The reasons for war, a survey with unflattering commentary

Not for the proud man apart   

From the raging moon I write   

On these spindrift pages   

Nor for the towering dead

With their nightingales and psalms   

But for the lovers, their arms   

Round the griefs of the ages,   

Who pay no praise or wages   

Nor heed my craft or art.

Dylan Thomas

Human animals face extinction from two things we do to ourselves:
   War and overpopulation.  Overpopulation is reviewed in another post.

Attempting to understand why we slaughter one another with such abandon, and have done for all of recorded history† is worthy, even if only becomes the attempt a part of the cultural background that impels us.

In the article reprinted here, two eminent scholars summarize the result of their investigation of the reasons we kill ourselves in War so freely.

The summery is interesting and compelling, if the authors' unstated assumptions are correct.  One assumption is that there is such an animal as a "rational man".  The criminal law and the tort law make the same assumption;  in both cases, an information revolution begin by Freud and Jung and others, which continues today, is ignored.  I cannot speak for economics, which is the authors' main thrust, but for law, the failure to consider current about human beings has lead to many bad decisions and bad results.  I suspect much the same is true or economics.

We have been three-quarters of a century without a world war, and without a nuclear war.  I suspect that billions of individual women and men, living in hundreds of communities, have individually decided that they will no longer support mass slaughter of our kind, and war as we have known it for all human history will morph into commercial struggles for power and wealth.

I also suspect that, while billions of us no loner support mass slaughter, most of us prefer a hierarchical structure to lead us, leading -- I fear -- to a universal dictatorship.

When billions of us, in time, watch power  accumulate in fewer and fewer hands, also reject government by monopoly or oligopoly, we will opt for a more open form of government, likely unknowable, currently.

Interlines to the summary are by me and are in italics,  and all images are posted by me and not by the authors.  Footnotes are omitted.  The full article is here.

† Here is a list of recent wars and casualties.  The list does not mention the wounded, nor the hundreds of millions of persons heart-broken and thrusting for revenge; or the hundreds of mullions of displaced persons.  See, e.g., here.

Wars and Casualties of the 20th and 21st Centuries
(slowly extending it back to the 19th century as i find data)by Piero Scaruffi
160 million people died in wars during the 20th century
(See also Modern Genocides)
TM, ®, Copyright © 2009 Piero Scaruffi All rights reserved.

1860-65: USA civil war (628,000) 1886-1908: Belgium-Congo Free State (8 million) 1898: USA-Spain & Philippines (220,000) 1899-02: British-Boer war (100,000) 1899-03: Colombian civil war (120,000) 1899-02: Philippines vs USA (20,000) 1900-01: Boxer rebels against Russia, Britain, France, Japan, USA against rebels (35,000) 1903: Ottomans vs Macedonian rebels (20,000) 1904: Germany vs Namibia (65,000) 1904-05: Japan vs Russia (150,000) 1910-20: Mexican revolution (250,000) 1911: Chinese Revolution (2.4 million) 1911-12: Italian-Ottoman war (20,000) 1912-13: Balkan wars (150,000) 1915: the Ottoman empire slaughters Armenians (1.2 million) 1915-20: the Ottoman empire slaughters 500,000 Assyrians 1916-23: the Ottoman empire slaughters 350,000 Greek Pontians and 480,000 Anatolian Greeks1914-18: World War I (20 million) 1916: Kyrgyz revolt against Russia (120,000) 1917-21: Soviet revolution (5 million) 1917-19: Greece vs Turkey (45,000) 1919-21: Poland vs Soviet Union (27,000) 1928-37: Chinese civil war (2 million) 1931: Japanese Manchurian War (1.1 million) 1932-33: Soviet Union vs Ukraine (10 million) 1932: "La Matanza" in El Salvador (30,000) 1932-35: "Guerra del Chaco" between Bolivia and Paraguay (117.500) 1934: Mao's Long March (170,000) 1936: Italy's invasion of Ethiopia (200,000) 1936-37: Stalin's purges (13 million) 1936-39: Spanish civil war (600,000) 1937-45: Japanese invasion of China (500,000) 1939-45: World War II (55 million) including holocaust and Chinese revolution 1946-49: Chinese civil war (1.2 million) 1946-49: Greek civil war (50,000) 1946-54: France-Vietnam war (600,000) 1947: Partition of India and Pakistan (1 million) 1947: Taiwan's uprising against the Kuomintang (30,000) 1948-1958: Colombian civil war (250,000) 1948-1973: Arab-Israeli wars (70,000) 1949-: Indian Muslims vs Hindus (20,000) 1949-50: Mainland China vs Tibet (1,200,000) 1950-53: Korean war (3 million) 1952-59: Kenya's Mau Mau insurrection (20,000) 1954-62: French-Algerian war (368,000) 1958-61: Mao's "Great Leap Forward" (38 million) 1960-90: South Africa vs Africa National Congress (?) 1960-96: Guatemala's civil war (200,000) 1961-98: Indonesia vs West Papua/Irian (100,000) 1961-2003: Kurds vs Iraq (180,000) 1962-75: Mozambique Frelimo vs Portugal (10,000) 1962-75: Angolan FNLA & MPLA vs Portugal (50,000) 1964-73: USA-Vietnam war (3 million) 1965: second India-Pakistan war over Kashmir 1965-66: Indonesian civil war (250,000) 1966-69: Mao's "Cultural Revolution" (11 million) 1966-: Colombia's civil war (31,000) 1967-70: Nigeria-Biafra civil war (800,000) 1968-80: Rhodesia's civil war (?) 1969-: Philippines vs the communist Bagong Hukbong Bayan/ New People's Army (40,000) 1969-79: Idi Amin, Uganda (300,000) 1969-02: IRA - Norther Ireland's civil war (3,000) 1969-79: Francisco Macias Nguema, Equatorial Guinea (50,000) 1971: Pakistan-Bangladesh civil war (500,000) 1972-2014: Philippines vs Muslim separatists (Moro Islamic Liberation Front, etc) (150,000) 1972: Burundi's civil war (300,000) 1972-79: Rhodesia/Zimbabwe's civil war (30,000) 1974-91: Ethiopian civil war (1,000,000) 1975-78: Menghitsu, Ethiopia (1.5 million) 1975-79: Khmer Rouge, Cambodia (1.7 million) 1975-89: Boat people, Vietnam (250,000) 1975-87: civil war in Lebanon (130,000) 1975-87: Laos' civil war (184,000) 1975-2002: Angolan civil war (500,000) 1976-83: Argentina's military regime (20,000) 1976-93: Mozambique's civil war (900,000) 1976-98: Indonesia-East Timor civil war (600,000) 1976-2005: Indonesia-Aceh (GAM) civil war (12,000) 1977-92: El Salvador's civil war (75,000) 1979: Vietnam-China war (30,000) 1979-88: the Soviet Union invades Afghanistan (1.3 million) 1980-88: Iraq-Iran war (435,000) 1980-92: Sendero Luminoso - Peru's civil war (69,000) 1984-: Kurds vs Turkey (35,000) 1981-90: Nicaragua vs Contras (60,000) 1982-90: Hissene Habre, Chad (40,000) 1983-: Sri Lanka's civil war (70,000) 1983-2002: Sudanese civil war (2 million) 1986-: Indian Kashmir's civil war (60,000) 1987-: Palestinian Intifada (4,500) 1988-2001: Afghanistan civil war (400,000) 1988-2004: Somalia's civil war (550,000) 1989-: Liberian civil war (220,000) 1989-: Uganda vs Lord's Resistance Army (30,000) 1991: Gulf War - large coalition against Iraq to liberate Kuwait (85,000) 1991-97: Congo's civil war (800,000) 1991-2000: Sierra Leone's civil war (200,000) 1991-2009: Russia-Chechnya civil war (200,000) 1991-94: Armenia-Azerbaijan war (35,000) 1992-96: Tajikstan's civil war war (50,000) 1992-96: Yugoslavian wars (260,000) 1992-99: Algerian civil war (150,000) 1993-97: Congo Brazzaville's civil war (100,000) 1993-2005: Burundi's civil war (200,000) 1994: Rwanda's civil war (900,000) 1995-: Pakistani Sunnis vs Shiites (1,300) 1995-: Maoist rebellion in Nepal (12,000) 1998-: Congo/Zaire's war - Rwanda and Uganda vs Zimbabwe, Angola and Namibia (3.8 million) 1998-2000: Ethiopia-Eritrea war (75,000) 1999: Kosovo's liberation war - NATO vs Serbia (2,000) 2001-: Afghanistan's liberation war - USA & UK vs Taliban (40,000) 2001-: Nigeria vs Boko Haram (1700) 2002-: Cote d'Ivoire's civil war (1,000) 2003-11: Second Iraq-USA war - USA, UK and Australia vs Saddam Hussein's regime and Shiite squads and Sunni extremists (160,000) 2003-09: Sudan vs JEM/Darfur (300,000) 2004-: Sudan vs SPLM & Eritrea (?) 2004-: Yemen vs Shiite Muslims (?) 2004-: Thailand vs Muslim separatists (3,700) 2007-: Pakistan vs PAkistani Taliban (38,000) 2012-: Iraq's civil war after the withdrawal of the USA (?) 2012-: Syria's civil war (130,000)Arab-Israeli warsI (1947-49): 6,373 Israeli and 15,000 Arabs dieII (1956): 231 Israeli and 3,000 Egyptians dieIII (1967): 776 Israeli and 20,000 Arabs dieIV (1973): 2,688 Israeli and 18,000 Arabs dieIntifada I (1987-92): 170 Israelis and 1,000 PalestiniansIntifada II (2000-03): 700 Israelis and 2,000 PalestiniansIsrael-Hamas war (2008): 1,300 Palestinians

The Reasons for Wars – an Updated Survey

Matthew O. Jackson and Massimo Morelli Revised: December 2009

Forthcoming in the Handbook on the Political Economy of War, edited by Chris Coyne, Elgar Publishing . . . .

Matthew O. Jackson: Department of Economics, Stanford University, Stanford, California 94305-6072, and external faculty member of the Santa Fe Institute, USA. Email: jacksonm@stanford.edu, http://www.stanford.edu/~jacksonm

Massimo Morelli: Columbia University and European University Institute. Email:

mm3331@columbia.edu http://www.columbia.edu/cu/polisci/fac- bios/morelli/faculty.html1. 

Introduction

Why do wars occur and recur, especially in cases when the decisions involved are made by careful and rational actors? There are many answers to this question. Given the importance of the question, and the wide range of answers, it is essential to have a perspective on the various sources of conflict. In this chapter we provide a critical overview of the theory of war. In particular, we provide not just a taxonomy of causes of conflict, but also some insight into the necessity of and interrelation between different factors that lead to war.

Is Dick Cheney a careful and rational actor?

Is Donald Rumsfeld?

Or George W. Bush?

Let us offer a brief preview of the way in which we categorize causes of war. There are two prerequisites for a war between (rational) actors. One is that the costs of war cannot be overwhelmingly high. By that we mean that there must be some plausible situations in the eyes of the decision makers such that the anticipated gains from a war in terms of resources, power, glory, territory, and so forth exceed the expected costs of conflict, including expected damages to property and life. Thus, for war to occur with rational actors, at least one of the sides involved has to expect that the gains from the conflict will outweigh the costs incurred. Without this prerequisite there can be lasting peace.Second, as cogently argued by Fearon (1995), there has to be a failure in bargaining, so that for some reason there is an inability to reach a mutually advantageous and enforceable agreement. The main tasks in understanding war between rational actors are thus to see why bargaining fails and what incentives or circumstances might lead countries to arm in ways such that the expected benefits from war outweigh the costs for at least one of the sides.

". . . and life." The last in a list, perhaps intended to be comprehensive, of factors to be considered when rationally deciding to engage in War.  Not listed are: 

maiming, burning,limbs twitted 'til sinus give way, insanity, often for what i left of life --

Goya wan't just making things up . . . .

Also not considered is starvation, the loss of habitat, he injury to many forms of non-human life, and, perhaps most relevant, grief that the death of a loved-one brings, with its resultant anguish, tears, and thrust for vengeance, perpetrating War. 

Đường phố lênh láng máu ở Syria | soha.vn

 The caption to this image is

"A man of sorrows and acquainted with grief.'

'"Remembrance Day takes on a profoundly different meaning

 once you have seen what people in the Armed Forces risk,

 their plight, their lives after servie,' 

writes Ben Quilty"

A good portion of our overview of the causes of war is thus spent discussing a framework of different bargaining failures. We emphasize that understanding sources of bargaining failure is not only useful as a categorization, but also because different types of failures lead to different conclusions about the types of wars that emerge, and particularly about things like the duration of war. We return to comment on this after discussing various reasons for bargaining failure. Below, we talk in detail about the following five reasons for bargaining failure:

1. Asymmetric information about the potential costs and benefits of war.

2. A lack of ability to enforce a bargaining agreement and/or a lack of the ability to credibly commit to abide by an agreement.

3. Indivisibilities of resources that might change hands in a war, so that not all potentially mutually beneficial bargaining agreements are feasible.

4. Agency problems, where the incentives of leaders differ from those of the populations that they represent.

5. Multilateral interactions where every potential agreement is blocked by some coalition of states or constituencies who can derail it.

To illustrate the importance of understanding which reason lies behind a conflict, note that if there is a lack of ability to enforce or commit to an agreement, then a war may last a long time. It will last until either one side has emerged victorious, or the situation has changed so that the costs of continued conflict have become overwhelmingly high for all sides. Such a lack of enforceable agreements is often one of the main ingredients leading to protracted wars. In contrast, suppose that enforceable and credible agreements are possible, but that the states start with asymmetric information, for instance, about the relative strength of one of the two countries. In such a case, there can be a bargaining failure which leads to war. However, in such a setting once war really begins the relative strengths of the countries can become clearer, and given that credible bargaining is possible and can avoid further costs of war the states could then reach an agreement to end the war. So, different durations of wars can correspond to different sources of bargaining failures. We expand on this below.

The chapter is organized as follows: For a clearer understanding of the boundaries of rationalist versus non rationalist explanations, we start by briefly discussing non-rationalist explanations in section 2. Section 3 provides a taxonomy of bargaining failures and how these relate to conflict; section 4 contains a discussion of which theories described in section 3 shed light on the observations of the democratic peace. In section 5 we report on the state of the literature on endogenous armaments and power and the implications for conflict and war.

2. The realm of rationality

Before proceeding to discuss various bargaining failures as causes of war, we discuss some of the alternative sources of conflict that are sometimes thought to fall into the realm of irrationality. We argue that many of these are more usefully viewed as being rational in nature, and hence the bargaining failure categorization still applies to many conflicts that are sometimes thought to be irrational. In order for our discussion to be as unambiguous as possible, we begin by clarifying what we mean when we dichotomize between rational and irrational actors. When we refer to a rational action by an agent we require that action to maximize the expected payoff to that agent out of the available actions and relative to the agent’s beliefs about the potential consequences of the actions. This does not necessarily require that the beliefs be accurate, nor that the payoffs of the individual agent correspond to what is best for the state or country that he or she might represent. This is a broader definition of a rationalist explanation than is usually understood in International Relations, where it is common to associate a rationalist approach to realist and neo-realist theories of conflict with unitary actors that are exclusively interested in material costs and benefits. Our broader definition should make it clear that what matters is that players, given the payoffs that they face from different outcomes, choose their actions to maximize it given their beliefs about the opponents’ actions, hence the qualification "material costs and benefits" is not necessary, nor it is necessary to confine the use of the rationalist approach to the world of unitary actors.

3 In order for this not to become a tautology, one has to be careful. An "irrational" act can always be rationalized simply by saying that it gave the agent taking it a high payoff for some intrinsic reason. Thus in order to have bite, the payoffs to agents for various actions have to have some natural specification. Although the distinction is thus partly semantic, or reliant on some idea of what natural payoffs should be, we still find it to be a useful dichotomy.

4

With this viewpoint in mind, let us discuss some causes of war that are often thought of as relying on some level of ``irrationality.’’ As we shall see, with our broad definition of rationality, even many of these may be interpreted as rational causes of war. This is not simply an issue of semantics, since the distinction has fundamental implications for how wars might be initiated, and if and how they can be avoided or terminated.

2.1 Religion

In principle, a war between two theocracies, or two states led by people of different religions, can be thought of as having rational explanations. It is a question of defining the objectives of the agents. For example, the goal might not be materially based, but might be based on the increasing the size of the population of one religion or eradicating another. In such situations, even with full commitment and bargaining opportunities, there might be no agreement that appeases an aggressor. One reason that one might place such motivations outside of the realm of "rationality," is that such objectives are often not put forth by a leader as if they are acting by choice, but instead leaders claim to be acting on behalf or under the direction of a higher being or religious code. Thus, the leaders in such settings do not necessarily view themselves as "optimizing" or "choosing" between paths but instead as following ordained directions. Perhaps even more importantly, from our perspective, such agents cannot be bargained with. That is, even if agreements are available and fully enforceable, such agents are driven by a specific goal that may be incompatible with the well-being or autonomy of another population. Thus, there is a critical distinction between a leader who is choosing and optimizing, even though the his or her rhetoric may be religious in nature, and a leader who believes that he or she acts simply as a channel for a higher being.

In this light, many wars that are thought of as being religious in nature can still be well-understood from a rational perspective. To make this point clear, let us discuss two prominent examples that are often considered to be at least partly religious wars: the crusades and the 30-year war.

5

Although the crusades were complicated by the fact that the aggressor was a coalition of national and sub-national armies, they fell under a common religious flag. Beyond the rhetoric, the commonality of interests within the Christian coalition can be doubted. As Fisher (1992) remarks about the interests of crusaders: "Undoubtedly, many of the Crusaders were inspired by a genuine religious motive next to their mundane concern for a share in the spoils. However, the idea of Christian unity failed again to achieve political reality. The Crusaders not only carved up the newly won territories in the East into petty principalities but also continued to struggle against each other in Europe. And they ultimately failed to hold the East precisely because they could not square their particular interests with the universal idea that had inspired them..." (Fischer 1992:438). "Thus, the politics of the Crusades, while showing that religious ideas can have some political effect, remained alliances circumscribed by the exigencies of power" (1992:443). Effectively, the crusades involved many factions and took place over many fronts and to a large extent involved attempts to gain or regain control of various territories, ranging from the Iberian peninsula, to Constantinople, to parts of the middle east including Jerusalem. The important aspect of this from our perspective is that the crusades took place at least partly due to a lack of ability to credibly commit to abide by agreements, to the multiplicity of factions involved on multiple fronts, and due to situations with great frictions in communication and in gaining information (e.g., see Runciman (1951-4)). Thus, the crusades can be partly understood from a combination of the rationalist perspectives that we discuss below.

Regarding the 30-year war, even though before 1618 there was an eruption of religious divisions within Europe emanating from multiple protestant reforms and movements, the religious motivation was used by some leaders to justify actions and to mobilize people, when again part of the instability derived from a multi-lateral power struggle and a lack of enforceable agreements. As argued by Gutmann (1988), a central reason for the failure of many settlement attempts was the difficulty of enforcing a new distribution of power that was so different from the official distribution of power defended by the papacy and imperial power. The Westphalia agreements that ended the war in 1648 cut the connections between some of the territorial and religious disputes, and the principles of autonomy and territory that were embodied in the agreement laid a foundation for modern states.

4 
To establish religious tolerance Catholics and Protestants were co-mingled within some of the same territories, and religious leaders were prohibited from having authority over people in separate territories. Thus, although the 30-year war involved religious motivations, the various factions were also motivated by territory, peace, and autonomy, and were eventually able to find a rather complicated agreement that was self-sustaining.

The long-standing conflict between Israel and the Palestinians could be viewed as another instance of a religious conflict that is often given non-rationalist explanations. However, it may more usefully be viewed through a rationalist lens. One of the central difficulties in resolving this middle-eastern conflict is in finding a stable agreement that is credible in the long run on behalf of the many different factions that comprise the two sides of the conflict. Even though the Oslo peace accords followed land for peace principles as one would expect in a rationalist dispute, when violence resumed the blame was given to "fundamentalism" on various sides (a typical non-rationalist explanation). The rationalist explanation for conflict based on multilateral bargaining, which we discuss below, is a more useful lens with which to view this conflict. In this case, both the Israelis and the Palestinians consist of many different constituencies and so although it appears to be a bilateral conflict it is in fact multilateral. In such settings, it can be that even with fully rational individual actors, agreements are not possible since the states end up being inconsistent in their decision-making as they are collectively aggregating the preferences of many different actors. This rationalist explanation is one that we discuss in more detail below.

2.2 Revenge

Revenge is another reason for war that one would instinctively place within the set of non-rationalist explanations of war. It is important, however, to distinguish an emotional version of revenge from a version of what someone might call revenge in the context of a repeated game: the punishment phase involved in trigger strategies of one kind or another. It is the emotional version that falls within the non-rationalist explanations. Revenge in emotional terms involves actions motivated exclusively by anger for a past action, and not motivated by the potential incentive consequences, nor decided ex ante as part of an optimal strategy. Wars driven by revenge are also rare, although famous examples include the motivation of the Achaeans' in the Trojan war, at least according to the description in the Iliad.

2.3 Ethnic cleansing and other ideological mass killings

As in our discussion of religion, one could in principle rationalize the incentives to eliminate another ethnic group or minority ideological group by a desire to obtain a larger share of the social cake, in the present and/or in the future (see e.g. Esteban and Ray (2008)). Such ideologies are generally uncompromising and not justified by reasoned choice but by appealing to other principles.

Hitler had the affirmation of the dominance of his race as a primary objective. However, as much as ethnic domination and insanity were part of Hitler's motivations, part of the understanding of the Second World War involves seeing why conflict was not avoided through concessions, and there rationalist explanations can help. As we mention below, for example, the failure of the Munich Agreement was due to credible commitment problems, and would have failed even if ethnicity and insanity were not in the picture.

3. Bargaining Failures and War

As mentioned in the introduction, we see two necessary ingredients for a war between rational agents. First, the costs of war cannot be overwhelmingly high. That is, for war to occur, at least one of the two parties must see a net potential gain from war under some circumstances. Second, there must be some impediment to bargaining, so that an enforceable and credible agreement cannot be reached. Effectively, rational decision makers weigh gains and losses from war given their objectives, beliefs, environment and constraints, and so if a mutually advantageous agreement is possible they should reach it. In an important paper, Fearon (1995) points out the criticality of bargaining failure for war. Basically, if rational agents come to the table with mutually consistent beliefs about the potential outcome of a costly war, then they should be able to reach a bargain to avoid it. In such a situation states can agree to split resources as they are expected to be split by a war, and then gain the extra surplus of the avoided destruction and costs of war.

Thus, to really understand the multitude of ways that wars may occur, it is illuminating to provide a taxonomy of bargaining failures and their roles in wars. As pointed out in Fearon (1995), there are various ways in which such bargaining might fail. It might be that the agents do not have the same beliefs or expectations about the potential outcome of a war. It could also be that they cannot commit to abide by an agreement and that there are no external means of enforcing an agreement. It might be that resources are indivisible and so there is no way to realize the split of resources that are expected as an outcome of a war. Beyond these three ways that are central to Fearon's analysis, we add another two. It might also be that the agents who bargain or make decisions do not have the same payoffs as the states at large, so that their incentives are distorted from what might be mutually beneficial to the populations. Finally, when considering multilateral bargaining, it might be that there is no outcome that is stable against coalitional deviations from groups of countries. In this section we elaborate on these five sources of bargaining failure, and we integrate this picture with some of the recent advancements in the theory of war.

3.1 Asymmetric information and bargaining failures

Asymmetries of information can arise from a variety of sources. It could be an asymmetry of information about the relative strengths of the countries either because of differences in what they know about each other's armaments, quality of military personnel and tactics, determination, geography, political climate, or even just about the relative probability of different outcomes.

The possibility of a bargaining failure due to asymmetric information has a solid foundation in economics, and was made very clear in work by Myerson and Satterthwaite (1983). To see the basic insights in the context of war, suppose that there are two countries and one of them, referred to as country A, has unknown strength. In particular, suppose that country A can either be strong or weak with equal probability in the eyes of the other country. Imagine that war involves a relatively small cost, that the victor in a war gains control of all resources, and that war results in one of the two countries conquering the other. Suppose that if country A is strong then it wins a war with probability 3/4 and if it is weak it wins with probability 1/4. So, in order to always avoid a war, an agreement must provide the strong version of country A with at least 3/4 of all resources less the cost of war (in expectation, presuming it maximizes expected payoff).

Now the asymmetry of information enters: a weak version of country A cannot be distinguished from a strong one by country B. Thus if the strong version of country A always gets at least 3/4 of the resources less the cost of war, then since a weak version of the country cannot be distinguished from a strong version by country B, a weak version of country A must also expect at least 3/4 of the total resources less the cost of war, as it can mimic a strong version of the country and get a high payoff without risk of war. This means that country B must get at most 1/4 of all resources plus the cost of war. If the cost of war is low enough, then the country B is better off simply going to war and taking its chances rather than reaching such an unfavorable bargain. This is obviously a highly stylized example, but it encapsulates the difficulties with bargaining in the face of asymmetric information. Generally, it may be difficult for a weak country to pretend to be a strong one, but there can still be some degree of asymmetric information across countries and even lesser asymmetries can make it impossible to find agreements that all parties will agree to in all circumstances.

It is important to note that imperfect information about the opponent's resolve or strength is a source of conflict that does not require any violation of common knowledge of rationality. The above reasoning is such that all the actors are fully rational, understand the setting, and fully comprehend all of its implications. It is also clear that the countries would like to avoid the difficulty. In particular, a strong version of country A would like to be able to distinguish itself from the weak version. If it could credibly demonstrate its strength, that would solve the problem. That is, if strength can be revealed peacefully and credibly (even at some minor cost), then there is a bargain which works as follows: if country A reveals strength, then it gets 3/4 of all resources and if it does not reveal its strength then it is presumed to be weak and only gets 1/4 of all resources. This solves the incentive problem as the weak version of country A can no longer pretend to be strong. Weakness is presumed unless evidence is presented to the contrary. This provides some insight into why countries might be willing to demonstrate arms (for instance publicly testing nuclear devices, holding military parades and exercises in observable settings, and so forth). There might be other settings where hiding strength is advantageous because bargaining is precluded, but in settings where binding agreements can be reached there are powerful incentives for the strongest types to reveal their strength to distinguish themselves from weaker types and to cement their bargaining position. Moreover, this is not limited to settings with just two potential strengths. Even with many different gradations of strength, the strongest wants to reveal itself, and then the next strongest will want to reveal itself, and so forth and this then unravels so all but the weakest types want to distinguish themselves. So this is robust to much richer information environments than the example above.

With such asymmetries of information, whether war will occur will depend on the extent to which the private information of individuals can be credibly revealed or not as well as how relevant the private information is to forecasting the outcome of a war. If it is really impossible to fully and credibly reveal information and such information is critical to predicting the outcome of a potential war, it can be that bargaining will fail and war must be expected with at least some probability. An early paper providing a model of war decisions with asymmetrically informed countries, and pointing out that an uninformed country may sometimes have to go to war to avoid bluffing behavior by an informed country, is Brito and Intriligator (1985).

The form of information asymmetry discussed above concerns potential outcomes of a war. A second information-based reason for a bargaining failure is that agents have inconsistent beliefs. For example, it could be that two states each are optimistic and are convinced that they will benefit from a war. In these cases war can erupt, as long as the inconsistency of beliefs is large enough to compensate for the cost of war. For instance, if both parties expect to win a war with a high enough probability, then there would not exist any agreement that avoids war

The possibility and examples of wars that are attributed to such miscalculations or errors due to lack of information or to different

A third form of information asymmetry concerns incomplete information about the motivations of other agents. Here it is believed that there is some probability that the other actor might be irrational.This includes spiraling models such as those discussed in Waltz (1959) and Schelling (1963), and more recently Kydd (1997). These ideas have been elaborated and extended upon by Baliga and Sjöström (2004) and subsequent works. The idea common to these works is that even a small probability of being faced by an armed irrational foe can lead a rational country to arm to some level. In turn, this now means that either a foe who is irrational, or a foe who thinks that I might be irrational will be arming, and this then leads me to arm even more, and this feedback continues to build. Depending on the specifics of the payoffs to arming and potential conflict, it can be that the rational countries each arm to very high levels and are ready to attack first because of the fear that the other side may attack first. In some cases, communication can help overcome this problem, since it can be in both countries’ interests to be known to be rational, but this depends on the specifics of the setting and the type of communication available,as Baliga and Sjöström (2009) show.

3.2 Commitment problems

Commitment problems are probably the single most pervasive reason for bargaining failure. This applies to many aspects of agreements that might avoid conflict, including promises to make future transfers and/or not to attack in the future. The implications of the inability to guarantee an agreement have been understood for centuries and, for instance, underlies the basic anarchic state of nature described by Hobbes (1651) in the Leviathan. As Hobbes states (1651, Chapter 13), "Because of this distrust amongst men, the most reasonable way for any man to make himself safe is to strike first, that is, by force or cunning subdue other men - as many of them as he can, until he sees no other power great enough to endanger him. This is no more than what he

needs for his own survival, and is generally allowed." Effectively there is nothing stopping someone from grabbing resources except fear of retaliation. Hobbes goes on to suggest that reasonable people can come to realize the inherent difficulties with anarchy and cede their rights to a Leviathan in order to live in peace. However, such social contracts do not generally appear in the international arena, and hence for an agreement to endure it has to be balanced in such a way as to be self-enforcing. In some cases, an outside authority, for instance an international organization such as the UN, can serve as an enforcer of an agreement, but the role of that international organization and its members’ incentives to really enforce the agreement are then part of a bigger picture where things need to be self-enforcing. Powell (2006) provides a rich set of illustrations of the pervasiveness of commitment problems.

What does self-enforcement entail? Effectively it must be that, in terms of our earlier discussion, the costs of war subsequent to whatever transfers of wealth or territory become overwhelmingly high. That is, for an agreement to be self-sustaining the states need to be sufficiently balanced in terms of strength and the allocation of resources, so that a war would not benefit any of the states in expectation. It can also be that even if one does not start at such a situation, then by giving up some resources one of the states becomes a less attractive target or a less threatening adversary and one reaches a situation where the costs of war outweigh the potential gains and so peace is self-enforcing. (See Bevia and Corchon (2009) for some discussion). Another way in which things might be made self-enforcing involves reputation. If a country faces potential conflicts with many other countries, then abiding by an agreement with one country can make it possible to credibly abide by agreements in other cases. Thus, it may be in a country's interest to abide by a collection of many agreements even when it might prefer to breach any single one of the agreements in isolation. Such self-enforcement involves embedding in a rich context and will depend on a variety of factors.

Let us also comment on some of the ways that a lack of commitment in agreements might lead to war.

3.2.1 Commitment not to attack after a received transfer or to deliver inter-temporal transfers

The most basic difficulty with a lack of commitment is the obvious one. A country delivering resources cannot trust that the other will not demand more or attack after receiving the resources. A notable example of such a failure of appeasement due to a lack of commitment is the Munich Agreement of 1938, after which Hitler invaded Czechoslovakia despite the agreement.

One idea that has been explored in terms of avoiding such difficulties is to make a series of transfers at a carefully determined rate over time that balances the incentives for conflict against anticipated future transfers. It is not always possible for such an approach to work, but it can in some circumstances, depending on which transfers are possible, how patient the countries are, how imbalanced they are, and how attractive or costly conflict is.

3.2.2 First Strike Advantages and Preemptive War

As the quote of Hobbes makes clear, one difficulty in attaining peace is that the natural anarchy in which international relations reside often leads to a first-strike advantage (preemption). That is, an element of stealth or surprise provides a significant advantage. If there were no first-strike advantage, and countries could have a well- founded expectation of the expected outcome of war, then there would be some mutual allocation of resources leading to a better outcome for all countries than war (presuming that the allocation does not further alter the expected outcome of war). That agreement becomes self-enforcing since it provides countries each with more than their expected resources after a war, and so war is worse for all involved in expectations. However, this presumes that the expected outcome of a war is the same independent of how the war starts. In many cases, the outcome depends on who initiates a war. A significant offensive advantage to war can lead war to be inevitable. As a simple illustration, imagine two evenly matched countries with an even split of resources and a cost to war. If war leads to an evenly matched outcome regardless of who attacks first or under what circumstances, then peace is self-enforcing. In contrast, if a country that strikes first gains a large advantage by doing so, and expects to gain resources with a high enough probability, then peace is destabilized. Each country would like to strike before the other, and also understands that the other also has an incentive to attack first, and so must react by expecting a war, and so war becomes inevitable. Various models of this appear in Powell (1991), Fearon (1995), Chassang and Padro i Miquel (2008), and Morelli and Rohner (2009).

3.2.3 Preventive war

Even in situations where countries are balanced in the short run, a country may fear that an opponent will become stronger over time and that the balance will be destabilized over time,, and may therefore wish to attack today to prevent being attacked by a stronger opponent in the future. Taylor (1954) is an early reference for this perspective, arguing that wars among great powers between 1848 and 1918 can be explained as preventive wars.

Interestingly, preventive incentives are not just an issue when countries are evenly matched and anticipate becoming unevenly matched in the future, but also when one country has a current arms advantage and worries that the other will catch up in the future and that the future situation will be unstable (possibly due to first-strike advantages, or some other considerations), and so wishes to attack while the balance is in their favor. This was an important concern during the early period of the cold war when the United States had nuclear weapon capabilities and the Soviet Union did not. There were debates about whether or not the U.S. should fight a preventative war during both the Truman and Eisenhower administrations. The fact that this did not happen has been argued to be due to a feeling that this was inconsistent with democratic principles (e.g., see the discussion in Silverstone (2007) and Levy (2008)), but from a purely rationalist perspective it might be that the fear of the future instability was insufficient to engage in a war at that time.

3.2.4 War as part of a dynamic bargaining process

Leventoglu and Slantchev (2007) report that almost 70 percent of conflicts end with a negotiated settlement, and almost no conflict ends with the complete elimination of one side, and hence the theory should explain why in many cases a commitment/self- enforcement problem disappears over time and a negotiated settlement eventually becomes feasible. They provide conditions, viewing war as part of a dynamic bargaining process, for a limited war to happen in equilibrium, and commitment to a negotiated settlement to appear after a period of war.

In summary, the pervasiveness of commitment problems comes from the lack of any external enforcement device in an international setting, and so any agreement is really only lasting if it is in the interest of all parties to continue to abide by it. A simple transfer of resources will not suffice unless it aligns incentives, or there are larger reputational concerns involved, or transfers are delicately arranged inter-temporally. There are many factors in such anarchic settings that naturally lead to instability such as preemptive and preventive motives, as well as the earlier mentioned asymmetries of information.

3.3 Indivisibilities and other physical impediments

Consider a situation where a fairly precisely balanced agreement needs to be reached in order to avoid conflict. If it is difficult to finely divide territory, or other natural resources in ways that strike the exact balance needed, that could lead to an inability to reach an agreement in the face of war. While indivisibilities are a seemingly important impediment to bargaining, Fearon (1995) dismisses them as a significant

explanation for war. Even if some resources are indivisible, it must be that there are no other resources that could be used to compensate. Agreements involving trade of large sections of land and money (e.g., the Louisiana Purchase), are plentiful, and the many dimensions through which wealth can be transferred from one state to another make it rare that a war occurs as a result from an inability to divide resources.

In terms of other impediments, delays in communication can make basic forms of bargaining difficult or impossible. While that is less of an issue in modern times, it was a substantial hurdle in times where armies might end up weeks or months in distance away from the leaders that commissioned them (as in the crusades). This leads to substantial delays in communication between the main parties involved in a potential conflict, and in such settings reaching an agreement that avoids conflict may be precluded even if such an agreement exists.

3.4 Agency problems

Even when decision makers are fully informed and have perfectly consistent beliefs, conflict may still be rationally chosen when there are differences in preferences between decision-makers and the rest of their country (a principal-agent problem). As explained in Jackson and Morelli (2007), when the decision-makers are biased relative to their countries war can occur, regardless of the availability of enforceable or binding agreements. The leader of a country might not face the same risks as the country's citizens, or it might be that the leader expects greater gains or glory from a war than the citizens.

Furthermore, as Jackson and Morelli (2007) point out, it can even be that a country would like to choose leaders that have different preferences from that of the country to improve their bargaining position. Overall the risk of war that this implies ex ante can be compensated by the ability of a hawkishly-biased leader to obtain better deals at bargaining tables. This means that even though democracies might be expected to have unbiased leaders who represent the preferences of the average citizen, unbiasedness cannot be guaranteed in democracies either..

Clearly, the ways in which leaders come to power differ across political regimes and this can affect the type of leader that emerges and the extent to which they represent the population as a whole. According to the selectorate theory in Bueno de Mesquita et al. (2003), democratic leaders need a larger coalition to support them relative to non democratic leaders. Keeping a larger coalition satisfied is more costly and hence losing a war is relatively more costly for democratic leaders, and generally makes them less prone to war.

3.5 Multilateral bargaining failures

As an illustration of the potential bargaining failures that arise in a multilateral setting, let us consider a simple three-state conflict. Suppose that there are three equally powerful countries with equal resources. Also suppose that if two countries cooperate, they can easily defeat the third with relatively low costs of war. In such a situation any pair of countries can expect to get almost all of the resources in the world by ganging up on the third. There is no bargain that is stable here.19 If the countries are about to sign an agreement, it must be that at least one of the countries gets at least a third of the total resources. The other two countries could gain by not signing the agreement, cooperating to defeat this third country and then afterwards splitting the resources evenly (and

citizens gain by having a leader whose bargaining position is more hawkish than they would have if they were bargaining on their own behalf.

What happens in settings with multilateral interactions will depend on the specifics of the bargaining process, the relative powers of different coalitions, and many other factors. What is clear, however, is that with three or more countries the fact that there is complete information, divisible outcomes, and an enforceable bargaining technology does not preclude war. This is an important and relatively unexplored territory in the theory of war, and given the innumerable wars that involve more than two states, understanding multilateral bargaining and war is an important area for future research.

Let us add a remark to this. Even when there are just two countries involved in a war, it might be multilateral considerations that derail peace. Although countries are sometimes discussed as if they are unitary actors, it is clear that they are composed of many actors with different objectives. As we know from the basics of collective decision making, an organization that is comprised of many actors does not necessarily act as if it were maximizing some objective function. Basic voting cycles can emerge and so a country composed of individually rational actors can exhibit intransitivities and other inconsistencies in its decision making that make the country difficult or impossible to bargain with.

4. Democratic Peace

As an example of how the various theories interact, let us consider a well-studied empirical regularity in international relations, namely the "democratic peace"; i.e., the observation that democracies rarely go to war with one another (e.g., see Doyle (1986) and Russett (1993)). The idea that incentives of aristocrats to go to war differ from that of democratic leaders is not new, and is well articulated by Kant (1795).

An important explanation of democratic peace is an agency one. As discussed above, Jackson and Morelli (2007) point out, when a leader has a disproportionately high share of benefits relative to costs from war when compared to the average citizen, then war can occur, but such a war will not occur if self-enforcing agreements are feasible and leaders are unbiased representatives. This "unbiased peace" result can be viewed as an explanation of democratic peace, since the checks and balances of a democracy can help reduce the chance of having a biased leader.

Conconi et al. (2009) extend this argument from Jackson and Morelli (2007), introducing an explicit election and reelection mechanism to control the bias of leaders. They show, theoretically and empirically, that the incentives of reelection can lead leaders to be unbiased in their decision-making, but that democratically elected leaders who do not face re-election can act similarly to autocratic leaders. Thus, the democratic peace observations are refined, and it is democratically elected leaders who face reelection who do not go to war with other democratically elected leaders also facing reelection. But autocrats or democratically elected leaders under the last term of a term- limit can diverge from the population’s interests and go to war. So it seems that a driving force behind the democratic peace is how a leader's incentives are kept in line with the population through potential reelection.

It is worth noting that the interactions between an executive’s behavior and election prospects can be quite complicated. For example, going counter to the incentives to avoid conflict when facing reelection, there are also ``wag the dog’’ sorts of situations, such as that described by Hess and Orphanides (1995, 2001), where an incumbent leader facing poor reelection prospects has greater incentives to initiate a war. The Hess and Orphanides (1995) explanation is that a conflict might reveal information about the leader’s abilities to the electorate that increases the probability of reelection. As Hess and Orphanides (2001b) suggest, such behavior can be correlated with recessions where an incumbent may be at a disadvantage.

Moving to the role of asymmetric information for the explanation of democratic peace, Fearon (1997) emphasized that so called "audience costs" (the cost of misrepresentations) are much higher in a democracy, and substantial audience costs can make signaling of information more effective in democracies, which in turn reduces asymmetries in information, and thus reduces the probability of war.

Fearon (2008) emphasizes another channel to rationalize democratic peace that involves commitment issues. The stronger country between two potential contenders usually has a higher GDP per capita. If it is democratic, then, even if the leader promises to a set of supporters some benefits from the war, it cannot avoid the possibility that eventually, once democratic rules apply to the unified country in case of victory, the GDP per capita of the winning country will go down. Hence voters of a richer democracy who believe that the unified country will lead to wealth redistribution should be against the war, and hence only weak contenders should remain interested in wars. However, weaker countries will generally have less interest in entering a conflict to begin with due to a low probability of success.

5. Endogenous Power

So far we have not talked much about the incentives of countries to arm. It is important to recognize that the probability of war depends on prior investments in arms, and that in turn the incentives to arm depend on how arms affect future incentives to go to war or to bargain. Thus, to fully understand decisions to go to war, such decisions cannot be divorced from the broader endogenous armament environment in which they reside.

There are studies of armament decisions in the case where conflict is inevitable (or bargaining is inevitable), such as that of Hirshleifer (1989, 1995) and Skaperdas (1992). The case were both decisions, whether to arm and whether to attack, are present

is analyzed in Powell (1993) and Jackson and Morelli (2009). The key difference in the analyses is the timing of the arming decisions. Powell's model leads to peace, and is one where countries move in alternating time periods and have their armament levels fixed for the intermediate periods. In such a setting, country 1 will set its arms at a level that it knows will be sufficient to deter the other country. The other country when called upon to move must then set its arms at a similar level, to deter future attacks when the first country can readjust its arms. This results in constant positive armaments and perpetual peace. In Jackson and Morelli (2009), the armament decisions are simultaneous, so that there is a sense in which countries cannot fully react to each other's armament levels but must anticipate them. In that setting, for a range of scenarios, countries randomize22 between a variety of strategies that must include hawkish, dovish, and deterrence armament levels. Of course, this is in the absence of commitment, as otherwise countries would sign binding agreements not to attack each other and the question of armament would become moot. The intuition behind why war is inevitable and some variation in arms levels necessarily result is fairly straightforward. A complete lack of arms on the part of both countries is not a stable outcome, since a country that anticipates that the other will be completely unprotected, would prefer to arm and attack (presuming that the costs of war are not overwhelmingly high, in which case perpetual peace and no arms are an equilibrium). Let us then consider the other extreme, where both countries arm to a high level, and mutually deter attacks. This also fails to be an equilibrium point. Given that there is a positive cost of war, it is true that if the countries are both heavily armed, then neither wants to attack the other. However, given the costs of war, deterrence is still assured if one country slightly reduces its arms. Given the savings of arms costs, then it cannot be an equilibrium for mutually high levels, as one country should lower its arms level to a slightly lower deterrence level. This incentive then ends up ratcheting down the arms levels, as it is always better to have slightly lower arms than the other country given that war will not occur if arms levels are close enough to each other. However, if we keep

ratcheting arms levels down so that the countries are not arming very much, then we return to the first reasoning that one of the countries should deviate to arm heavily and go to war. So, there is no stable pair of arms levels, and the equilibrium must involve some randomization, and over at least several types of arms levels. Jackson and Morelli also investigate comparative statics when there is a probability that the countries will have an opportunity to bargain (in a credible way) to avoid a war. Increasing the probability of a bargaining opportunity leads countries to make less use of deterrence armament strategies and more use of hawkish and dovish strategies, so the possibility of potential conflict increases. The idea is that as bargaining becomes more likely, deterrence strategies become less valuable, all else held equal, as do hawkisk strategies, while dovish strategies become more valuable. To ensure stability, one needs to increase the use of dovish strategies, which then reestablishes value to the hawkish strategies and increases their use, which then also maintains a reason to have at least some deterrence activity. The overall comparative statics that come out of this are that there is a lower probability of war due to the increased bargaining opportunities; however, there is a higher probability of war conditional on bargaining not being feasible.

Interestingly, peaceful outcomes are not necessarily the efficient ones in such endogenous-arms settings. Arms are wasteful, and so having many periods of peace but with costly armament levels can be worse than simply having an early conflict and then thereafter living in a unified country with peace without the need for arms.

Another interesting case is the one in which arms remain unobservable even after the investment phase has ended and war is about to start. Meirowitz and Sartori (2008) analyze this case, and they also show that war cannot be avoided even if bargaining technologies exist. In their case, the source of the positive probability of war is asymmetric information, whereas in Jackson and Morelli's observable case the bargaining failures fall under the category of commitment and enforcement frictions.

Beyond these models of endogenous power, there are also models such as that of Chassang and Padro i Miquel (2008). They do comparative statics in weapons stocks which sheds light on the incentives to arm and which sorts of arms and levels of arms countries might seek. They note that the advantage to a first strike affects the perceived value to the attacker, while the incentive to preemptively attack rather than risk being attacked is something that comes out of the payoffs of the potential target. Increasing the arms of just one country increases the advantage over an opponent in terms of a striking first which can increase preemption tensions. However, mutual increases in arms can lower the risk of suffering an attack and lower preemption tensions. They use this to note that extremely destructive weapons such as nuclear weapons can produce a more even balance and result in mutual deterrence, while arming with less destructive weapons might provide enough of an asymmetry that it results in incentives to strike first, and thus also an incentive to preemptively attack.

6. The duration of wars

As mentioned briefly in the introduction, part of the importance of understanding the various reasons for war is that different scenarios lead to different sorts of outcomes. To see this most starkly, consider a situation where a war starts due to a lack of commitment. In such a case a war can be protracted. A peace agreement only becomes attainable after the balance of power has shifted so that it becomes in both sides’ interest to agree to peace. This can take a long time. In contrast, if bargaining is possible, but fails due to asymmetric information about the relative strengths of countries, then a bargain should be reached as soon as the relative strengths of the countries becomes clear. This may take some time, but might happen much more quickly, and with lower costs, than it would take for the balance of power to shift significantly enough to lead to self-enforcement.

Exactly how long the war might last when there is no ability to commit can depend on many factors. In the case of asymmetric information, a model that offers predictions in this regard, where a country learns about another country's strength or resolve over time, is often referred to as a war of attrition (e.g., see Smith (1974) and Bishop and Cannings (1978)). The time at which a war of attrition ends depends on the specifics of the gain from winning, the costs of staying in the conflict, the patience of

the actors, and the level of uncertainty. The basic structure is one where two opponents incur costs at some rate per unit of time as long as they stay in conflict. The first one to give up loses and the other one wins. The uncertainty can be about the value to the other from winning, or the cost of conflict, or the patience. As the conflict goes on, it reveals that the other has not yet given up indicating a higher patience, lower cost, or greater patience. Eventually one of the two sides gives up.

It should be clear that our discussion applies to many sorts of conflicts and not just to inter-state wars. For example it applies to things like civil wars, coups and revolutions, and even strikes and other social and economic conflicts. Indeed, even some of the literature that is specifically aimed at understanding conflict in one arena can shed light on others. As an example, Acemoglu, Ticchi and Vindigni (2009) provide an explanation for the long duration of some civil wars. They explain that a government can fear having too strong a military, as a strong military can initiate coups especially in contexts where a government cannot commit to maintaining the resources directed to the military once a conflict ends. Understanding this interaction between a government and a military thus provides an additional lens into arming decisions, which then not only affects the number of conflicts that take place internally and their duration (as for instance a weaker military may take a long time to eradicate a rebel group), but then also has implications for the likelihood and potential duration of external conflicts.

7. Concluding remarks

We have presented a rich framework within which we can understand the prerequisites for war. Although our discussion has drawn mainly from the large literature on inter-state war, many of the same issues are at play in civil wars and other forms of conflict. Again, there must exist incentives for conflict and some barriers to the ability to reach an enforceable bargain. Some revolutions and coups arise from an agency problem either on the part of the current ruler or the leader of the attack. Some civil wars erupt because of ethnic or religious diversities manifesting themselves in the form of multilateral bargaining failures.

Although the theoretical understanding of the various causes of wars is developing well, and there are innumerable case studies of war and analyses of particular conflicts, systematic empirical work that analyzes the origins of wars across many cases is still relatively lacking. A richer understanding of the origins of wars would help further advance the theory, and would help in sorting more frequent and important causes from those which are less so; and ultimately would help in developing policies aimed at avoiding the costs of conflict.

Bibliography

Abreu, Dilip (1988): "Towards a Theory of Discounted Repeated Games," Econometrica, Vol. 56(2), 383-96.

Acemoglu, D., S. Johnson, J.A. Robinson (2001) "The Colonial Origins of Comparative Development: An Empirical Investigation," The American Economic Review, Vol. 91, No. 5, pp. 1369-1401.

Acemoglu, D., S. Johnson, J.A. Robinson (2005) "Institutions as a Fundamental Cause of Long-Run Growth," Chapter 6 in The Handbook of Economic Growth 1A, edited by P. Aghion and S. Durlauf, Elsevier.

Acemoglu, D., D. Ticchi, and A. Vindigni (2009) ``Persistence of Civil Wars,’’ mimeo: MIT

Alesina, A. and E. Spolaore (1997) "On the Number and Size of Nations," Quarterly Journal of Economics, Vol. 112 No. 4, 1027-1056.

Aumann, R.J. (1976) "Agreeing to Disagree," The Annals of Statistics, Vol. 4, No. 6 (Nov., 1976), pp. 1236-1239

Baliga, Sandeep and Tomas Sjöström (2004): "Arms Races and Negotiations," Review of Economic Studies, 71(2), 351-69.

Baliga, S. and T. Sjöström (2009) "The Strategy of Manipulating Conflict,"

27

Bevia, C. and L. Corchon (2009) "Peace Agreements without commitment ," mimeo: Universitat Autonoma de Barcelona.

Bishop, D.T. and Cannings, C. (1978) A generalized war of attrition. Journal of Theoretical Biology 70: 85-124

Blainey, Geoffrey. (1973): The Causes of War, New York: the Free Press.

Bloch, F., S. Sanchez-Pages, and R. Soubeyran (2006) "When does Universal Peace Prevail? Secession and Group Formation in Conflict," Economics of Governance, Vol. 7, 3-29.

Brito, D.L. and M.D. Intriligator (1985) `` Conflict, War, and Redistribution, ‘’ The American Political Science Review, Vol. 79, No. 4, 943-957.

Bueno de Mesquita, B. (1981): The War Trap, Yale University Press. Bueno de Mesquita, B., J.D. Morrow, R.M. Siverson and A. Smith (2003): Logic of

Political survival, MIT Press.

Bueno de Mesquita, E.A. (1999): "An Institutional Explanation of the Democratic Peace"

Bueno de Mesquita, Bruce., and Randolph M. Siverson (1995): "War and the Survival of Political Leaders: A Comparative Study of Regime Types and Political Accountability," American Political Science Review, 89(4), 841-55.

Cai, H. (2003) "War or Peace." in Contributions to Economic Analysis and Policy, BE Press, Vol 2, No 1, Article 2.

Chassang, S. and G. Padró i Miquel (2008) "Conflict and Deterrence under Strategic Risk," NBER Working Paper # 13964.

Chassang, S. and G. Padró i Miquel (2008b) "Mutual Fears and Civil War," BREAD Working Paper # 165.

Chassang, S. and G. Padró i Miquel (2009) "Economic Shocks and Civil War," forthcoming: Quarterly Journal of Political Science.

Chiozza, Giacomo and Hein E. Goemans (2004): "International Conflict and the Tenure of Leaders: Is War Still Ex Post Costly?" American Journal of Political Science, 48(3), 604-19.

Clausewitz, Carl von (1832)[1976]: On War, edited and translated by Michael Howard and Peter Paret, Princeton University Press.

28

Conconi, P., N. Sahuguet, and M. Zanardi (2009) "Democratic Peace and Electoral Accountability," mimeo: ECARES and CEPR.

Crawford, V. (1982), ``A Theory of Disagreement in Bargaining,’’ Econometrica 50,607- 637.

Downs, George W. and David M. Rocke (1994): "Conflict, Agency, and Gambling for Resurrection: The Principal-Agent Problem Goes to War," American Journal of Political Science, 38(2), 362-80.

Ellingsen, T., and T. Miettinen (2008), "Commitment and Conflict in Bilateral Bargaining", American Economic Review 98, 1629-1635.

Esteban, J. and D. Ray (2008) "On the Salience of Ethnic Conflict," American Economic Review, 98, 2185-2202.

Fearon, J.D. (1995) "Rationalist explanations for war," International Organization, 49:3, 379-414.

Fearon, J.D. (1996) "Bargaining over Objects that Influence Future Bargaining Power," mimeo.

Fearon, J.D. (1997): "Signaling Foreign Policy Interests: Tying Hands versus Sinking Costs," Journal of Conflict Resolution, 41(1), 68-90.

Fearon, J.D. (2008) ``A Simple Political Economy of Relations among Democracies and Autocracies,’’ mimeo: Stanford University.

Fisher, Markus (1992) Feudal Europe, 800–1300: Communal Discourse and Conflictual Practices. International Organization 46:427-66.

Garfinkel, M. (1990): "Arming as a Strategic Investment in a Cooperative Equilibrium," American Economic Review, Vol. 80-4, 50-68.

Garfinkel, Michelle R. and Stergios Skaperdas (2000): "Conflict without Misperceptions or Incomplete information: how the Future Matters," Journal of Conflict Resolution, Vol. 44-6, 793-807.

Gartzke, E.A. (1999): "War Is in the Error Term," International Organization, 53(3), 567-87.

Gutmann, Myron P. (1988): "The Origins of the Thirty Years' War," Journal of Interdisciplinary History, Vol. 18, No. 4, The Origin and Prevention of Major Wars (Spring), pp. 749-770 Published by: The MIT Press

29

Hess, G. and A. Orphanides (1995) ``War Politics: An Economic rational-voter framework’’ in the American Economic Review, 85: 4, 828-846.

Hess, G. and A. Orphanides (2001) ``War and Democracy,’’ Journal of Political Economy, 2001, vol. 109, no. 4, 776-810.

Hess, G. and A. Orphanides (2001b) ``Economic Conditions, Elections, and the Magnitude of Foreign Conflicts,’’ Journal of Public Economics, 80:1, 121-140.

Hirshleifer, Jack (1989): "Conflict and rent-seeking success functions: Ratio vs. difference models of relative success, " Public Choice, Public Choice 63: 101-112.

Hirshleifer, J. (1995) "Anarchy and its Breakdown," The Journal of Political Economy, Vol. 103, No. 1, 26-52.

Hobbes, T. (1651) Leviathan, or the Matter, Forme, and Power of a Commonwealth, Ecclesiasticall and Civil,

http://www.earlymoderntexts.com/f_hobbes.html [Focus on Chapters 13, 14, 15, 17-24, 26, 29, (they are short)]

Intriligator, Michael D. (1964) "Some Simple Models of Arms Races," Rand Report Intriligator, Michael D. (1975) "Strategic Considerations in the Richardson Model of

Arms Races," The Journal of Political Economy, Vol. 83, No. 2, 339-354 Intriligator, Michael D. (1982) "Research on Conflict Theory: Analytic Approaches and

Areas of Application ," The Journal of Conflict Resolution, Vol. 26, No. 2. Jackson, M.O. and M. Morelli (2007) "Political Bias and War," American Economic

Review, 97:4, 1353-1373. Jackson, M.O. and M. Morelli (2008) "Strategic Militarization, Deterrence and War

between Nations," http://www.stanford.edu/~jacksonm/gunsbutter.pdf

Jervis, Robert (1976): Perception and Misperception in International Politics, Princeton: Princeton University Press.

Jervis, Robert (1978): "Cooperation under the Security Dilemma," World Politics, 30(2), 167-214.

Jones, S. (1989), "Have your Lawyer Call my Lawyer: Bilateral Delegation in Bargaining Situations", Journal of Economic Behavior and Organization 11, 159-174.

30

Jones, D.M., S.A. Bremer, and J.D. Singer (1996) "Militarized Interstate Disputes, 1816- 1992: Rationale, Coding Rules, and Empirical Patterns," Conflict Management and Peace Science, VOL. 15, NO. 2, 163-215.

Jordan, James, (2006): "Pillage and Property", Journal of Economic Theory, Volume 131, Issue 1 , November, Pages 26-44.

Kaiser, D. (1990): Politics and War: European Conflict from Philip II to Hitler Kant, I. [1795] (1991) "Toward Perpetual Peace: A Philosophical Sketch," in Reiss,

Kant's Political Writings, Cambridge University Press, pp. 93-130. Kaplan, M.A. (1957) System and Process in International Relations, New York: Wiley.

Kirshner, Jonathan (2000): "Rationalist Explanations for War?" Security Studies, 10(1), 143-50.

Kocherlakota, N.R. (1996) "Implications of Efficient Risk Sharing without Commitment," The Review of Economic Studies, Vol. 63, No. 4., pp. 595-609.

Krasner, Stephen D. (1995): "Compromising Westphalia," International Security, Vol. 20, No. 3 pp. 115-51.

Kydd, Andrew H. (1997). "Game Theory and the Spiral Model," World Politics Vol. 49(3), 371-400.

Kydd, Andrew (2000) "Arms races and arms control: Modeling the hawk perspective," American Journal of Political Science, 44 (2), 228—244.

Lake, David A. (1992): "Powerful Pacifists: Democratic States and War," American Political Science Review, 86(1), 24-37.

LeBreton, M. and S. Weber (2003) "The Art of Making Everybody Happy: How to Prevent a Secession," IMF Staff Papers, Vol 50, No. 3, 403-435.

Leventoglu, Bahar, and Branislav Slantchev (2007): "The Armed Peace: A Punctuated Equilibrium Theory of War", American Journal of Political Science, vol. 51 no. ...

Levy, Jack (2008) "Preventive War and Democratic Politics" in International Studies Quarterly 52:1.

Luard, Evan (1986): War in International Society, Yale University Press.

Madarasz, K. (2009) "Bargaining and Conflict with Biased Beliefs," mimeo: Budapest University of Economic Sciences.

31

32 Mansfield, E.D., and J. Snyder (1995): "Democratization and the Danger of War,"

International Security Vol. 20 no.1, 5-39. McBride, M. and S. Skaperdas (2009) "Conflict, Settlement, and the Shadow of the

Future," mimeo: University of California, Irvine.

Meirowitz, A.H. and K. Ramsay (2009) "Institutions, Investing, and Fighting: A game- free result about the odds of war," mimeo: Princeton University.

Meirowitz, A.H. and A.E. Sartori (2007) "Secrecy and War: The Origins of Private Information", Quarterly Journal of Political Science.

Midlarsky, M.I. (): Handbook of War Studies, Ann Arbor: University of Michigan Press. Morelli, M. and D. Rohner (2009) ``Natural Resource Distribution and. Multiple Forms

of Civil War.,’’ mimeo: Columbia University. Mueller, J. (1990) Retreat from Doomsday: The Obsolescence of Major War, Basic

Books. Myerson, R.B., and M.A. Satterthwaite (1981) "Efficient Mechanisms for Bilateral

Trading," Journal of Economic Theory, Vol. 29. No. 2. Niou, Emerson M.S., Peter C. Ordeshook, and Gregory F. Rose (1989): The Balance of

Power; Stability in International Systems. Cambridge University Press. O’Neill, Barry (1990) ``A Survey of Game Theory Models on Peace and War,’’

Handbook of Game Theory, ed. By R. Aumann and S. Hart, Elsevier: NY. Piccione, Michele and Ariel Rubinstein (2007) "Equilibrium in the Jungle," The

Economic Journal, Vol 117-522, 883-96. Powell, R. (1993) "Guns, Butter and Anarchy," American Political Science Review, 87,

115-132;

Powell, R. (1999) In the Shadow of Power. Princeton: Princeton University Press;

Powell, R. (2004) "The inefficient use of power: Costly conflict with complete information," American Political Science Review 98(2), 231-242.

Powell, R. (2006) "War as a Commitment Problem," International Organization, 60, 169– 203.

Querou, N. (2009) ``Commitment and the optimality of negotiation or conflict,’’ mimeo. Querou, N. (2009b) ``Group bargaining and conflict,’’ mimeo.

Powell, R. (2009) ``Persistent Fighting to Forestall Adverse Shifts in the Distribution of Power,’’ mimeo: UC Berkeley.

Ray, D. (2007) A Game-Theoretic Perspective on Coalition Formation, Oxford University Press: U.K..

Ray, D. (2009) "Costly Conflict under Complete Information," mimeo: NYU. Richardson (1960) Arms and Insecurity, Boxwood.

Runciman, Sir Stephen (1951-1954) A History of the Crusades, 3 vols., Cambridge Press. democratic peace:

Russett, B. (1993) Grasping the Democratic Peace: Principles for a Post-Cold War World, Princeton University Press: Princeton N.J.

Russett, Bruce, Christopher Layne, David E. Spiro, Michael W. Doyle (1995), Correspondence section of International Security, 19(4), 164-84.

Sanchez-Pages, S. (2005): "Conflict as a Part of the Bargaining Process," Economic Journal, forthcoming.

Schelling, Thomas C. (1963): The Strategy of Conflict, London and New York, Oxford University Press.

Schelling, Thomas C. (1966): Arms and Influence, New Haven and London, Yale University Press.

Schultz, K. (1998) "Domestic opposition and signaling in international crises," American Political Science Review 92 (4), 829—844.

Schwarz, M. and K. Sonin (2008) ``A Theory of Brinkmanship, Conflicts, and Commitments,’’ Journal of Law, Economics, and Organization. 24(1): 161-183.

Silverstone, Scott A. 2007. Preventive War and American Democracy, New York: Routledge.

Slantchev, B. (2003) "The power to hurt: Costly conflict with completely informed states," American Political Science Review 97 (1), 123—133.

Slantchev, B. (2005) "Military coercion in interstate crises," American Political Science Review, 99 (4), 533-547.

Slantchev, B. (2007) ``Feigning Weakness,’’ mimeo: University of California at San Diego.

33

34 Skaperdas S. (1992) "Cooperation, Conflict, and Power in the Absence of Property

Rights," The American Economic Review, Vol. 82, No. 4, 720-739.

Smith, A. (1998): "Fighting Battles, Winning Wars; Opening up the Black Box of War: Politics and the Conduct of War." The Journal of Conflict Resolution, Vol. 42, No. 3, 301-20.

Smith, A. and A. Stam (2003) ``Bargaining and the Nature of War,’’ Journal of Conflict Resolution, 20:10, 1-30.

Smith, J.M. (1974) Theory of games and the evolution of animal contests. Journal of Theoretical Biology 47: 209-221.

Tarar, Ahmer (2006): "Diversionary Incentives and the Bargaining Approach to War," International Studies Quarterly, Vol. 50-1.

Taylor, A.J.P. (1954): The Struggle for Mastery in Europe, 1848 to 1918, Oxford, Clarendon Press.

Ticchi, D. and A. Vindigni (2009) "War and Endogenous Democracy," mimeo: Princeton University.

Wagner, R.H. (1986): "The Theory of Games and the Balance of Power," World Politics, 38(4), 546-576.

Wagner, R. H. (2000) "Bargaining and War," American Journal of Political Science, 44(3), 469-84.

Waltz, K. (1959): Man, the State, and War, Columbia University Press: N.Y.