue individually, I select the major items for discussion. In the end, I hope these responses help guide my fellow students of military effectiveness as we pursue the next phase of research.

Assessing The Cohesion of the French Army

Brendan Green believes I spend too much time describing the cohesion of the different armed forces in my cases studies and too little time explaining the sources of their cohesion. He would have preferred that I give more attention to the independent variables in my theory: regime control and military autonomy. In his view, because of this focus, my description of the cases can appear unclear. For instance, he sees little difference in the organizational autonomy between the French armed forces of 1914 and 1940. The position he advances represents a sensible approach, but is by no means the only way to do business. Let me first explain my research focus and then detail the key differences between the French cases.

There are few set recipes in social sciences. We can all agree that certain ingredients are crucial to good research, like establishing correlations and tracing causation when testing theories. Much of what we do empirically reflects the theoretical objective. Here my aim is to test the plausibility of my argument, cohesion theory, balancing the amount of time I spend on the theory’s independent and dependent variables. To do that, I provide a detailed list of indicators of the independent variables to describe regime control and organizational autonomy, my effort at both precision and succinctness. In this way, I can focus much of the historical narrative on tracing the effects of these variables on military cohesion in war. My goal is to unpack how and why these variables interact to create four ideal types of militaries, and then, to outline why each of these militaries performs differently in combat. I thought it crucial to emphasize the distinct dimensions of cohesion: staying power and battlefield performance. In the end, it is a matter of taste. Although another researcher might decide on a different recipe, my choice to emphasize the dependent variables seems a reasonable course to demonstrate my theory’s plausibility.

Our disagreement about the French cases illustrates this point. All of the ingredients to adjudicate this dispute are present in theory. On the particulars of the French cases, I think what separates 1914 from 1940 is the following: a concerted effort by politicians to assuage civil-military tensions and class divisions before the First World War had a positive effect on military cohesion.[46] France fielded a professional military in 1914 because the  National Revival ameliorated, but did no repair, these cleavages. It may not have solved all of the problems between France and its armed forces, but it certainly helped.[47] These same problems resurfaced before the Second World War and were compounded by a sense of war weariness, which plagued French society as well as its military. For these reasons, France fielded an apathetic military in 1940.[48] As Eugen Weber explains:

There were patriots in France, and they were many, but somehow patriotism was dead. Its demise was not immediately evident, but it had died in the trenches, on the Marne, at Verdun. Patriotism now spoke of pacifism, expressed reluctance to fight—fight for others, then for France as well—and it reflected an ever more weary, ever more desperate search for impossible solutions to insoluble problems. The years before 1914 had been bright with red, white and blue of patriotism and its more blustering incarnation: nationalism. The years before 1939 favored other colors: the red of socialism and communism, the white of pacifism.[49]

Weber describes how war fatigue and its traditional cleavages hurt the cohesion of the French armed forces in 1940:

French and foreign observers were struck by the caution of the fighting troops. In 1934 a confidential report on army morale referred to the debility (atonie) that deprived it of vitality, found weariness among officers, unenthusiastic docility among other ranks. Others concurred, ‘Reconnaissance units show excessive timidity,’ noted Americans, ‘while infantry advances at a slow pace.’ The German attaché was more cruel: ‘Has the French infantry forgotten how to attack?’[50]

In sum, France had no National Revival to bolster its divided and war-weary army before facing a second German invasion in 1940. The combination proved too lethal.

Questions about Cohesion Theory

Caitlin Talmadge raises three intriguing conceptual issues with my theory. First, she believes that I underestimate the cohesion of militaries from democracies. Talmadge points to both the United States and Israel as examples of armed forces that have exhibited strong cohesion. She suspects some additional factor outside of my theory could play a role.

Although I am skeptical that democracies produce the most cohesive armed forces in war, my argument is not that democracies lack any will to fight. Under the right circumstances, they can demonstrate strong cohesion. Here the different facets of cohesion are important. I define cohesion as one part staying power and one part battlefield performance. The former reflects the military’s ability to avoid disintegrating when fighting losing wars, and the latter refers to determination and flexibility in individual battles.  According to cohesion theory, the armed forces of democracies will lack the staying power of more repressive regimes, like Nazi Germany and the Soviet Union, because they cannot rely on coercion to keep their militaries and populations fighting to the death.

Instead, their strength in terms of military cohesion lies mostly with battlefield performance, the second component of cohesion. Depending on the level of organizational autonomy given to their armed forces, democracies will either produce professional or apathetic militaries. A professional military will show exhibit strong battlefield performance, fighting hard in individual battles. In addition, its staying power will decay slowly when victory no longer seems likely. [51] Apathetic militaries, in contrast, will fight with weak determination and flexibility on the battlefield, and their staying power will dissipate quickly when victory looks impossible.

High stakes in a conflict, including defending one’s homeland, are an additional factor outside my theory that could also push armed forces to fight hard. In the book, I describe this as an important background condition that can motivate militaries exhibit greater cohesion. Fighting to defend its homeland undoubtedly bolsters the cohesion of Israel’s professional military. The stakes of a conflict could represent the omitted variable Talmadge suggests also deserves attention when explaining cohesion.

Second, Talmadge argues that my theory does not account for military organizations that possess autonomy from outside interference, but do not focus on warfighting. Again, I think there is some merit to this objection. It is probably true that some militaries do not use their autonomy for preparing for military operations. However, most armed forces focus on training for warfighting since that mission provides them with a rationale to increase their budgets.[52] As cohesion theory contends, ethnic or class cleavages can interfere with this focus on warfighting. Civil-military tensions can also derail this mission.

Finally, Talmadge argues that I do not separate skill from will on the battlefield as neatly as my theory purports. On this point we are in agreement. Cleanly separating out these two concepts is very difficult. Nevertheless, I do not think it is correct to assume that the will to fight of a military derives mostly from trust that officers are skilled. That route to building cohesion seems plausible but indirect. There are more direct paths to building cohesion. Inside the armed forces, cohesion comes the shared experiences through training, which in turn instills strong norms of collective loyalty. Outside the armed forces, a strong regime might try to impose an ideology of unconditional loyalty on its society, which in turn creates a pool of military recruits willing to take great risks in war and pressure others to do the same.[53]

Future Research

The reviewers also point to some new directions for future research. Phil Haun notes that during war a country’s armed forces can change from one ideal type to another type. The example he identifies to illustrate this point is the Red Army in World War II. Haun rightly notes that the amount of control that Joseph Stalin and the Soviet state exercised over the military changed over the course of the war. Later in the conflict, facing less coercion, and armed with more autonomy, the Red Army fought with more determination and flexibility.[54] An authoritarian Red Army seemed to have become a messianic Red Army, mirroring its German opponent. The possibility of militaries changing in the middle of a war seems plausible and deserves further consideration.

Austin Long asks about the applicability of cohesion theory to modern armed forces. He wants to know if the theory applies to militaries that are not large, conscription armies. The cases examined in Endurance and War involve mass mobilized militaries, with conscripts filling their ranks. Today, many modern armed forces, especially the world’s most powerful militaries, are smaller, comprised of volunteers. The All Volunteer Force (AVF) of the United States, for example, has retained its cohesion through long, stalemated conflicts in Iraq and Afghanistan. Long also thinks cohesion theory can shed light on the cohesion of some non-state actors.

As I speculate in the conclusion of Endurance and War, the staying power of the U.S. military probably benefited greatly from its being a volunteer force. Conscript militaries come with stronger ties to society at large. Their staying power is vulnerable to a disgruntled home front that concludes a war is no longer worth fighting. In contrast, armed forces made up of enlistees, or volunteers, will insulate themselves from domestic pressures to end a war when victory appears unlikely. This difference, more than any other factor, probably explains why U.S. armed forces could endure long conflicts in Iraq and Afghanistan but not the Vietnam War.

Cohesion theory can also help explain variation in the cohesion of non-state actors. I agree with Long that ISIS, or the Islamic State, resembles Nazi Germany’s messianic Wehrmacht, both in its staying power and battlefield ruthlessness. Recent work on the war in Afghanistan suggests cohesion theory can play a role explaining combat performance of the Taliban. [55] Finally, the book’s argument should also help U.S. officials as they try to create cohesive armed forces in Afghanistan and Iraq. Cohesion theory, unfortunately, would predict that in both of these countries the United States is fighting an uphill battle to overcome persistent societal cleavages that will undermine efforts to build national armed forces with a strong will to fight.



[1] See, for example, Edward A. Shils and Morris Janowitz, “Cohesion and Disintegration in the Wehrmacht in World War II,” Public Opinion Quarterly 12:2 (Summer 1948): 280-315; Allan R. Millett and Williamson Murray, eds., Military Effectiveness, 1st ed., 3 vols. (Boston: Allen and Unwin, 1987-1988); and Omer Bartov, Hitler’s Army: Soldiers, Nazis, and War in the Third Reich (New York: Oxford University Press, 1992).

[2] James T. Quinlivan, “Coup-proofing: Its Practice and Consequences in the Middle East,” International Security 24:2 (Fall 1999): 131-165; and Kenneth M. Pollack, Arabs at War: Military Effectiveness, 1848-1991 (Lincoln: University of Nebraska Press, 2002).

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

[4] Stephen Biddle, Military Power: Explaining Victory and Defeat in Modern Battle (Princeton: Princeton University Press, 2004).

[5] Stephen Biddle and Robert Zirkle, “Technology, Civil-Military Relations, and Warfare in the Developing World,” Journal of Strategic Studies 19:2 (June 1996): 171-212; Risa Brooks, Shaping Strategy: The Civil-Military Politics of Strategic Assessment (Princeton: Princeton University Press, 2008); and Caitlin Talmadge, The Dictator’s Army: Battlefield Effectiveness in Authoritarian Regimes (Ithaca: Cornell University Press, 2015).

[6] Some recent contributions on regime type and military effectiveness include Michael C. Desch, Power and Military Effectiveness: The Fallacy of Democratic Triumphalism (Baltimore: Johns Hopkins University Press, 2008); Alexander B. Downes, “How Smart and Tough Are Democracies? Reassessing Theories of Democratic Victory,” International Security 33:4 (Spring 2009): 7-51; and Jessica L.P. Weeks, Dictators at War and Peace (Ithaca: Cornell University Press, 2014). On civil-military relations, see Ulrich Pilster and Tobias Böhmelt, “Coup-Proofing and Military Effectiveness in Interstate Wars, 1967-99,” Conflict Management and Peace Science 28:4 (September 2011): 331-350; and Talmadge, Dictator’s Army.

[7] John A. Nagl, Learning to Eat Soup with a Knife: Counterinsurgency Lessons from Malaya and Vietnam (Chicago: University of Chicago Press, 2005); United States Department of the Army, The U.S. Army/Marine Corps Counterinsurgency Field Manual (Chicago: University of Chicago Press, 2007); Jason Lyall and Isaiah Wilson III, “Rage against the Machines: Explaining Outcomes in Counterinsurgency Wars,” International Organization 63:1 (Winter 2009): 67-106; Jason Lyall, “Do Democracies Make Inferior Counterinsurgents? Reassessing Democracy’s Impact on War Outcomes and Duration,” International Organization 64:1 (Winter 2010): 167-192; and Austin Long, The Soul of Armies: Counterinsurgency Doctrine and Military Culture in the US and UK (Ithaca: Cornell University Press, 2016).

[8] On small unit cohesion, see Shils and Janowitz, “Cohesion and Disintegration in the Wehrmacht.” On democracy, see Reiter and Stam, Democracies at War, chapter 3. For a critique of democracy as an explanation for how hard soldiers fight, see Risa A. Brooks, “Making Military Might: Why Do States Fail and Succeed: A Review Essay,” International Security 28:2 (Fall 2003): 149-191.

[9] Haun states that “Stalin allowed the Red Army the autonomy to develop into a professional force,” but according to Castillo’s theory, given the high level of regime control in the Soviet Union, an increase in military autonomy would generate a messianic military.

[10] Dale C. Copeland, The Origins of Major War (Ithaca: Cornell University Press, 2001); Dale C. Copeland, Economic Interdependence and War (Princeton: Princeton University Press, 2014); Sebastian Rosato, Europe United: Power Politics and the Making of the European Community (Ithaca: Cornell University Press, 2011); Roger D. Petersen, Resistance and Rebellion: Lessons from Eastern Europe (Cambridge: Cambridge University Press, 2006); Alexander B. Downes, Targeting Civilians in War (Ithaca: Cornell University Press, 2012); Keir A. Lieber, War and the Engineers: The Primacy of Politics over Technology (Ithaca: Cornell University Press, 2008).

[11] Jasen J. Castillo, Endurance and War: The National Sources of Military Cohesion (Stanford: Stanford University Press, 2014). Further citations in the text.

[12] S. L. A. Marshall, Men Against Fire: The Problem of Battle Command (Norman: University of Oklahoma Press, 2012); Edward A. Shils and Morris Janowitz, “Cohesion and Disintegration in the Wehrmacht in World War II,” Public Opinion Quarterly 12:2 (1948): 280-315; Omer Bartov, Hitler’s Army: Soldiers, Nazis, and War in the Third Reich (Oxford: Oxford University Press, 1992).

[13] Testimony of John Kerry before the Senate Foreign Relations Committee, April 22, 1971, Legislative Proposals Relating to the War in Southeast Asia, Hearings before the Committee on Foreign Relations, United States Senate, Ninety-Second Congress, First Session (April-May 1971), Washington: Government Printing Office, 1971.

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

[15] Castillo does not engage in an explicit treatment of Czar Nicholas II’s army in the book (presumably for lack of space), but it is the implicit foil for his chapter on Stalin’s army.

[16] Marc Bloch, Strange Defeat: A Statement of Evidence Written in 1940 (New York: Norton, 1968).

[17] See, e.g. Jack L. Snyder, The Ideology of the Offensive: Military Decision Making and the Disasters of 1914 (Ithaca: Cornell University Press, 1989), 70–90.

[18] David G. Herrmann, The Arming of Europe and the Making of the First World War (Princeton: Princeton University Press, 1997), 202-203.

[19] Carl Clausewitz, On War Peter Pare ted. (Princeton: Princeton University Press, 1976), 89.

[20] Condoleezza Rice, “The Problem of Military Elite Cohesion in Eastern Europe: The Case of Czechoslovakia,” Air University Review (January-February 1982), available online at:

[21] See for example, Teresa Rakowska Harmstone, et al., Warsaw Pact: The Question of Cohesion (Ottawa: Department of National Defence, 1984).

[22] It is well outside the scope of this essay but it is an interesting question as to why fourfold typologies are so common.  It could be that striving for parsimony pushes graduate students towards two variable explanations, for the typology is most common, it seems to me, in theories generated by dissertation writing (including my own).  But surely three variables is not absurdly complex?

[23] Caitlin Talmadge, The Dictator’s Army: Battlefield Effectiveness in Authoritarian Regimes (Ithaca: Cornell University Press, 2015).  As Talmadge is also a participant in this roundtable I will not discuss her work in more detail, but it is a very useful complement to Castillo.

[24] See Rod Nordland, “Troop ‘Surge’ in Afghanistan Ends with Mixed Results,” New York Times, 21 September 2012.

[25] Kristin M. Hall, “Army’s 101st pays high price for Afghan surge year,” Associated Press, 2011, See also Daniel Bolger, Why We Lost: A General’s Inside Account of the Iraq and Afghanistan Wars (New York: Houghton Mifflin, 2014).

[26] Hall, “Army’s 101st pays high price for Afghan surge year.”

[27] For example, even in 1970, as U.S. forces were withdrawing from Vietnam, four battalions of the 101st suffered over 70 killed and 400 wounded in just over three weeks in battles in the A Shau Valley.  See Keith W. Nolan, Ripcord: Screaming Eagles under Siege, Vietnam 1970 (San Francisco: Presidio Press, 2000). It is worth noting in passing that I bought this book from a PX at Camp Victory, Iraq in the fall of 2007. The armies of today no doubt reflect on the struggles and cohesion in their forebears.

[28] See for example Ned Parker, et al., “Special Report: How Iran’s military chiefs operate in Iraq,” Reuters, 24 February 2015.

[29] See Austin Long, “Whack-a-Mole or Coup de Grace? Institutionalization and Leadership Targeting in Iraq and Afghanistan,” Security Studies 23:4 (2014): 471-512.

[30] See, for example, Ann Harrison, “Openness and Growth: A Time Series, Cross-Country Analysis for Developing Countries,” Journal of Development Economics 48:2 (1996): 419-447.

[31] See for example Murong Xuecun, “A Land China Loves and Hates,” New York Times, 13 October 2015.

[32] A few exceptions are Dan Reiter and Allan Stam, Democracies at War (Princeton: Princeton University Press); Dan Reiter, “Nationalism and Military Effectiveness: Post-Meiji Japan,” in Creating Military Power: The Sources of Military Effectiveness, ed. Risa A. Brooks and Elizabeth A. Stanley (Stanford: Stanford University Press, 2007), 27-54; and Barry R. Posen, “Nationalism, the Mass Army, and Military Power,” International Security 18:2 (Fall 1993). Historians have addressed the issue of cohesion in particular cases, as discussed in Castillo, Endurance and War, 9-12.

[33] For example, Edward Shils and Morris Janowitz, “Cohesion and Disintegration in the Wehrmacht in World War II,” Public Opinion Quarterly 12:2 (Summer 1948): 280-315; and S.L.A. Marshall, Men Against Fire (Gloucester: Peter Smith, 1978).

[34] For just a few of the many examples, see Risa A. Brooks, Shaping Strategy: the Civil-Military Politics of Strategic Assessment (Princeton: Princeton University Press, 2008); Barry Posen, The Sources of Military Doctrine: France, Britain, and Germany Between the World Wars (Ithaca: Cornell University Press, 1984); Elizabeth Kier, Imagining War: French and British Military Doctrine between the Wars (Princeton: Princeton University Press, 1997); John Mearsheimer, Conventional Deterrence (Ithaca: Cornell University Press, 1983); Stephen Biddle, Military Power: Explaining Victory and Defeat in Modern Battle (Princeton: Princeton University Press, 2004); and Caitlin Talmadge, The Dictator’s Army: Battlefield Effectiveness in Authoritarian Regimes (Ithaca: Cornell University Press, 2015).

[35] The phrase comes from Ernest May, Strange Victory: Hitler’s Conquest of France (New York: Hill and Wang, 2000).

[36] On South Vietnam, see among many sources George Veith, Black April: The Fall of South Vietnam (New York: Encounter Books, 2012); and Stephen Hosmer, Konrad Kellen, and Brian Jenkins, The Fall of South Vietnam: Statements by Vietnamese Military and Civilian Leaders, report R-2208-OSD (HIST) (Santa Monica: RAND Corporation, December 1978). On Iraq, see Keren Fraiman, Austin Long, and Caitlin Talmadge, “Why the Iraqi Collapsed (and What Can Be Done about It),” Monkey Cage blog post at The Washington Post, 13 June 2014; and Caitlin Talmadge and Austin Long, “Why the U.S. (Still) Can’t Train the Iraqi Army,” Monkey Cage blog post at The Washington Post, 22 September 2015.

[37] See, for example, Reiter and Stam, Democracies at War; Biddle, Military Power; and Stephen Biddle and Stephen Long, “Democracy and Military Effectiveness: A Closer Look,” Journal of Conflict Resolution 48:4 (August 2004).

[38] Examples include Risa Brooks, “Making Military Might: Why Do States Fail and Succeed? A Review Essay,” International Security 28:2 (Fall 2003); Michael Desch, Power and Military Effectiveness: The Fallacy of Democratic Triumphalism (Baltimore: Johns Hopkins University Press, 2008); and Alex Downes, “How Smart and Tough Are Democracies? Reassessing Theories of Democratic Victory in War,” International Security 33:4 (Spring 2009).

[39] Chaim Herzog, The Arab-Israeli Wars: War and Peace in the Middle East from the 1948 War of Independence to the Present (New York: Vintage Books, 2005).

[40] Stephen Ambrose, Band of Brothers: E Company, 506th Regiment, 101st Airborne: from Normandy to Hitler’s Eagle’s Nest (New York: Simon & Schuster, 1992).

[41] Samuel Huntington, The Soldier and the State: The Theory and Politics of Civil-Military Relations (Cambridge: Harvard University Press, 1957).

[42] Talmadge, The Dictator’s Army, chapter 2.

[43] Lien-Hang T. Nguyen, Hanoi’s War: An International History of the War for Peace in Vietnam (Chapel Hill: University of North Carolina Press, 2012), 34-35; William S. Turley, “Civil-Military Relations in North Vietnam,” Asian Survey 9:12 (December 1969), 879-899; William Turley, “The Vietnamese Army,” in Communist Armies in Politics, ed. Jonathan R. Adelman (Boulder: Westview Press, 1982), 63.

[44] The Military History Institute of Vietnam, Victory in Vietnam: the Official History of the People’s Army of Vietnam, 1954-1975, trans. Merle L. Pribbenow (Lawrence: University of Kansas Press, 2002).

[45] Although the imagery is compelling, I am fairly certain he has not done this.

[46] Jasen J. Castillo, Endurance and War: The National Sources of Military Cohesion (Stanford: Stanford University Press, 2015), 130.

[47] David B. Ralston, The Army of the Republic: The Place of the Military in the Political Evolution of France, 1871-1914 (Cambridge: MIT Press, 1967), 373. Historians disagree about the scope of this revival and whether or not its effects were universally positive. But as Douglas Porch notes, “The French army’s stand on the Marne in September 1914 demonstrated that it contained enough competent generals, a skilled general staff and a complement of soldiers fired by patriotism and a spirit of sacrifice to stem the German advance.” See Porch, March to the Marne: The French Army, 1871-1914 (New York: Cambridge University Press, 1981), 213.

[48] Castillo, Endurance and War, 30 and 100.

[49] Eugen Weber, The Hollow Years: France in the 1930s (New York: W.W. Norton, 1994), 17.

[50] Ibid., 249.

[51] This argument reflects existing research finding that support for war in democracies decreases over time and when casualties mount. See D. Scott Bennett and Allan C. Stam III, “The Duration of Interstate Wars 1816-1985,” American Political Science Review 90:2 (June 1996): 239–257; D. Scott Bennett and Allan C. Stam, “The Declining Advantages of Democracy: A Combined Model of War Outcomes and Duration,” Journal of Conflict Resolution 42:3 (June 1998): 344-366. See also Adam Berinsky, “Assuming the Costs of War: Events, Elites, and American Public Support for Military Conflict,” Journal of Politics 69:4 (November 2007): 975-997.

[52] On the military preference for offense as a means to acquire more resources see Barry R. Posen, The Sources of Military Doctrine: France, Britain, and Germany Between the World Wars (Ithaca: Cornell University Press, 1984), 49-50; and Jack L. Snyder, The Ideology of the Offensive: Military Decision Making and the Disasters of 1914 (Ithaca: Cornell University Press, 1984), 24-29.

[53] Castillo, Endurance and War, 8-12 and 22-38.

[54] David M. Glantz and Jonathan House, When Titans Clash: How the Red Army Stopped Hitler (Lawrence: University of Kansas Press, 1995).

[55] Theo Farrell and Michael Semple, “Making Peace with the Taliban,”’ Survival 57:6 (2015): 79-110; Theo Farrell and Antonio Giustozzi ‘The Taliban at War: Inside the Helmand Insurgency, 2004-2012,’ International Affairs 89:3 (2013), 845-871.