The ORE of Politics will be available for subscription in late September. Speak to your Oxford representative or contact us to find out more.

Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, POLITICS ( (c) Oxford University Press USA, 2016. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy).

date: 17 August 2017

The Selectorate Theory and International Politics

Summary and Keywords

The Selectorate Theory is based upon one simple, perhaps even commonplace assumption: Once in office, leaders want to remain in office. They have a variety of tools to enhance their longevity in office, but the theory hypothesizes the leader’s allocation of two types of goods will be paramount in their efforts. One good is private, meaning that it is enjoyed by those to whom it is allocated and not to others. Such goods would include money, jobs, opportunities for corruption, but their hallmark is that they are not shared. These goods may be given to one individual or to a group, but they are not shared outside those to whom they are given. The second type of good is public and is shared by all those in the state. These goods would include potable water, clean air, education, and, importantly, national defense. There is little unique about the Selectorate Theory’s understanding of these goods, as they approximate ideas from economics.

The importance and values of these two goods depend critically on the political institutions of the state. The Selectorate Theory identifies two political institutions of dominant importance: The Selectorate, from which it takes its name; and the Winning Coalition. The former consists of all those people who have a role in selecting the state’s leader. This group may be large, as in the electorate in democratic states, or small, as in the case of an extended family or a junta. In unusual circumstances it can even be a group outside the state, as when a foreign government either imposes or influences choices made inside the state. The winning coalition may be large, but not larger than the selectorate, or it may be as small as an extended family or a junta, groups that essentially constitute the selectorate.

Variations in these two institutions can have important consequences for how the state conducts its foreign policy. For example, leaders in states with small winning coalitions should be able to take greater risks in their policies because if these fail, they will be able to mobilize and distribute private goods to reinforce their position. If these goods are not readily available, it is possible to purge non-critical supporters and redistribute their goods to others.

These institutions are also important in identifying the kinds of issues over which states are more or less likely to enter into conflict. States with small winning coalitions are more likely to enter into disputes over things that can be redistributed to supporters, such as land or resources. Large winning coalitions will have little use for such goods, since the ratio of coalition size and goods to be distributed is likely to be exiguous.

The Selectorate Theory also provides a firm analysis of the foundations for the idea of the Democratic Peace, which has been generally either lacking or imprecise.

Despite its clarity, some interpretations of the Selectorate Theory have led to mistaken inferences about what it says. We discuss several of these and close with a consideration of the need for improvement in the measurement of key variables.

Keywords: selectorate, leaders, survival, winning coalition, empirical international relations theory


The article presents an account of the Selectorate Theory’s (hereinafter ST) ability to explain how its concepts and inferences help in the explanation of international political phenomena. The article proceeds as follows. First, it offers a brief description of the essentials of the ST.1 Then it describes its use in the context of international politics before presenting an assessment of the correspondence of predictions from ST to empirical results in the literature. It then responds to some theoretical criticisms of the ST before concluding with an outline of further research that would mitigate an admitted shortcoming of the current state of ST research.

The Selectorate Theory (Short Course)

The Selectorate Theory, which was spelled out in greatest detail in The Logic of Political Survival (2003; herein after LPS), is built upon several assumptions. One significant part of political analysis explains how political actors, both groups and individuals, seek to obtain power, usually seen as office holding. The ST runs off the fundamental assumption that once political leaders have achieved office, they want to stay there and act to do so. In a sense, this is one of the recognized, but usually unarticulated, assumptions that motivate much of the analysis of the behavior of office holders. In order to gain office and stay there, leaders require the support of a group we term the winning coalition, which is drawn from a larger group, the selectorate, thus supplying the name for the theory. The latter consists of all those members of the polity who have the opportunity to select the leader: that is, the politically relevant portion of the population. If their choice wins, they become the winning coalition. This may seem an obvious general description of politics in democratic states, but the ST has several novel features that broaden its application to all states. First, the selectorate can be large, as in modern democratic states; or it can be small as in some monarchies (e.g., Saudi Arabia) and anywhere in the range of states between these two. Thus the types of political systems are not lumpy, but vary along two continua. Membership in it can be constitutionally defined, as in democracies; or it can be based on individual attributes, such as lineage (e.g., Saudi Arabia). It can be based on other factors as well, such as membership in a military organization, a tribal affiliation, or wealth, among others. It can even be limited to a group of religious leaders, such as the cardinals who vote on the selection of a pope. The important point is that all political systems have such a group. This group, of course, is not necessarily constant, but can change. For example, democracies have often enlarged the group of those who can participate in the selection of the leader through increasing the size of the electorate, but other changes can be more abrupt, such as when there is a sharp change in a political system through a coup, which raises the ratio of the winning coalition size to selectorate size either by shrinking the selectorate or expanding the coalition; revolution, in which previously excluded groups enter the polity as members of the selectorate; and, for some, as replacement members of the winning coalition.

In order to stay in power, political leaders rely on mixtures in the allocation of two types of goods. The first of these is public goods or policies from which no one in society can be excluded. These include, for instance, clean air, universal availability of public education, and national defense; these goods are public in that their distribution is intended to be widespread and inclusive. The second set of goods is private, and their distribution is targeted to individuals and groups in an exclusive manner in which the good is narrowly enjoyed. Such goods could include money, land, and opportunities for corruption, among others.

No political system functions purely by allocating only one type of good or the other. All polities spend resources on a mixture of the two types of goods. That is, no regime can maintain itself solely on the allocation of private goods, although a few have come close.2 By the same token, governments that provide large amounts of public goods also manage to offer key supporters private goods in the form, for example, of patronage, tax advantages, and other focused public policies. In short, all regimes have mixed policies. The nature of the mix of these goods is determined by the size of the winning coalition while the total amount of revenue spent is determined by the size of both the winning coalition and selectorate. For the purposes of exposition we stylize these groups into three types: Both the selectorate and the winning coalition are large, both are small and the winning coalition is small, and the selectorate is large. These would approximate in some respects political systems that are democratic, a monarchy, or a junta and an authoritarian one-party state, respectively. We say approximate those types because the lumpiness of the types can obscure important variations. We may tend to think of democracies as political systems with large selectorates and large winning coalitions, and in some cases this is justified. However, there are democratic political systems in which the selectorate is relatively large, but the rules of selection often have resulted in small winning coalitions. Consider the United Kingdom and it its first-past-the-post electoral system in which a candidate for parliament need not win a majority in their constituency. In fact, no U.K. government has held office with a majority of the popular vote since 1945, and over that period it was not unusual for governments to be formed on the basis of vote percentages in the low 40s, with as little as 25 percent being theoretically possible. Minority governments are usually based on even smaller winning coalitions, and some of them may have winning coalitions as small as those in some autocratic states such as Franco’s Spain or Mexico under the PRI.

The essential point is that the mix of goods the leader chooses to allocate will depend on which pattern of institutions obtains in the state. When the winning coalition and the selectorate are both large, leaders and their supporters can most efficiently maintain themselves in power through a mix of policies that will favor public goods; relying entirely, or even mostly, on private goods will not be effective since there are too many people and not enough goods. To be sure, when the winning coalition is large some individuals and groups will receive private goods in the form of coveted official positions through appointment or policies that favor their group, such as a protective tariff, but the overwhelming choice will be public policies intended to benefit all. When the winning coalition and selectorate are small, support for the leader can be maintained efficiently by a mix favoring the allocation of private goods, since there is little incentive to use public goods as doing so would reduce the resources available to continue the support of the winning coalition and provide a challenger with an opportunity to “buy off” one or more coalition members with larger private benefits. When the winning coalition is small and the selectorate is large, leaders are inclined, once again, to use private goods to maintain their support. It is worth noting that the ST anticipates that under this circumstance the allocation of the private goods to the supporting coalition will be minimal because its members will be particularly loyal since defection would force them out of a small favored group into a large less-favored group with a low chance of regaining their former position.

But what is the relationship of the ST to the study of international relations? If the study of international relations is defined as what happens between states, then the contribution of the ST is that it explains how those events are conditioned by the distribution of political power and the allocation of resources among groups within the state. That distribution will circumscribe the range and type of policies a leader will have an incentive to pursue so as to remain in office, as we explain in the next section.

The ST and IR

Over the last two decades there has been a significant turn inward in the study of international politics. Prior to this the study of international politics was dominated by the neorealist school that had as one of its fundamental premises that the state’s form of internal political and economic organization is irrelevant to explaining the choices leaders must make in a system that is characterized by anarchy and an uneven distribution of power (Waltz, 1979). As part of his argument Waltz denigrated what he termed the “reductionist” elements internal to the state, such as its type of political or economic system. The main argument of neorealism was brought into serious question by the observation that while democratic states had disputes and conflicts with each other, the level of violence between them did not reach a sufficient magnitude to be termed a war. This fact, of course, was disputed by neorealists (Gowa, 1999; Spiro, 1994), but in repeated studies the overall generalization has been robustly sustained. Contrary to neorealism, it is evident that variation in the political institutions of the state makes a profound difference in the policies that states adopt.3 Not only has this proven true regarding the use of force but also in explaining foreign aid giving and receiving; the credibility of threats to sanction; the incidence of coups and revolutions; the orientation toward free trade or protectionism; and a host of other topics in international relations.

But states do not act. People act, and in recognition of this the ST goes inside the state to examine the relationship of the interaction of a leader’s desire to remain in power with the relative sizes of the selectorate and the winning coalition. Put simply, the distribution of those two institutions will induce leaders to choose those policies that are consistent with the interests of the winning coalition. The impact of those institutions can be seen in the ST’s implications for the choice leaders make in the cases of serious conflict and war.

If leaders desire to stay in power, and if they realize that the policies they select will influence how long they remain, then they will choose their policies carefully. This simple rationale leads to the idea of selection effects, the primary assertion of which is that the conflicts and wars we see in history are not the product of a random process, but are those that were chosen to occur by the participants. Because of this, leaders will have an incentive to choose a mix of those policies that will be in accord with the interests of the members of the winning coalition. How is this manifested?

The ST argues that, compared to leaders who answer to a small number of supporters, those who answer to a large winning coalition require greater confidence in their prospects of victory to go to war; they make greater efforts to secure victory should they not prevail quickly; and they do not reduce their military allocations after war as rapidly. A large winning coalition pushes a leader to produce public goods such as success in foreign policy, while a leader with a small winning coalition can devote resources to private benefits for supporters to compensate for failures in foreign policy. These generalizations follow from the fact that, all else equal, the resources committed to private goods are necessarily spread more thinly as the winning coalition increases in size while the value of public goods to members of the winning coalition does not decrease as the winning coalition increases in size.

Leaders are assumed to maximize their prospects of retaining power and to leave as many resources over for their discretionary use as is compatible with keeping their coalition loyal. In this way, leaders minimize the risk that someone in their coalition will join a rival’s coalition and that they will be removed from power. In wartime, as in peacetime, the theory indicates that leaders in large coalition polities (like many democracies, Lake & Baum, 2001) seek gains primarily in the form of public goods (such as policy concessions and compliance from vanquished states), while in small coalition systems (like many autocracies) leaders primarily pursue a variety of private, material goods, such as money, land, monopolies, and rents; in the extreme, this is a pattern often referred to as kleptocracy.

The ST’s application to international relations is not limited to conflict and war, although both are key elements in what transpires between states, but, as we will discuss, has application to international trade (Milner & Kubota, 2005; Wruuck, 2013), environmental policy (Bernauer & Koubi, 2013), international treaties and norms (Morrow, Bueno de Mesquita, Siverson, & Smith, 2016), foreign aid (Bueno de Mesquita & Smith, 2009) and other aspects of foreign policy.

It is worth noting that although neorealism has been undermined, there is nothing in the ST that says that power is not important. In fact, it is the case that states must maintain their power as best they can because a failure to do so exposes the state to the predations of others, a situation that would endanger the winning coalition whether it is small or large. After all, it is through state power that leaders are able to produce the benefits they need to provide. Consider the following dilemma leaders may face that would be mitigated, or not, by the capabilities of the state.

The leader … who adopts policies that reduce the security of the state does so at the risk of affording his or her political opponents the opportunity of weakening the leaders grasp on power. Put differently, a leader’s search for the security of the state intertwines with the search for policies that will maintain the leader in power against domestic opposition. The desire to remain in power thus provides the linchpin between the threat and uncertainties of the international system and the inevitable imperatives of fending off the domestic opposition.

(Bueno de Mesquita & Siverson, 1995, p. 853)

The Empirical Assessment of the ST

In this section we will discuss the correspondence between the ST and data. The assessment of any scientific theory rests in large part on whether the predictions generated from the theory match the results of empirical research.

One of the advantages of the ST is its ability to generate hypotheses about how leaders will behave under specified circumstances. It does so with greater rigor than is found in most of what passes for theory in the study of international politics, by which we mean loose verbal arguments. The ST, on the other hand, offers the possibility of directly inferring hypotheses from assumptions and then testing them against data. Sometimes the hypotheses have a familiar place in the literature, which broadens the domain of the theory, but it is more impressive and a stricter test when the theory can be used to generate novel hypotheses that are not found in the literature and then to test these using well-established data. One of the first empirical papers relying on the ST did exactly that.

Drawing upon the logic of the ST, Bueno de Mesquita, Morrow, Siverson, & Smith (2004) generated five hypotheses about the behavior of states in international wars that are unrelated to any then extant hypotheses. While a full account of all the hypotheses is too lengthy for this venue, a discussion of the first gives an idea about the novelty of what they reveal. Bueno de Mesquita and colleagues hypothesize, “At the end of a war, larger coalitions demobilize … more slowly than smaller coalitions” (p. 375). On the face of it, this would appear to be inconsistent with what a reasonable person might expect; after all, large coalitions might be expected to yield to the wishes of those who want soldiers home. In actuality, the data analysis of the pattern of military spending after the war ends is consistent with the hypothesis. Why is this so? The logic of the argument, put briefly, is that when wars are fought to attain policy objectives, as is the case with large winning coalition states, in contrast to, say, land, it is necessary to devote additional resources to assuring that the conquered state implements the policies the victor wants.

The ST also provides a theoretical explanation for the democratic peace that had been lacking. In Ballots and Bullets (1999) Gowa casts a skeptical eye on the existence of the democratic peace. While the data analysis presented there is flawed (Siverson, 2000), a telling question is raised about the lack of any theory of the micro-foundations for the democratic peace. This was not an ungenerous assessment up to the time it was observed. However, at about the same time. Bueno de Mesquita and colleagues (1999) proposed an institutional explanation for the democratic peace that relied entirely on the logic of the ST. They reasoned that when democratic leaders are faced with the prospect of war they devote more resources to the war than autocrats. This is so because leaders in large states with large winning coalitions must produce successful policy to stay in power while autocrats can rely on sharing private goods to buttress their position if they lose. A consequence of this is that democratic states are uninviting targets. In addition, because a democratic leader needs success to remain in office, they choose their targets with a view that failure raises the chances of losing office. Autocrats, on the other hand, do not need a large advantage, because, again, they can use private goods to reward their supporters if their outcome is a loss. Because of the interrelationship of the democratic bias toward careful choice and the allocation of resources, democratic states rarely become involved in failed, lethal disputes.

This dynamic casts interesting light on the normative explanation of the democratic peace advanced by Maoz and Russett (1993), who argue that when a serious dispute takes place between two democratic states, the normal deliberative processes in each state delay mobilization and allow diplomats more time to find solutions. However, given the dynamic described in the previous paragraph, democratic leaders ought to have significant incentives to settle disputes quickly, since escalation is likely to be costly and may produce a damaging outcome for the leader. In fact, in a test of the ST’s ability to capture the expected selection effects, a study of the time from the onset of a dispute to its end, as given in the MID data set, democratic dyads settle their disputes significantly faster than other types of dyads (Bueno de Mesquita et al., 2004).

Neorealism is focused on states having only one interest; they seek to increase their security so as to survive in an anarchic world. Without gainsaying that power is important, it is also true that states have other interests. Neorealism is silent on these, but the ST is not. First presented in Chapter 9 of LPS, Morrow and colleagues (2006) elaborate the material presented there to draw out several hypotheses about the kinds of issues that will involve states of various types into international disputes. Leaders of states with small winning coalitions have incentives to pursue issues of conflict that involve territory since land can provide private goods in the form of a variety of resources. On the other hand, leaders of states with large winning coalitions and selectorates have no incentive to pursue conflict over territory since the private goods that the territory would convey offer little of value. These leaders will be motivated to pursue issues that involve the political policies of other states. This inclination has a significant entailment. If they defeat a state in a conflict, how does the victor ensure that the vanquished conforms to the victor’s policy preferences? The answer to this is to pursue the war to remove the other state’s leader and winning coalition and replace them with more pliant actors. Using a more nuanced data set than that originally reported in LPS, the results show that the larger the winning coalition, the less interest the leader has in territorial expansion and the more likely the leader will seek removal of the enemy leader as a solution to the commitment problem. In addition, the selection institutions of the opponent also influence a state’s war aims. The relative capabilities of the two states interact with their selection institutions. While, in general, states with large winning coalitions will not seek territory, they are more likely to seek the acquisition of territory when they are weaker than their opponent, since that would tend to reduce the imbalance in resources between them. Such states are also likely to try to overthrow the enemy leader when they are stronger than the enemy state because the costs of deposition are relatively low.

Critiques of ST

The ST has not been without its critics. Sharp criticism has been directed at the overall conception of the theory, and at least one critique has argued that the measurement of the winning coalition is just a crude measure of democracy. Although neither of these is focused on international politics, because they are directed at the heart of the theory and its empirical support, they warrant discussion. At least one criticism is directed at alleged shortcomings in the robustness of the ST to explain the conflict behavior of autocratic states. We begin by considering the more general criticisms.

Gallagher and Hanson (2015) examine the literature regarding Selectorate Theory, especially in the context of authoritarian regimes. They conclude that the theory is flawed on three dimensions. They believe that because authoritarian regimes lack formal institutions for leader transitions, that therefore the theory cannot explain politics in such settings. Further they contend that the theory fails to adequately address those who are not enfranchised. Finally, they are critical of the measures of coalition and selectorate size.

They are not the first to highlight weaknesses in the measures of winning coalition and selectorate size. Indeed, Bueno de Mesquita, Smith, Siverson, & Morrow (2003) emphasize that the measures are crude and that they and others should look for ways to improve measurement. Although Gallagher and Hanson (2015) seem to believe that such efforts have not been made, this is incorrect. For instance Bueno de Mesquita and colleagues completed an application of ST at the district level in Tanzania at the request of the World Bank, and Bueno de Mesquita and Smith (2015a) published an updated version of that study in which coalition size was measured in terms of the district-level vote support for the second largest party in each parliamentary district with results sufficiently informative that the World Bank presented them to the Tanzanian government. Measuring winning coalition size as the second-place vote in each parliamentary district in Tanzania, they found that public goods provision (e.g., distance to a health clinic) was substantially higher in large coalition districts than in small. Conversely, private goods allocation (e.g., marketable maize vouchers) rose sharply as the size of the winning coalition shrank. This helped the World Bank better understand why their aid programs in Tanzania did not meet their performance goals. Bueno de Mesquita and colleagues also undertook a pilot study in which they used an expert survey instrument to more precisely measure coalition and selectorate size. Lacking funds to pay experts to fill out the survey for all regimes, they mounted an online expert survey but without compensation; only a few country experts were willing to put in the time to complete the survey. While a full data set was not developed, the proof-of-concept has been well established for 10 or 15 countries and for 50 or so years. We return to this topic next.

Their other critiques reflect a poor understanding of the logic of Selectorate Theory. For instance, in defense of their claims about the inadequacy of the theory to explain authoritarian transitions, they cite Svolik’s (2012) finding that two-thirds of non-democratic leaders lose office by non-constitutional means such as coups and revolts as if this evidence is inconsistent with Selectorate Theory expectations. Yet, the theory has been used to explain mass uprisings, coups, purges, democratization, and oppression as vehicles of leader survival or transition (Bueno de Mesquita et al., 2003; Bueno de Mesquita & Smith, 2009, 2010, 2015b) with substantial empirical as well as logical success in authoritarian settings. They also cite evidence that within the range of coalition size short of democracy, the theory does not explain variation in leader tenure. Here the authors show a failure to understand the central equilibrium logic of the theory although Chapter 7 of LPS is dedicated to explaining why the theory should not be expected to explain variance in leader survival, but rather variance in what they do to survive. Although a more recent modification of the formal theory (2015b) allows leaders to lose office even in the absence of uncertainty, the LPS version does not, because that was not the theory’s central concern. As Chapter 7 highlights, if all leaders made the same policy and private goods choices, then the theory would show great differences in survival contingent on the content of those choices. But, since the theory assumes leaders are survival oriented, in equilibrium they make the resource allocation choices that maximize their survival given their institutional constraints: that is, the size of the winning coalition and selectorate. Hence, in equilibrium, as the coalition on which they depend gets smaller and the selectorate gets larger, they need to spend less to sustain coalition loyalty (opening the door to kleptocracy or civic-minded pursuits) and what they spend on the coalition heavily emphasizes private goods; when the coalition is large, then to survive as long as possible (in the absence of errors or shocks), they emphasize public goods and as the ratio of winning coalition to selectorate increases, they spend more, leaving less for their discretionary use, to survive. So, in equilibrium, leaders do what maximizes their survival. Therefore, the theory (absent uncertainty or shocks) is focused on the actions taken to survive and not survival time per se as thoroughly explained in Chapter 7 of LPS.

Questioning the role that democracy plays in the statistical analysis in LPS, Clarke and Stone (2008) present an extensive reanalysis of a substantial part of the empirical data analysis that supports the ST. Their basic argument is simple: In the statistical analyses present in LPS, democracy is used as a control variable to estimate the effects of the winning coalition and selectorate that are separate from democracy. The specification of this set of estimates is at question, since, as Clarke and Stone argue, the residualization of democracy “exaggerates the importance of their variables of interest by increasing the size of the estimated coefficients and decreasing the estimated standard errors” (p. 387). They contend the results reported in LPS are thus artefactual.

Morrow, Buena de Mesquita, Siverson, and Smith (2008) engage this argument by first admitting that Clarke and Stone are correct and, second, by offering a reanalysis of the models in question in a manner that removes the identified problem. The details of their approach are given in their response to Clarke and Stone (p. 396), but their reported results are exceptionally strong across all the analyses they report, with the size of the winning coalition being substantially important in the models. Their analyses, for example, lead to the conclusion that “the effects of W remain both in the theoretically predicted direction and strong when we control for elements of democracy … outside our theory.” They also note, “The elements of democracy outside our theory do not predict the provision of public goods and private benefits as consistently or as well as the size of the winning coalition does” (p. 397).

The final critique we consider is focused on the subject of international conflict. In an exploration of the conflict proclivities of authoritarian regimes, Weeks (2014) offers three reasons why the ST does not offer a coherent account for why leaders in a system with a small winning coalition will be more inclined than other types of states to enter conflict, a key prediction of the ST. The first of these objections asserts that in some political systems the loyalty principal, described above, will not operate because member of the winning coalition may have independent power bases that make them less vulnerable to removal and hence less loyal; they may be in a position to sanction the leader, as opposed to simply falling in line. However, members of the winning coalition who represent a bloc of individuals presumably use their position to provide benefits to that bloc, and exclusion from the winning coalition would end that and likely erode their own basis of support. Weeks (2014) offers examples of where the loyalty principal did not work and leaders fell from power. The case of Nikita Khrushchev is given by Weeks (2014) as an example of a failure of the loyalty to restrain other members of the winning coalition. This is an accurate description of the event, but there is nothing in the ST to suggest that loyalty is blind to the ambitions of the other members of the winning coalition who may have their own designs on gaining leadership. Can we imagine a small winning coalition whose members lack political ambition? Weeks’s position is based on a simple misconstruction of the ST. It is completely incorrect to assert that when the winning coalition is small and the selectorate is large, “members of the winning coalition have strong incentives to stay loyal to the leader regardless of his or her performance in providing public goods such as national security” (p. 7, emphasis added). First, under the institutional arrangements described, the winning coalition will have an interest in receiving private goods. National security will rise to be a dominant issue only when the welfare of the winning coalition is in doubt. Under such a circumstance there is nothing in the ST that would prevent the winning coalition from deposing the leader and making (or trying to make) a deal with those threatening the nation, much as when the Italian Fascist Grand Council dumped Mussolini and made arrangements with the Allies (Ikle, 1971).4 This is not unique. In 1964, when Guatemala was invaded by a U.S.- backed force, the Guatemalan army simply stopped fighting, then deposed the president and installed in his place the army commander, who tried unsuccessfully to make a deal with the invading force that would allow him to remain in office. Once a leader is shown to be vulnerable, the winning coalition may have little motivation in remaining loyal. Indeed, Bueno de Mesquita and Smith (2015b) have demonstrated the precise political conditions under which coalition members will remain loyal, will launch a coup, or will sit on their hands, allowing a revolution to succeed.

Moreover, if a leader fails to provide the necessary private goods, there is nothing in the ST saying that members of the winning coalition will not choose another leader. If the circumstances described by Weeks were obtained, no autocratic leader would leave office except by death. While some do leave office in that manner, many more are deposed, often by those close to them who seek a realignment in the distribution of private goods. Following a successful coup, for example, military budgets tend to increase (Leon, 2014)

A second argument offered by Weeks is that the ST is based on the assumption that “conditional on coalition size all actors perceive the world in the same way” (p. 8). If so, this assumption then neglects the uncertainty of foreign policy decisions. The text offers a footnote referring the reader to page 75 of LPS, where there is the assertion that the members of the winning coalition are assumed to be identical except in their affinities. The remainder of the paragraph in LPS containing that statement makes it clear that the members of the polity are assumed to be identical to avoid discussion of particular characteristics of individual political leaders and generate an abstract model covering the broad aspects of how political systems run. Of course, people in political life differ. Simplifying assumptions are necessarily used to gain generality. Still, LPS (p. 64 and pp. 473–474) touches on group or bloc interests as creating a mechanism to alter goods allocation and a means by which small coalition regimes might nominally appear to rely on a large coalition. Recent research, in fact, expands on the construction of the winning coalition to allow for special interest groups in a bloc voting environment in which goods are aimed at those groups both under conditions in which the group benefits are rival and when they are non-rival (Smith, Bueno de Mesquita, & LaGatta, 2017).

Finally, Weeks offers the assertion that the ST simply assumes a conflict exists between states and “does not explain how such disputes develop but rather how domestic institutions affect the decision to use force to settle disputes once they have arisen” (p. 8). To be sure, the ST, like many other theories of conflict, does not explain when a conflict will take place between two states, but as shown, it has quite a bit to say about what kinds of issues will motivate conflict over territory or policy and how domestic institutions influence the maintenance of international norms.

There are, however, two shortcomings of the ST in LPS. First, the data used to construct the two key concepts of the winning coalition and the selectorate are less strong than they should be, as the authors readily admit (Morrow et al., 2008) As they stand, they are built on several aspects of the Polity database that are used to measure the extent to which states are autocratic or democratic; both of these are concepts that are distinct from those in the ST. Why is this so? The answer is quite simple: Given available resources (namely, nil), it was the only practical way to proceed at the time the measures were needed. The choice could have been made to present only the theory and leave the problem of testing the hypotheses until later. But besides its logical consistency, confidence in a theory must ultimately be assessed by its correspondence to data, and we believed it was better to present the tests with data that was plausible, if not perfect. Given the crudeness of the data, it is testimony to the power of the theory that the empirical results have been so strong. Nonetheless, more can be done to sharpen these measures, a topic we will discuss shortly.

Second, we also need to acknowledge that while the ST is a very abstract theory, this virtue cuts two ways. Its strength is to be found in its ability to be applied to any political system and its public policy components, but its generality also means that the effects of institutional variation across political systems is obscured or at least blurred.

The Future of the Selectorate Theory

While there has been a significant amount of work inspired by or drawn from the ST, there are areas where more work is needed. The self-criticism at the end of the prior segment points to a glaring shortcoming on the ST. There are ways to overcome this, but they hinge importantly on the length of time period the data might cover. The longer the period, the more arduous the job and the greater the uncertainty about the accuracy of the measures, unless there is a willingness to accept 0, 1 distinctions, in which case it might not be a gain over current measures. If we suppose that data are to be used to test hypotheses, rather than to describe something, then a data set beginning in, say, 1945 or 1950 would provide a large amount of information. How might the data be gathered?

First, with sufficient resources, an effort could be mounted to create measures of the selectorate and winning coalition in a manner similar to what has been done with the polity data that has been remarkably successful in sustaining itself. While this idea may be attractive, the length of time it has taken to collect, refine, check, and report the data has been decades.

A second alternative might be faster and less expansive. Following the lead of Minhas, Ulfelder, and Ward (2015) and his associates it is possible to scrape data from the Internet that would provide measures of the selectorate and the winning coalition. They proceeded by drawing upon several recognized political system categorizations to provide them with agreed-upon typologies of political system types. Using these as ground truth, they examined news stories about those states, found modal terms, and then used those terms to identify governments with similar profiles. The difficulty with this is the lack of a ground truth estimate of the selectorate and the winning coalition. A way out of this is to draw a sample of state years and code them for the sizes of their winning coalition and the selectorate and then use those values as ground truth for the web scraping of the relevant content. Alternatively, data from the expert panel noted above, although limited, could also be used to estimate ground truth and thus enable analysis.

In the meantime, it should be noted that in recent research (Morrow et al., 2016) on international norms, the values of the winning coalition have been estimated in an improved, more nuanced way, and we continue to experiment with the rigorous development of estimates of coalition size and selectorate size.


Bernauer, T., & Koubi, V. (2013). Are bigger governments better providers of public goods? Evidence from air pollution. Public Choice, 156(3–4), 503–609.Find this resource:

Bueno de Mesquita, B., Koch, M. T., & Siverson, R. M. (2004). Democratic dyads and dispute duration: An empirical assessment of an institutional explanation of the democratic peace. Conflict Management and Peace Science, 21(Winter), 255–268.Find this resource:

Bueno de Mesquita, B., Morrow, J. D., Siverson, R. M., & Smith, A. (1999). An institutional explanation for the democratic peace. American Political Science Review, 93(December), 791–808.Find this resource:

Bueno de Mesquita, B., Morrow, J. D., Siverson, R. M., & Smith, A. (2004). Testing hypotheses about political institutions and war fighting. World Politics, 56(April), 363–388.Find this resource:

Bueno de Mesquita, B., & Smith, A. (2009). Political survival and endogenous institutional change. Comparative Political Studies, 42(2), 167–197.Find this resource:

Bueno de Mesquita, B., & Smith, A. (2010). Leader survival, revolutions and the nature of government finance. American Journal of Political Science, 54(4), 936–950.Find this resource:

Bueno de Mesquita, B., & Smith, A. (2015a). Tanzania’s economic and political performance: A district-level test of selectorate theory. In S. Wolf (Ed.), Politics in South Asia (pp. 31–48). Heidelberg, Germany: Springer Verlag.Find this resource:

Bueno de Mesquita, B., & Smith, A. (2015b). Political succession: A model of coups, revolution, purges and everyday politics. Journal of Conflict Resolution, 61(4) (September), 1–37.Find this resource:

Bueno de Mesquita, B., Smith, A., Siverson, R. M., & Morrow, J. D. (2002). Political institutions, policy choice and the survival of leaders. British Journal of Political Science, 32(4), 559–590.Find this resource:

Bueno de Mesquita, B., Smith, A., Siverson, R. M., & Morrow, J. D. (2003). The logic of political survival. Cambridge, MA: MIT Press.Find this resource:

Clarke, K. A., & Stone, R. (2008). Democracy and the logic of political survival. American Political Science Review, 102(August), 387–392.Find this resource:

Finer, S. E. (1990). The history of government from the earliest times. New York: Oxford University Press.Find this resource:

Gallagher, M. E., & Hanson, J. K. (2015). Power tool or dull blade? Selectorate theory for autocracies. Annual Review of Political Science, 18, 367–385.Find this resource:

Gowa, J. (1999). Ballots and bullets: The elusive democratic peace. Princeton.Find this resource:

Ikle, F. C. (1971). Every war must end. New York: Columbia University Press.Find this resource:

Lake, D. A. (1992). Powerful pacifists: Democratic states and war. American Political Science Review, 86(March), 24–37.Find this resource:

Lake, D. A., & Baum, M. A. (2001). The invisible hand of democracy: Political control and the provision of public service. Comparative Political Studies, 34(6), 587–621.Find this resource:

Leon, G. (2014). Loyalty for sale: Military spending and coups d’etat. Public Choice, 159, 363–383.Find this resource:

Maoz, Z., & Russett, B. M. (1993). Normative and structural causes of democratic peace. American Political Science Review, 87(September), 624–638.Find this resource:

Milner, H. V., & Kubota, K. (2005). Why the move to free trade? Democracy and trade policy in the developing countries. International Organization, 59(1), 107–143.Find this resource:

Minhas, S., Ulfelder, J., & Ward, M. D. (2015). Mining tests to efficiently generate global data on political regime type. Research and Politics, 2(3)(July), 1–8.Find this resource:

Morrow, J. D., Bueno de Mesquita, B., Siverson, R. M., & Smith, A. (2008). Separating the effects of W from the other elements of democracy. American Political Science Review, 102(August), 393–400.Find this resource:

Morrow, J. D., Bueno de Mesquita, B., Siverson, R. M., & Smith, A. (2006). Selection institutions and war aims. The Economics of Governance, 7(January), 31–52.Find this resource:

Morrow, J. D., Bueno de Mesquita, B., Siverson, R. M., & Smith, A. (2016). The domestic foundations of international norms: Non-intervention versus territorial integrity. Manuscript submitted for publication.Find this resource:

Siverson, R. M. (2000). Review of Ballot and bullets. International History Review, 22(September), 744–746.Find this resource:

Siverson, R. M. (1995). Democracies and war participation: In defense of institutional constraints. European Journal of International Relations, 1(December), 481–490.Find this resource:

Smith, A., Bueno de Mesquita, B., & LaGatta, T. (2017). Group incentives and rational voting. Journal of Theoretical Politics, 29(2), 299–326.Find this resource:

Spiro, D. (1994). The insignificance of the democratic peace. International Security, 19, 50–86.Find this resource:

Svolik, M. (2012). The politics of authoritarian rule. Cambridge, U.K.: Cambridge University Press.Find this resource:

Waltz, K. (1979). The theory of international politics. Menlo Park, CA: Addison-Wesley.Find this resource:

Weeks, J. (2014). Dictators at War and Peace. Ithaca, NY: Cornell University Press.Find this resource:

Wruuck, P. (2013). The political economy of anti-dumping. Heidelberg, Germany: Springer Verlag.Find this resource:


(1.) The Selectorate Theory is really a misnomer, since the size of the winning coalition might be argued to be more important. Terming it the W/S Theory would be more accurate, but clumsy.

(2.) For example, Finer (1990) asserts that by his calculations the Caliphate in the period 918–919 had central government revenues of 15.5 million dinars and expenditures of 10.5 million on the caliph’s personal household.

(3.) The exceptionalism of democratic states is not limited to the Democratic Peace but to other characteristics as well, such as in comparison to autocracies, their tendency to win wars (Lake, 1992), to suffer few battle deaths in wars they initiate (Siverson, 1995), and to settle their disputes more quickly (Bueno de Mesquita et al., 2004).

(4.) The July 20th plot on Hitler’s life was also motivated by a hope of forming a government that could make a separate peace with the Allies.