The Institutional Features of the Prisoners of War Treaties

The Institutional Features of the Prisoners of War Treaties by James D. Morrow Formal treaties on the treatment of prisoners of war (POWs) were negotiated and ratified during the 20th century. These treaties create a system for the treatment of POWs with universal and detailed standards and decentralized enforcement. I explain the form of the POW system as a rational institutional response to four strategic problems posed by prisoners of war: monitoring under noise, individual as opposed to state violations, variation in preferred treatment of POWs, and raising a mass army. In response to these four problems, neutral parties help address the problem of monitoring the standards. Ratification of the treaties screens out some states that do not intend to live up to the standards of the treaties. The two-level problem of state and individual violations is addressed by making states responsible to punish the actions of their own soldiers. By protecting soldiers taken prisoner, the agreements help states raise manpower for their militaries during wartime. The case supports many, but not all, of the hypotheses of the project. In particular, it suggests different strategic roles for membership and centralization in international institutions. The twentieth century saw the legalization of a system for the treatment of prisoners of war (POWs). This system had notable successes in improving the treatment of POWs in some cases, while in other cases, the system failed to induce states to abandon the abuse and murder of soldiers who surrendered to their forces during war. This paper seeks to explain the form of the legal rules and the system they induce to handle POWs, with a secondary concern of understanding why that institution fails and succeeds in different cases. International institutions vary widely in their forms. International law falls to the loose end of the spectrum of international institutions. Compared to other international institutions surveyed in this volume, the laws of war do not require recurrent decisions on proper policies as the International Air Transport Association did nor do they reach judgments about the facts in individual cases as dispute resolution panels do.i Instead, treaties on POWs and other laws of war set standards and prescribe mechanisms for their members to use when they are at war with one another. Enforcement of the standards is left to the parties themselves. Within this volume, this chapter covers an example of how normative values legalized into a treaty shape state behavior. It also addresses the question of this project; why do these treaties take the form they do? Informal understandings on the treatment of prisoners of war are as old as history. In this century, formal negotiations have developed those understandings into institutional arrangements that prescribe appropriate treatment and provide ways for states to verify that their soldiers taken prisoner are treated according to the agreement. This paper seeks to explain those arrangements as a rational response to the strategic problems that prisoners of war present. POWs and their treatment pose the following strategic problems: monitoring under noise, individual as opposed to state violations, variation in preferred treatment of POWs, and raising a mass army. Rational responses to these strategic problems characterize the POW system. The paper seeks then to understand not just the legal principles on the treatment of POWs but also the system of monitoring and enforcement built on those principles. These strategic problems are the specific realization of the independent variables in the Rational International Institutions Project (RIIP) in this issue-area .ii The first three strategic problems correspond to the following independent variables of the RIIP framework: uncertainty about behavior (UNCERT(B)), distribution under uncertainty about preferences (DIST and UNCERT(P)), and enforcement under uncertainty about behavior (ENF and UNCERT(B)), respectively. The fourth strategic problem concerns the relations between the state and its citizens that lays beyond the scope of the RIIP framework. I then test the hypotheses of the RIIP project about how these independent variables determine the membership (MEMB), centralization (CENT), and flexibility (FLEX) of the POW system. Briefly, I find that the POW system corresponds to a rational design to respond to the four strategic problems. The system has a general standard of treatment that avoids specific negotiations of the applicable standard for particular wars. Ratification of the standard screens out some states that have no interest in following the standard. The standards produce general reciprocal responses which are irregular and disproportionate to apparent violations of the standard. When the system breaks down, it fails at the individual level as well as the state level. The power to monitor the agreement is devolved away from the warring parties. The cases broadly supports the hypotheses of the RIIP project on membership (MEMB), centralization (CENT), and enforcement (ENF). The specific hypotheses on how uncertainty about preferences (UNCERT(P)) and distributional issues (DIST) affect membership, how uncertainty about behavior (UNCERT(B)) and the number of actors (NUMB) affect centralization, and how uncertainty about the state of the world (UNCERT(S)) and the number of actors affect flexibility are all supported in this case. The case does not support the framework hypotheses on the effect of enforcement on membership or centralization nor the effect of distribution on flexibility. Instead, this case suggests greater levels of complexity for these hypotheses. For example, the POW system is centralized in the determination of its standards but decentralized in its enforcement. The POW case also suggests that rational design of institutions depends on the strategic problems posed by an issue, and consequently the hypotheses of the RIIP project should hold only when the strategic problems in their logic are found in the case being examined. The next section discusses some general issues about the laws of war as an international institution. I then present the four strategic problems and rational institutional responses to each. A description of the POW system in practice allows a comparison of the predicted form of the institutions with their reality. The hypotheses of the RIIP project are evaluated for this case. I then examine some alternative arrangements for handling POWs and some alternative explanations. How Can the Laws of War Work? Political institutions must be self-enforcing to be sustained. In the language of game theory, institutions must form an equilibrium of a game, both in the sense that a particular institution induces equilibrium behavior and in the sense that the particular institution must be in equilibrium within the set of all possible institutions, including none.iii Kenneth Shepsle calls these the questions of institutional equilibrium and equilibrium institutions.iv The POW treaties are one example of the laws of war, prewar agreements about acceptable conduct during wartime. Such prewar agreements operate as institutions by shaping the decisions of actors during wartime. The agreements codify standards of treatment for POWs and rules for verifying that those standards are being carried out. The institutional equilibrium for the laws of war is the wartime behavior of states given the existing treaties. Such behavior covers not just treatment of POWs by states and individuals, but also how actors develop and use the system for the treatment of POWs. An agreed standard can shape what strategies (in the broadest sense of strategy as being all actions undertaken during war) states use to prevail in a war. A prewar agreement to abjure certain strategies can be upheld during war when reciprocity and audience costs make both sides unwilling to be the first to use a banned strategy.v The treaties of the laws of war are a public means for states to accept and understand their obligations during wartime. The agreement does not prevent the parties from acting in their best interest; instead, it sways the actors over what strategies they use in their pursuit of victory. If there are enforceable prewar agreements to restrict violence during wartime, then there are likely to be many different enforceable agreements. Precise standards matter, and there are many different precise standards that are enforceable. The choice among these many different standards is the question of equilibrium institutions. A rational approach to international institutions requires that existing institutions be Pareto optimal in the set of enforceable institutions; some party would be worse off if an institution was changed to another enforceable institution. Otherwise, no actor would object to a change of the institution, and it would not persist. Later, I will consider some alternative arrangement for handling POWs and compare them to the existing institutions to assist in the judgment of why the latter exists. The laws of war rely on reciprocity for enforcement. Violations of the rules may be deterred by reciprocal response. However, reciprocity can be implemented in many ways: what actions trigger a response, who should respond to an unacceptable action, and what responses are properly reciprocal rather than violations themselves? Reciprocity then requires shared understandings about appropriate treatment and responses that are institutional in nature. The understanding of how reciprocity will be employed on an issue shapes behavior on that issue (institutional equilibrium), that understanding can be changed if none oppose a change (equilibrium institutions). The laws of war can be thought of as the codification of the shared understanding at the heart of reciprocal enforcement of standards. Having discussed the institutional features of the laws of war in general, I now turn to the specifics of prisoners of war. Actors create institutions to address problems they face, and the character of those institutions reflect those problems and how they could be addressed. The next section lays out four strategic problems presented to states by the POW issue and discusses rational responses to those problems. Four Strategic Problems The issue of how to handle POWs raises four strategic problems that shape the institutions addressing their treatment. In this section, I describe each of these strategic problems and discuss the institutional response to each problem in the literature on rational institutions. This section details what institutional form we should expect on the POW issue and explains the logic underlying those forms. The four strategic problems, as we shall see, are closely related in practice; I separate them here in order to apply results in the literature that analyses these problems separately. Monitoring Under Noise Institutions built on reciprocity require the actors to monitor each other's actions so they can respond to violations of an agreement. Noise--uncertainty about behavior in the RIIP framework (UNCERT(B))--makes monitoring a significant issue for institutions because actors cannot determine exactly what one another has done. Instead, actors must draw inferences about others' actions from outcomes. Because outcomes result in part from factors outside the control of the actors, drawing such inferences is not straightforward. A classic example of uncertainty about behavior in economics is cartel enforcement if the members of the cartel can only observe the market price.vi They would like to know if any member of the cartel has cheated on the agreement by producing more their agreed share. However, others' production cannot be observed directly. If one member overproduces, the market price should drop. Production alone, however, does not determine price; a drop in demand could also cause the price to drop. Should the members of the cartel respond to a drop in price by raising their own production, the appropriate reciprocal response if a member has cheated on their agreement? Alternatively, problems of uncertainty about behavior can sometimes be addressed by the creation of neutral actor to collect and disseminate information.vii The information provided by the neutral actor can help alleviate some of the problems of monitoring, provided that such a neutral actor can be found and the parties have incentives to comply with its requests for information. The RIIP hypothesis on uncertainty about behavior and centralization is based on this argument. Noise arises in the POW issue for two reasons. First and foremost, states cannot observe in detail whether one another is complying with the standards of POW treatment because, by definition, POWs are in the hands of the other side. Japan refused the Red Cross access to American and Commonwealth soldiers taken prisoners in the first few months of the Pacific front of the Second World War. Consequently, it took months and even years before the home countries had a clear idea about how the Japanese were mistreating POWsviii. Here a neutral actor to collect and disseminate information can help. Second, much that occurs on a battlefield lies outside the view of commanders, and so they rely on reports from lower-level personnel about the conduct of their own soldiers on the field. In the case of atrocities, few soldiers are willing to report that they personally have committed such acts. Common accounts of summary killings of prisoners report that unspecified others carried out the act, and often such reports are indirect rather than eyewitness reports of the killing. Within camps, individual camp commanders and guards have some autonomy in how they operate. Factors outside the control of a detaining power may make it impossible for them to provide full support for POWs. The vagaries of war make it possible for a state to kill its own men taken prisoners inadvertently in the course of military operations. During the Second World War, submarines of the United States Navy sank Japanese ships transporting Americans held prisoners to Japan.ix A neutral actor is not likely to be helpful in addressing this source of noise because of the large amount of action to observe and the danger of combat to observers. Both of the above problems create noise; POWs may not receive treatment up to the standards of the treaties even though the detaining power has tried to live up to its treaty obligations. Uncertainty about behavior has consequences about resolving uncertainty about preferences. A government at war attempts to judge the preferences of its opponent--that is, whether the opponent intends to honor its treaty obligations--by observing the opponent's behavior. Uncertainty about behavior can make the inference from behavior to preference difficult. Errors of both overreaction and underreaction to reports of violations are possible under noise, and any system must address this inferential problem and the appropriate response in the face of it. The rational response to uncertainty about behavior requires the movement from direct and immediate reciprocity to more general reciprocity involving "bright lines" of acceptable outcomes.x Tit-for-tat responses to noise can lead to feuds of reciprocal punishments triggered by outside influences, rather than a defection from the agreement. Instead, actors should ignore small violations of the agreement and only respond to large violations of the accepted standard. A common standard of levels of violations deemed minor and acceptable allows the actors both to judge one another's likely responses to their own actions and avoid reciprocal feuds triggered by small amounts of noise. Because reciprocal punishment are not always carried out in response to violations that appear minor, such punishments must be disproportionate in order to carry the same deterrent effect as direct and immediate reciprocal sanctions. Then uncertainty about behavior has two primary effects on reciprocal enforcement of an agreed standard: one, the sides adopt a common standard of acceptable behavior to judge significant defections, and two, punishments become irregular and disproportionate to violations. Individual as Opposed to State Violations An effective agreement on the treatment of POWs must operate not only at the state level but also at the individual level. The greatest risk of being killed as a POW occurs between the time that the soldier indicates he wants to surrender to the time he enters a holding area behind enemy lines.xi Soldiers of even the best disciplined armies kill men attempting to surrender for a variety of reasons, including personal revenge, combat stress, and an immediate concern not to be bothered with the presence and care of prisoners.xii The use of surrender as a ruse for surprise attack occurs at times. Factors that could be described as cultural can also make the practical act of surrender difficult.xiii For instance, German military law forbid soldiers to surrender until all their ammunition was expended; American soldiers were often enraged by Germans who attempted to surrender once Americans had closed to their position under fire. Similarly, German soldiers often killed any American prisoner in possession of any captured German items under the assumption that the possessor had killed the German soldier who had been issued the item. The consequence of these difficulties and dangers of the act of surrender is that an effective agreement on prisoners of war must operate at the individual level as well as the state level. The POW treaties do specify some important elements of conduct on the battlefield, such as use of uniforms to identify soldiers and their army. Nevertheless, much of the practical agreement is ad hoc given the elements discussed above. The consequences of the failure of such an individual-level agreement are stark. Atrocity breeds retaliation. Further, the success of the state-level agreement plays a large role on the battlefield. Rumors about treatment of POWs spread rapidly within armies and change the willingness of soldiers to surrender. When POWs are treated poorly by a state, soldiers fighting against it are less likely to attempt to surrender, preferring to fight on even in unfavorable situations.xiv Such resistance makes the soldiers of the first state to be less likely to grant quarter to those who do attempt to surrender. This was the case during the Second World War on both the Eastern Front where Nazi Germany fought the Soviet Union and in the Pacific where Japan and the United States fought. In the former case, state policy reinforced the tendency on the battlefield towards no quarter. In the latter case, American policy did not encourage acts of brutality; however, the dynamics of the battlefield did drive widespread brutality by GIs and Marines.xv The possibility of individual violations of a treaty standard creates an enforcement problem (ENF) under uncertainty about behavior (UNCERT(B)) in the terms of the RIIP framework. Although the framework focuses on enforcement problems for actors directly involved in the institution, these individual level violations also pose an important enforcement problem. If individual violations are not restrained, they can lead to widespread violations and a collapse of the system of enforcement at the individual level, as described above. The institutional logic of controlling individual behavior parallels social institutions for the control of ethnic rivalry and conflict.xvi One way to control behavior between the groups entrusts individuals of each group to make appropriate responses to violations by members of the other group. The fear of an overall breakdown and the loss to all involved acts as a deterrent of bad behavior. Another approach centralizes the responsibility for punishing violators within their own group. The central authority within each group then wishes to punish violators from its own group to retain the cross-group agreement. Because a violator's own group may have better information about what was done and who did it and the ability to punish the violator personally, the deterrent is stronger when enforcement is devolved onto each group to police its own members. Further, the former system is very likely to break down in the face of substantial noise because violations breed cross-group retaliation. Under some circumstances, the threat of a complete breakdown can lead to effective restraints on violence across armies. The "Live-andLet-Live" principle found on some sections of the Western Front in World War I is an example of such even in the face of efforts by both sides' leadership to break down such agreements.xvii However, a system that devolves enforcement of individual violations controls the noise produced by such violations more effectively and is less likely to break down into general crossgroup violence.xviii An institutional response to the two-level problem devolves responsibility for punishing individual violations on the militaries of the violators. A devolved system of enforcement will not prevent all individual violations; rather, the prevention of widespread violations is the goal. The punishment of individual violations by the violator's own national military is a sign of the efficacy of such an institution. When soldiers are not held accountable for their actions or when state policy encourages atrocities, the battlefield is more likely to break down into general violations. This centralization of monitoring and disciplining of individual violations follows the same argument as the RIIP hypothesis on centralization and enforcement; centralization–here of the ability to punish individual violations–increases with the severity of the enforcement problem. Variations in Preferred Treatment of POWs A common standard of treatment of POWs requires agreement on many aspects of the handling of prisoners. However, states disagree about how prisoners should be treated. Each would like to see its own preferred standard be enforced and may choose to violate an agreed standard to do so. Other states are willing to live within an agreed standard even though they may prefer some other specific standard. At an extreme, a state may choose not to sign an agreement because it disagrees with specific provisions in the draft agreement; the Soviet Union did not sign the 1929 Geneva agreement on POWs because it allowed captor nations to treat officers and soldiers differently. In short, the adoption of any standard creates a distributional problem,xix and furthermore, differences in preferences about treatment create uncertainty about other states' motivations and so uncertainty about their future actions. In the RIIP framework, this is a distributional problem (DIST) exacerbated by uncertainty about preferences (UNCERT(P)). To give the reader a sense of the range of how states think POWs should be treated, consider the strategic advantages states at war can gain through their treatment of the men they take prisoner. Bad treatment of POWs they hold encourages soldiers of the opposing side not to take prisoners themselves, making it harder for your soldiers to surrender. Mistreatment does have consequences on the battlefield as rumors of how the other side treats POWs spread. Soldiers generally believe that reciprocity will hold; one German soldier reacted after watching the SS massacre about 300 Russian POWs, “It was already clear to us that it would have repercussions. That our prisoners [in Russian hands] would be treated in the same way.”xx States may wish to treat prisoners they take poorly in order to fortify their own soldiers' willingness to fight hard on the battlefield. In some cases, POWs have been recruited into the army of the detaining power, although coercion is often present in such recruiting appeals, particularly when joining the enemy army is a way out of terrible treatment in POW camps.xxi POWs are commonly used as a labor force, although they are banned by treaty from work in a state's war effort. Prisoners often welcome work, particularly agricultural work, as a way out of a dreary existence in camps. The question is what work and under what conditions. At the extreme, the Germans and Japanese during the Second World War used some POWs as close to slave labor in mines and railroad construction. The loss of life for those forced to work in those conditions was extremely high. German treatment of Soviet soldiers used as mine labor was so bad that the Nazis had to improve the prisoners' diet and accommodations, and limit their work hours just to get any valuable work out of them.xxii Useful military information can also be extracted from prisoners, both on the battlefield (where such acts are more common) and behind the lines. Keeping prisoners up to the standards of the POW treaties is costly to the detaining power, and so it is tempting for states to cheat on the standards. There are also important ideological and moral differences over the treatment of prisoners. Japan sought to inculcate their soldiers with the doctrine that troops who surrendered would be considered dead for all purposes by the home country.xxiii This doctrine helped to create the exceptional willingness to die in combat shown by Japanese soldiers during the Second World War. It also led to a general contempt towards soldiers of other nations who surrendered to the Japanese. In contrast, democratic states generally provide good treatment of POWs as an expression of the value they place on the protection of the individual, despite the political debate it triggers about whether POWs are being treated too well under the circumstances. Finally, racial attitudes direct state policy towards the mistreatment of POWs, most notably Nazi racist policies in Eastern Europe during the Second World War. This wide range of strategic consequences from the treatment of POWs leads to a wide range of plausible positions that states can take on the treatment of POWs. Some try to provide a reasonable existence to the men they take prisoner, while other states seize the advantages of mistreating POWs. State leaders make judgments about how their state will treat POWs given their state's strategic situation and values. In terms of institutional design, a state’s preferences reflect the considerations underlying these judgments. These differences in state preferences create the dilemma of inferring future actions from unknown preferences. State leaders can try to infer other's preferences from observed events. Often, actors would like information on others' preferences, and are willing to transmit such information about their own preferences, before acting. One institutional response to this dilemma is to create systems that allow states to signal their preferences to one another or force them to screen themselves in or out of a group. Typically, such signals or screens require costs to provide an incentive for actors with different preferences to separate themselves.xxiv Such costs could arise within the process itself though the consequences of separation, making costless actions--"cheap talk"--into effective signals.xxv Outside parties can then judge better the preferences, and likely future actions, of a state.xxvi Signaling or screening costs in international politics are commonly attributed to audience costs.xxvii The signal may set up dynamics by itself that lead to other actors imposing costs on the state leader who sent the signal. Such audiences could be external or internal. Other states might use violations of treaty obligations to judge the reliability of future promises; interested domestic parties could choose to remove their leader after he or she fails to uphold a state obligation.xxviii Such audience costs could be sufficient to make treaty obligations binding in some cases. A treaty would screen out some states that are unwilling to live up to the obligations of the treaty, and so inform other states that ratifying states were more likely to carry out their obligations under the treaty. The adoption of a single standard of conduct through a treaty creates a screen to help separate those states who are willing to live with the agreed standard from those who are not. Further, a uniform standard also solves the distributional problem that setting a standard poses. Once a standard is set, the question moves from “which standard is appropriate” to “which states are willing to comply with this standard?” A uniform treaty then addresses problems of both distribution and uncertainty about preference inherent in the question of variation of preferred treatment of POWs. The arguments above reflect the logics to that behind the RIIP hypotheses on how a greater number of actors cause centralization and how uncertainty over preferences produces restrictive membership. A single standard of conduct centralizes the judgment of who accepts a standard. Restricting membership in the system to ratifying states helps to reduce uncertainty over preferences for appropriate treatment of POWs. Raising a Mass Army Modern warfare is fought by mass armies, mobilized out of the citizenry of the nation. Conscription raises mass armies, and most armies since the Napoleonic Wars have relied on some form of conscription, particularly during wartime. Understandably, many able-bodied citizens are reluctant to face the risks of combat. Draft evasion and desertion are serious threats to raising a mass army and sustaining it in combat. The well-known logic of public goods applies here; all citizens enjoy the benefit of a victorious army, while those killed or maimed in combat and their families bear the cost. Nevertheless, large numbers of citizens are willing to fight for their country when drafted, and others are willing to volunteer to fight (although the likelihood of being drafted does drive some enlistments in wartime). Margaret Levy calls this behavior contingent consent.xxix Citizens are more willing to serve, and less likely to resist conscription, when they perceive that the state treats them fairly. Such fairness is judged by treatment of potential inductees, war aims, and citizens' overall view of the legitimacy of their government. Enforcement against those who try to evade the system helps to create a sense that the system treats all fairly. Quasi-voluntary compliance then combines the cooperation of citizens with enforcement against those who do not cooperate. All types of political systems rely on a combination of citizen compliance and state coercion to fill out their mass armies, although democracies rely on coercion less than other systems do. The implicit bargain between a state and its citizens extends to the treatment of citizens once induced into the military. Standards for the handing of prisoners of war should reflect the need of states to uphold their end of that implicit bargain. Institutionalizing those standards at the international level increases the credibility of a state's promise to protect its citizens serving under arms to the greatest extent possible in the vagaries of war. Clearly, all four of these strategic problems are interrelated. Uncertainty about behavior plays a key role in both the character of reciprocal responses and in the handling of individual versus state violations. Different degrees of commitment to their own citizens inherent in the bargain to serve in the military drives part of the uncertainty about preferences that makes a screening system useful. Pulling together the characteristics that would form a rational international institution on the POW issue, we find that there should be a common standard that is agreed in advance of conflict. This standard serves as bright line to determine what constitutes a violation and also as commitment by member states to good treatment of soldiers taken prisoner. Ratifying this standard before war begins serves as a screen on which states are willing to uphold that standard; states who refuse to ratify the agreed standard indicate their unwillingness to live up to it. One standard also solves the distributional problem posed by different state preferences for the treatment of POWs. Enforcement mechanisms for the standard must address the noise problem at both the state and individual level. Retaliation for violations is likely to be irregular and disproportionate to the violations. Evidence of the failure to uphold the standard should be appear at both between states and on the battlefield. During wartime, individual violators will be tried and punished only by their own military. Of course, many individual violations will go unpunished if not unidentified. The Prisoner of War System The Geneva Conventions of the treatment of prisoners of war is the centerpiece of the institutions to deal with POWs.xxx The rules codified in the treaties are applicable to all wars between members of the treaty. The treaties create a common standard which is subject only to limited and specified revision by individual pairs of warring states. For example, warring states may agree to exchange prisoners during wartime, but they are under no obligation to work out an exchange agreement. The POW treaties cover just about every facet of treatment of prisoners from the time of capture to repatriation after the war is over. Diet, discipline, the right to escape, the type of work that POWs can perform, and who can qualify as a POW are among the topics covered in the 1949 Geneva Convention. The detail that is specified on each of these issues has increased through the series of the three Geneva Conventions. The relevant treaties have been negotiated at multilateral international conferences open to representatives of all states. National ratification signals acceptance of the resulting standards, and records of ratification are centralized now in the United Nations. After each of the World Wars, the treaties have been renegotiated to account for experiences with the prior treaty during the wars. The series of POW treaties have been negotiated in open multilateral conferences. All states have been invited to participate, although the major powers understandably have dominated the negotiations. The Red Cross has also served as an important nonstate actor present at the negotiations. Enforcement of those treaties is decentralized. The warring parties alone have the ability to counter behavior that violates the treaties. Although member states at war have the power to prosecute and punish violators of the treaties from the other side, they rarely choose to, in part out of concern for retaliation against their own soldiers held captive. Even trials for criminal acts committed during captivity are treated cautiously. For instance, there were several cases where Nazis held prisoner in the United States killed other German POWs for acts they deemed disloyal to the Nazi regime. The United States did prosecute the killers, although it chose not to carry out the death penalties until after the war was over.xxxi Reciprocity is the unstated but recognized tool of enforcement. When the rules are generally observed in a conflict, protests of mistreatment are the first step and often suffice to remedy particular cases of mistreatment. Sometimes, very direct reciprocal sanctions are used by the parties. After the Dieppe Raid in 1942, a number of Germans taken prisoner by the Canadians during the raid were found with the hands tied, a violation of the rules. In response, the Germans then bound a specified number of Commonwealth soldiers they held prisoner, leading to counter sanctions by the British against Germans they held prisoners.xxxii In conflicts where major violations of the rules occur at the state level, reciprocity in general terms occurs. Both sides typically mistreat prisoners in these wars, with the notable exception being treatment of Japanese taken prisoner by the United States and Commonwealth forces during the Second World War. Breakdowns of the agreement of treatment of POWs often also leads to direct retaliation by soldiers of both sides on the battlefield; surrendering becomes a much riskier proposition than in other wars (not that the act of surrendering is ever free of the risk of killing by the captors). Because POWs are held behind enemy lines, the treaties provide for independent monitors of camp conditions. The Protecting Powers, the neutral states that operate as diplomatic liaisons for one warring state within the territory of the other, are the primary monitors of the agreement. Representatives of each Protecting Power have the responsibility to compile lists of soldiers taken prisoner, to convey mail to and from POWs, and to monitor conditions in camps, including discipline of POWs, and must be given free rein by states holding POWs to do these tasks. Once war begins, each Protecting Power establishes a POW bureau to act as a clearinghouse for information on prisoners. In practice, the Red Cross also performs many of the same roles, particularly when the sides find it difficult to appoint a Protecting Power. Its unique role as a humane agency that ministers to POWs, particularly those wounded in combat, places the Red Cross in the appropriate position to serve as a monitor. In either case, the collection of information is taken away from the national agents of either warring party. Member states also have the responsibility to educate their own soldiers about their rights as POWs and their responsibilities to troops of the other side trying to surrender under the treaties. Member states are also obligated to enforce the rules against their own soldiers who violate the provisions of the treaty. Such punishments do occur on the rare occasions that it can be clearly established that a soldier has violated clearly communicated military policy on the treatment of POWs. Member states also have the right to try violators from other states as war criminals, although the trials must be open to monitors from the Protecting Power and the accused must be treated as a POW until convicted. As noted earlier, such trials are rare. Membership in the treaties is open to all states by signing and ratifying the treaty. As is typical in international laws, ratifying states can object to parts of the treaty by filing a reservation at the time of ratification. They can also make clarifying statements about how they interpret parts of the treaty or object to the membership of another state in the treaty at the time of ratification. Joint membership by warring parties has been a strong signal that both parties will generally honor their obligations under the treaties. When the treaties have been broadly ignored, at least one warring side was not a party to the treaty, such as the Soviet Union and Japan in the Second World War and North Korea and China during the Korean War.xxxiii The notable exception to this generalization is the Iraq-Iran War, where both sides broadly violated the agreement despite both states being members.xxxiv Death rates of POWs held is one measure of adherence to the standards of the treaties because substandard treatment leads to death of prisoners. Table 1 presents death rates for the major combatants by front in the Second World War. The death rates on the Eastern Front and in the Pacific war were substantially higher than in Western Europe. The difference in death rates between Soviet versus American and Commonwealth soldiers held prisoner by Germany are quite stunning and are matched by descriptions of the treatment of these two types of prisoners. In some wars between one ratifying state and one state that did not ratify the treaty, the ratifying state generally upheld its obligations under the treaty in the face of violations by the nonmember state. Examples of this are the Pacific Theater of World War II where Japan did not live up to its treaty obligations while the United States did for the small number of Japanese taken prisoner. Table 1 about here Another sign of a breakdown of the agreement is retaliatory escalation on the battlefield of summary killings of men attempting to surrender. Reports of such killings are more common in the wars where the agreements have been violated, such as the Eastern Front) and the Pacific Theater of the Second World War.xxxv In all cases, all warring states engage in some violations of the treaty; it is not unusual for soldiers, even in the best disciplined armies, to shoot men from the other side while the latter are attempting to surrender.xxxvi The critical differences are whether the states involved engage in the violations as a matter of state policy or fail to control widespread violations of the agreements by their soldiers. Nazi Germany, the Soviet Union, and Japan all had army policies that encouraged killing of men attempting to surrender and to punish their own men who had surrendered. The best-known of these policies was the orders that the German army was to execute any Soviet Commissar captured during Operation Barbarossa.xxxvii In August of 1941, Stalin issued Order 270 that declared that Soviet soldiers who surrendered were "traitors to the motherland" and that they were subject to execution when returned and their wives to imprisonment.xxxviii Japanese military training emphasized that soldiers who surrendered would be considered never to have existed in the eyes of their families and the nation.xxxix Testing the Hypotheses Does the POW system match the expectations derived from models of institutional responses to the four strategic problems? As expected, the system creates a common standard with little room for ad hoc adjustment for individual cases. Punishment of apparent violations at both the state and individual level is irregular; many violations go unpunished. The punishment of individual violations predominantly falls on the state of those violators. The ratification process screens out some states who do not intend to follow the agreed standards during wartime. When state policy violates the standard or when individual violations either go unpunished or are encouraged, breakdown in the agreement on the battlefield also occurs. The act of surrender in such settings is unusually hazardous. Finally, the power to monitor compliance with the agreements is devolved to independent agents from the warring parties. There are two areas where the POW system may not match the expectations from the strategic problems. First, it is difficult to judge a "disproportionate" response to the atrocities committed on the battlefield by soldiers of a state that refuses to follow the standards. Were atrocities by United States Marines in the Pacific Theater of the Second World War a "proportionate" response to the Bataan Death March? Second, states of victims from the victorious states in the Second World War punished individual violators of the standard through the war crimes trials after the war. Such trials are legally permissible under the Conventions, but rarely done during wartime. The system does not rely on such postwar trials. However, those trials were ad hoc; they were not recognized as part of the system before or after they occurred. The International Criminal Court seeks to codify such postwar trials and integrate them into the laws of war generally. It is not clear yet whether such a system will work or even be adopted. Does the POW system match the hypotheses of the RIIP project? Of course, a single case can neither prove nor disprove those conjectures, and conjectures, by their very essence, are open to modification in the light of further evidence and argument. I use this case to shed some light on the conjectures concerning three of the dependent variables--membership, centralization, and flexibility. I refer to specific hypotheses using the codes in the introductory chapter; MEMB|ENF is the hypothesis about enforcement as independent variable relates to membership as dependent variable. Membership Membership of an institution, according to the project, should be determined by the severity of the enforcement and distributional problems and the level and type of uncertainty. The membership rules of the POW system are not restrictive, only ratification of the treaties is required for membership. Ratification of the treaties does appear to screen out some states which have no intention of following the standards, so the actual membership is not universal. One can imagine more restrictive membership tied to stronger enforcement of the system, as is the case in the recent Chemical Weapons Convention. There member states may not trade restricted chemicals, both poisonous ones and their chemical precursors, to nonmembers. This restriction provides a positive incentive to sign the treaty, which in turn has much stricter international inspections than earlier treaties. The case of the POW treaties has a mixed record on the conjectures on membership, primarily because the logic of that system differs from the logic underlying the hypotheses of the framework. There is an enforcement problem on POWs between states at war that have ratified the treaties. Then the framework expects that membership should be restrictive (MEMB|ENF) to exclude possible defectors and free riders. As pointed out above, the restrictions on membership are weak, only ratification is required. However, there really is not a question of free riding on the POW standards; the identification of a common standard and which states are willing to live by that standard is the issue. Membership in the POW treaties does reduce uncertainty by screening out some states that are not willing to live within the standards set out by the treaty, supporting the conjecture on membership and uncertainty over preferences (MEMB|UNCERT(P)). The conjecture that inclusive membership increases with the severity of the distributional problem (MEMB|DIST), is weakly supported here. Opting in is a signal that a state will abide by the rules of the system, which solves the distributional problem of agreeing on a particular standard. However, the logic of membership in the POW issue is different from the logic behind this conjecture in the theme paper. In the latter, inclusive membership allows for tradeoffs to solve distributional problems; here, the distributional problem is solved by only admitting states who signal their willingness to abide by the standards of the treaty.