Modalities of Power and Peace: The Implications of the Nature of the State


Studies of post-conflict states have examined the influence of regime type, political inclusion, and state strength on the durability of peace and conflict recurrence but have failed to consider how the type of power exercised within the state apparatus may impact the durability of peace in that state. This paper argues that the nature of the state plays a significant role in determining the likelihood of conflict recurrence within post-conflict states. The personalization or institutionalization of state power influences the strength of the state apparatus and its ability to address challenges to its authority and avoid conditions of multiple sovereignty that can lead to renewed civil conflict. A hazard analysis of 76 states that experienced civil conflict in the period of 1987-2010 establishes the existence of a relationship between the nature of the state and peace durability. Personalization of state power has a positive and significant effect on peace duration. Post-conflict personalist regimes are less likely to experience renewed conflict than their more institutionalized counterparts. These findings imply that in semi-authoritarian states personalization of power may increase stability and prospects of lasting peace. 

This material is based upon work supported by the National Science Foundation under Grant No. 1062343

Table of Contents: 


    The nature of the state has long been neglected in the realm of peace and conflict research. The proliferation of intrastate conflict following the end of the Cold War has led many scholars to note the tendency for conflicts to recur, trapping states in what Collier and Sambanis (2002) deem a “conflict trap.” As a result, a number of studies have examined the factors that may influence the risk of conflict recurrence, such as economic development, political institutions, and peacekeeping. Many studies have posited that democracy and the institutions of the state may affect the durability of peace. However, the relationship between the nature of the state, which influences how its organization and institutions operate, and the durability of post-conflict peace has not been examined. Despite the “Third Wave” of democratization following the fall of the Soviet Union, authoritarian and weakly democratized regimes have persisted and failed to follow predicted paths toward democratic consolidation. This suggests a need to examine the aspects of the state beyond the degree of autocracy or democracy in order to better understand why some states are at risk of recurring conflict while others emerge unscathed. Could the nature of the state, namely its degree of personalization or institutionalization of power within the state apparatus, affect the durability of post-conflict peace? This paper aims to extend the prior work regarding the forms of power employed within the state and to establish a relationship between the nature of the state and peace durability.

    Prior Examinations of Peace Durability

    Peace durability remains a focus of much study in the field of peace and conflict research. Factors influencing the durability of peace may arise from the origins of conflicts, the execution of conflicts, the termination of conflicts, and events during the post-conflict period. Conditions such as ethnic fractionalization, the presence of natural resources, low levels of economic development, and central government weakness may impact civil war onset. Given that these conditions may persist or even be exacerbated as a result of a civil war, it is reasonable to assume that they may also negatively impact peace durability (Elbadawi and Sambanis 2002). According to Collier and Hoeffler (2004), opportunities for rebellion (“greed”), arising from availability of financing (from lootable resources), low opportunity costs (a factor of economic development), and perceived military advantage greatly increase the risk for civil war. Elbadawi and Sambanis (2002) likewise demonstrate the importance of economic development in their finding that high levels of economic development can mitigate the risk of civil war due to ethnic fractionalization.  Concerning ethnic fractionalization, Collier and Hoeffler (2004) find that grievances provide a much less significant conflict risk factor, but ethnic fractionalization has a more significant effect on civil war onset than most grievances studied.

    Fearon and Laitin (2003), like Collier and Hoeffler, study the effect of opportunity and grievance on the risk of civil war outbreak. However, they posit that factors favoring insurgencies, mainly central government weakness, play the largest role in determining the risk of civil war. According to their findings, ethnic fractionalization has much less significance in determining the risk of civil war than per capita income. This concurs with the Elbadawi and Sambanis (2002) finding that economic development decreases the risk of civil war regardless of ethnic divisions. Fearon and Laitin (2003) also find that democracy and discrimination against linguistic and religious minorities have little effect on the risk of civil war. Political instability, mountainous territory, population size, and the presence of oil resources are shown to have a significant impact on the risk of civil war. Mason and Krane (1989) show that when government-sponsored violence becomes indiscriminate, most likely to occur under weak authoritarian regimes, it becomes rational for the non-partisan population to support the insurgency. This shift in support becomes rational because even non-support of the insurgency will not protect the subject population from indiscriminate state violence that makes no distinction between the partisan and non-partisan, whereas support for the insurgency may garner some form of protection from the rebel groups. Tilly (1978), examining the onset of revolution, likewise stresses that governance plays a role in conflict onset through his concept of “multiple sovereignty.” Multiple sovereignty is the condition in which groups mobilize and challenge the state and the government is unable to successfully suppress these groups, leading to the onset of civil conflict. While there is debate over the strength of relationships, ethnic fractionalization, lootable resources, levels of economic development, and government weakness, all appear to impact the risk of civil war onset and, by extension, post-conflict peace durability.

    Characteristics of the civil war also impact peace durability, namely its duration and cost in terms of lives lost. Luttwak (1999) argues that wars that end in one side’s victory are much more conducive to lasting peace. Mason and Fett (1996) demonstrate that duration of conflict impacts the likelihood of a negotiated settlement, with longer conflicts more often settled through negotiation rather than military victory. However, Mason et al. (2011) show that, while peace achieved through negotiated settlement is initially more fragile than that achieved through outright victory, the difference disappears within a short amount of time. Over a number of years, peace achieved through negotiated settlement becomes more stable than that achieved through government victory, which counters Luttwak’s thesis. Doyle and Sambanis (2000) find that wars ending in peace treaties are positively associated with peace-building success. Mason et al. also find support for a “war weariness” effect; wars of longer duration lower the risk of conflict recurrence. Their results suggest that more costly wars are more likely to recur, possibly due to more entrenched distrust between opposing sides. Doyle and Sambanis (2000) likewise find a negative relationship between high levels of battle deaths and peace-building success. Smith and Stam (2004) suggest that opposing sides learn their relative strengths through the execution of a conflict, and thus longer conflicts lead to the convergence of both sides’ beliefs regarding their relative strengths. Short conflicts deprive both sides of this opportunity and thus could increase the likelihood of conflict recurrence.

    A number of studies have examined the effect of the nature of the termination of conflicts on the post-conflict duration of peace. Beyond the difference between wars ending in settlements or victory, the kind of settlement reached, the unfolding of the peace process itself, and the presence of third parties impact the durability of peace. Hartzell and Hoddie (2003) found that the incorporation of higher numbers of power-sharing measures into peace settlements decreases the risk of civil war recurrence, and that the inclusion of a third-party guarantor likewise decreases the risk of recurrence. Stedman (1997) advocates for the proper management of spoilers (parties who feel threatened by the peace process and seek to undermine it) as a necessary condition for lasting peace. Fortna (2003; 2004) demonstrates that the presence of a peacekeeping mission positively influences peace duration; she additionally shows that peacekeeping missions do not “cherry-pick” the easy cases, thus substantiating their effectiveness.

    Beyond the peace process, the structure of the post-conflict state and conditions within that state may also affect peace duration. Democracy has been commonly associated with peace, and Walter (2004) finds some support for this idea in her findings that strong democracies have a much smaller likelihood of collapsing into conflict compared with other regime types (387). However, the effect does not hold true for weak democracies: both Hegre et al. (2001) and Mason et al. (2011) find that weak authoritarian regimes and weak democracies are both more prone to conflict recurrence than strong democracies and autocracies. Paris (2004) argues that liberalization should occur after “institutionalization,” the achievement of some degree of social order and economic recovery. Beyond institutional formations, some debate exists over the claim that geographically separating the two opposing sides after a civil war will positively impact peace duration. Kaufman (1996) promotes partition of the state as a solution to ethnic conflict that will better support a stable peace, however Walter (2006) suggests that partition may have adverse effects on peace by increasing the likelihood of new challenges to the state. Finally, Kang and Meernik (2005) address the need for investment and economic aid to promote post-conflict recovery following civil war. They emphasize the significance of economic development in positively impacting peace duration and demonstrate the positive effect of short-term foreign aid and democracy in achieving this aim.

    Prior Considerations of the Nature of the State

    A number of comparative studies have examined the state and its institutions; however, the majority of these studies have focused on regime types determined by institutional constructions and the effect of these systems on democratization. A few studies have examined the nature of the state, but these studies have generally focused on the transition to democracy and have been limited to specific geographic regions. Thus the impact of the nature of the state on peace and conflict has been largely unaddressed by the literature to date.

    Linz and Stepan’s (1996) seminal work discusses impact of prior regime type a state’s transition toward democracy. Their analysis of regime types includes a “sultanistic” party type, but their overall analysis of regimes focuses primarily on institutional structure and does not discuss the implications of regime type on the durability of peace. Geddes (1999), Levitsky and Way (2002), and Hadenius and Teorell (2007) also discuss the implication of regime type on a state’s potential democratic transition, each drawing from one another and creating their own modified regime typologies. However, while some of these regime types, particularly Geddes “personalist” authoritarian regimes, overlap somewhat with classifications based upon the nature of the state, they focus almost entirely on institutions. These typologies are then applied to democratization and do not touch upon their potential influence on peace durability. The most prominent works on regime type have concentrated on the relationship between prior institutional structures and paths toward democratization, failing to consider the impact of the nature of governance in these states and the potential implications of these regime typologies for peace and conflict.

    Recently, increasing scholarly attention has been paid to the nature of the state in the realm of democratic transitions. Theobald (1982) expands upon Weber’s notion of patrimonialism and discusses its modern forms and roots, though his observations of patrimonialism focus mostly on its links to an underdeveloped economy and seek to explain why patrimonialism still exists rather than exploring the implications of its existence. Bratton and Van de Walle (1997) extensively examine the nature of neo-patrimonial regimes in Africa, delineating types of neo-patrimonialism and varieties of institutional structures within neo-patrimonial regimes. According to their definition, neo-patrimonial regimes are characterized by “the incorporation of patrimonial logic into bureaucratic institutions” (62) through the domination of the state apparatus by an individual leader, the appointment of officials to bureaucratic office based upon personal relationships and loyalty, and the use of positions of power for accumulation of wealth and status. Ishiyama (2002) modifies and extends Bratton and Van de Walle’s typologies to the post-communist regimes of Central Asia and outlines potential paths for the transition to democracy within these states. Snyder (1992) also studies transitions from neo-patrimonial regimes, focusing specifically on neo-patrimonial dictatorships, and uses three key relationships—that of the ruler and the military, the ruler and elites, and domestic actors and foreign powers—to explain outcomes of transitions. These works, while greatly improving upon prior definitions of neo-patrimonialism, do not address the implications of this concept for peace durability. Ishiyama and Bratton and Van de Walle’s works focus on democratic transitions and are limited to particular geographic regions, and Snyder’s work, while positing a potential outcome of revolution, does not extend his study beyond this discussion.

    Corporatism is one of the few other theoretical forms of the nature of the state considered at length in current literature, and much of this literature is devoted to the debate over the very definition of corporatism. Schmitter (1979) provides one of the most cited and controversial definitions of corporatism and corporatist regimes. His definition has been modified and refined by both Wiarda (2004) and Adams (2004) to describe a corporatist form of governance as one characterized by the incorporation of interest groups into a strong directing state that often controls or regulates the activities of said groups. While these definitions greatly assist in conceptualizing an often amorphous term, none of these authors expand upon the implications of these corporatist regimes for peace durability, democratic transitions, or other phenomena. O’Donnell (1977) creates a definition and typology for a particular sort of corporatist regime, the bureaucratic-authoritarian state, and discusses the implications of this type of regime for transitions toward democracy; however, he does not discuss any potential relationship between the bureaucratic-authoritarian state and post-conflict peace durability. Additionally, all of the aforementioned works focus exclusively on Latin America for their case studies and exploration of the implications of corporatism.

    Goodwin’s (2001) study stands as one of the few works that addresses the nature of the state and its impact on revolutions, which can be extended to the relationship between the nature of the state and the onset of civil war. Drawing from Weber’s definitions of patrimonial and bureaucratic forms of domination, he develops a model of state vulnerability to revolution based on the degree of patrimonial or bureaucratic state organization, inclusivity or exclusivity of the state’s political regime, and state infrastructural power. He posits that weak, exclusive, patrimonial states are most susceptible to the development of revolutionary movements and defeat by said movements. Drawing on Goodwin’s work, this paper intends to move beyond the impact of the nature of the state on the risk of conflict onset to the implications of the nature of the state for durability of peace.

    Defining the Nature of the State

    Most discussions of the nature of the state begin with Weber’s tripartite conceptualization of legitimate authority, contrasting specifically his conceptions of traditional authority and rational-legal authority. From Weber’s traditional authority come the modern concepts of neo-patrimonial and Linz’s “neosultanistic” regimes in which power is derived from a particular leader, whereas bureaucratized or institutionalized regimes are embedded in the concept of rational legal authority derived from rules that constrain the state. The definitions of these types of regimes have been highly contested throughout the prior literature on the nature on the state. A number of scholars have focused on corruption, patronage, and clientilism as the most defining features of the neo-patrimonial state, leading to definitions that encompass nearly all nondemocratic regimes (Albrecht and Schlumberger 2004; Eisenstadt 1973; Erdmann and Engel 2007; Lemarchand and Legg 1972; Theobald 1999). However, as Mesquita et al. (2000, 2003) demonstrate, any regime with a small winning coalition, a key feature of most autocratic regimes and some nonconsolidated democratic regimes, will display some degree of patronage as private goods and political appointments provide a means of buying coalition loyalty at a lower risk than enactment of possibly unsuccessful public policy. Therefore, nearly all types of autocratic and semi-authoritarian regimes will be characterized by patronage, clientilism, and corruption to varying extents. However, the means by which an individual gains access to the benefits of patronage varies greatly between these two sorts of regimes. Thus, a central tenet of this paper is that the nature of regimes must be understood in terms of degrees of personalization or institutionalization of power, conceptualized as modalities of powerby this author.

    The concept of modalities of power can best be understood as a continuum, on which different sorts of regimes classified in prior literature lie. Chehabi and Linz’s (1998) “neosultanistic” regimes lie at the extreme of personalization, where all power flows from the individual leader who rules without constraint, directly appointing officials based on loyalty that is derived from an application of coercive threats and incentives. Extreme clientilism is practiced within a very small base of supporters (Chehabi & Linz, 1998). Neo-patrimonial regimes lie slightly closer to the center, displaying some aspects of institutionalization while still maintaining a high degree of personalization. Power still flows from the individual leader, and this leader governs the state as his personal property, using the state apparatus as a means to accumulate personal wealth and maintain a network of patronage, by which they reward loyal supporters with private goods and positions of power (Bratton & Van de Walle 1997). However, the leader does not rule completely free of constraints and circles of patronage are wider than those of “neosultanistic” regimes (Chehabi and Linz 1998).

    On the other end of the modalities of power continuum lie more institutionalized regimes, where, in the ideal case, power is dependent upon the highest office regardless of the individual occupying said office. These sorts of regimes include corporatist authoritarian regimes and more pluralistic authoritarian regimes, in all of which the state apparatus is constrained by a set of rules and an individual’s advancement through the bureaucracy and military follows a predictable trajectory. According to Huntington (1965), institutionalized regimes are characterized by adaptability, complexity, autonomy, and coherence. The corporatist authoritarian regime, including O’Donnell’s (1979) bureaucratic-authoritarian state, and its pluralistic counterpart, deemed simply “authoritarian” by Linz and Stepan (1996), possess channels of representation for multiple interest groups that create both complexity and autonomy by preventing the capture of the state by one societal group. These regimes thus adhere closely to Huntington’s standards of institutionalization. The mechanisms of mediation between the state and society are key to institutionalized regimes. In corporatist authoritarian regimes, interest groups are incorporated into the state and controlled channels are created for the representation of societal groups within the state apparatus (PC Schmitter 1979; Wiarda 2004). More pluralistic authoritarian regimes are also characterized by the incorporation of interest groups into the state apparatus; however, the state does not grant certain groups a monopoly on representation of a particular functional unit of society within the state, allowing for a higher degree of elite and interest group competition (Linz and Stepan 1996; Schmitter 1979). These institutionalized regimes can vary widely from the ideal form, with some falling near the center of the modality of power continuum as individuals show the potential to wield political power based upon personal charisma; however, in institutionalized regimes, power constraints and systems of checks and balances are in place to prevent the domination of the state by one individual.

    Conflict Recurrence and the Nature of the State: A Theoretical Explanation

    The new concept of modalities of power serves to illuminate the process by which post-conflict states become vulnerable to conflict recurrence. Several scholars have discussed the impact of weak governance on the onset and recurrence of civil conflicts, finding that civil conflicts erupt when rebel groups have motive, in the form of grievances, and opportunity to challenge the state (Collier and Hoeffler 2004; Fearon and Laitin 2003). However, few have gone beyond stating that “anocracies”—regimes that are neither strong democracies nor strong autocracies— are more vulnerable to civil conflict (Gurr 1974; Hegre et al. 2001). For a discussion of the process by which states become vulnerable to civil conflict, one must turn to the literature regarding revolutions. Tilly (1978) posits that civil conflict occurs under conditions of “multiple sovereignty” (200) situations where opposition groups have the motive and opportunity to mount an armed challenge to the state, significant portions of the population support these challenging groups, and the government is unable or unwilling to act to suppress these groups. Post-conflict states are especially vulnerable to the development of multiple sovereignty because conflicts typically weaken state infrastructure and create divisions and grievances within society. Among these post-conflict states, the nature of the regime, described through modalities of power by this author, may further determine these states’ level of vulnerability to the development of multiple sovereignty and the recurrence of civil war.

    Given the assumption that, as stated by Goodwin (2001), states may create and shape their own challengers by means of their governance, the modality of power within a regime has critical implications for a state’s vulnerability to conflict recurrence. According to Tilly’s theory of multiple sovereignty, there are three key processes in the creation of a scenario in which conflict recurs. First, an opposition group forms due to the presence of grievances and opportunity and lays an alternative claim on the government. Second, this group mobilizes, gaining increasing support from large segments of the population, often due to increasing societal grievances. Third, the government is unable or unwilling to suppress this group, leading to the outbreak of civil conflict. The modality of power within a regime directly influences the processes of opposition group formation and mobilization as well as the state’s capacity to respond to these groups.

    Due to their near-complete dependence on patron-client relationships and distribution of private goods in exchange for legitimacy, highly personalized regimes are generally characterized by high levels of corruption (Chehabi and Linz 1998). As a consequence, grievances may form among the disenfranchised population, who may view the leadership as profiting from their exclusion from the political system and may rally to form opposition groups. Of greater significance, in highly personalized regimes, appointments to office within the state bureaucracy and military occur based on personal loyalty to the leader, as opposed to competence, and assessments of loyalty also determine an individual’s access to private goods via patronage networks. Bratton and Van de Walle (1994) posit that during transitions from neo-patrimonial regimes, fragmentation of elites develops over access to patronage, as elites and members of the state apparatus who have been deemed “less-loyal” may be excluded from high positions or private goods. Chehabi and Linz (1998) describe similar processes of marginalization within neosultanistic regimes. Extending this argument to situations beyond regime reform, these excluded elites are less likely to support the current regime in the case of an armed challenge and much more likely to defect to the opposition. Furthermore, the regime lacks the channels of mediation between state and society to ascertain the level of popular support for policies enacted, to allow articulation of demands and interests of societal groups, or to address grievances that develop (Eisenstadt 1973). The lack of mechanisms for mediation forces the conveyance of grievances to occur outside the state apparatus and state control, thus encouraging the development of extra-state organizations to convey demands. These organizations may develop into full-fledged opposition groups that could potentially mount an armed challenge to the state.

    By contrast, in more institutionalized regimes, the state apparatus functions on rational-legal rules that determine political appointments primarily based on competence and merit. These regimes are, thus, less likely to suffer elite fractures because advancement and access to the benefits of office are possible regardless of the particular individual in power. Corporatist and pluralistic authoritarian regimes possess mechanisms of mediation between the state and society, providing the regime with information regarding popular support for public policy and creating opportunities for the articulation of societal grievances within the official state apparatus. The regime may then take some form of action to ameliorate grievances and preempt the formation or expansion of opposition groups based on grievances. These regimes also have the means to incorporate these groups into the state apparatus, a political strategy that Goodwin (2001) argues serves to de-radicalize mobilized groups (46).

    Personalized regimes may create increased opportunities for the development of armed challenges to the state. Due to fear of rivals, neo-patrimonial and neosultanistic leaders have a high incentive to weaken the state apparatus, particularly military institutions to prevent usurpation of power. By frequently removing military or other elites who could potentially threaten their grasp on power, neo-patrimonial leaders serve to protect their personal rule (Goodwin 2001). As a consequence they undermine the effectiveness of their regime as a whole to confront potential challengers and further compound problems of state weakness, which may embolden opposition groups to challenge the state.

    Leaders of institutionalized regimes also have an incentive to attempt to eliminate potential rivals; however, there are more constraints on the power of the individual leader that serve to prevent internal purges (Bratton and Walle 1994; Goodwin 2001). The strength of the actors in a more institutionalized regime is such that a leader who attempts to eliminate key members of the elite would be more likely to be removed from power than in a personalized regime. The power of the leader flows from the institutional office itself rather than the leader as an individual, creating a disincentive to weaken state institutions as that would weaken the power of the office. More institutionalized regimes also have the benefit of political appointments based upon competence, whereas personalized regimes make appointments based on loyalty, creating means for highly incompetent officials to reach key positions of power, further weakening the effectiveness of the state.

    The modality of power in a regime also affects the ability of the state to manage opposition groups that have already begun to mobilize and challenge the state. Personalized regimes, as mentioned before, are more likely to possess weakened institutions which may be staffed by incompetent officials, thus limiting their effectiveness in confronting a challenge to the state. However, beyond this, the personalization of power in neo-patrimonial regimes creates a zero-sum political game where the leader’s primary objective is political survival because a fall from power will likely mean the end of that leader’s free life. Opposition groups are unlikely to settle for anything less than the removal of a personalist leader from office and quite possibly his trial for crimes in office (Chehabi and Linz 1998). Personalized regimes are constrained in their ability to co-opt opposition groups because the private goods which leaders use to maintain political legitimacy are limited and cannot be dispersed widely without risking regime collapse (Bratton & Van de Walle 1994). Thus, personalized regimes often resort to coercive force to repress opposition to the regime; however, due to the often weak, corrupt, and incompetent nature of the military in these states, force may be applied indiscriminately and may not be enough to completely defeat the opposition. In these cases, as demonstrated by Mason and Krane (1989), indiscriminate force may serve to further mobilize opposition groups and increase their base of support within the governed population, which can create conditions of multiple sovereignty and facilitate a recurrence of civil war.

    By contrast, more institutionalized corporatist and pluralistic regimes are more likely to possess stronger, more effective institutions that are better equipped to address grievances. The institutional strength and diversified infrastructure of corporatist and pluralistic regimes also gives these regimes more flexibility to co-opt opposition groups. Because power flows from the highest office, not the individual, unpopular leaders can be replaced without regime failure. Furthermore, because interest groups are regularly incorporated into the state apparatus, a similar strategy of incorporation could be used to co-opt and de-radicalize opposition groups (Goodwin 2001). Should the regime choose to use force to suppress opposition, the stronger state institutions, including the military, are more likely to effectively apply overwhelming and discriminate force, which could prevent escalation of low-level conflict to civil war.

    This process also bears heavily on the willingness and ability of the government to suppress mobilized groups challenging the state. In highly personalized regimes, the military has been characterized as “hardly more than the praetorian guard of the despot,” which makes maintaining a monopoly of coercive force in the state territory difficult (O’Donnell, Schmitter, and Whitehead 1986). As mentioned before, neo-patrimonial and neosultanistic regimes are also prone to elite fractures that may make members of the government apparatus unwilling to defend the current regime or even defect to the opposition. Thus a situation in which the government is unable or unwilling to suppress an opposition group may be more likely to occur. In institutionalized regimes, the military is likely to be much more professionalized, more coherent, and stronger as an institution (Chehabi and Linz 1998). As a consequence, the government will likely be more willing and able to suppress opposition groups.


    In personalized regimes, societal grievances may develop and remain unaddressed, motivating opposition groups to form and mobilize outside the bounds of state control. The loyalty-based patronage system in these regimes may also cause elite fractures, weakening state institutions and creating the potential for defection of officials to opposition groups if the state is challenged. Should opposition groups attempt to challenge the state, they will likely be met with indiscriminate coercive force from the state due to state institutional weakness and the neo-patrimonial leader’s need to suppress potential rivals. Indiscriminate violence furthers the process of opposition group mobilization, and state institutional weakness and elite defection creates a scenario in which the government is unable to suppress these opposition groups, establishing conditions of multiple sovereignty. This yields the following hypothesis:

    Hypothesis 1: Higher degrees of personalization of power within a state will result in lower durability of peace.

    By contrast, in institutionalized regimes, mediation mechanisms between the state and society allow for the management of societal grievances and the co-optation of opposition groups. At the same time, competence-based appointments strengthen state institutions, and rules governing institutional operations constrain individual leaders and prevent the weakening of state institutions due to a leader’s attempt to suppress rivals. Should opposition groups challenge the state militarily, stronger state institutions are more likely to effectively apply coercive force to suppress these groups. From these arguments it is possible to draw the following hypothesis:

    Hypothesis 2: Higher degrees of institutionalization of power within a state will result in higher durability of peace.

    Research Design and Methodology

    These hypotheses are tested using a sample of 76 post-conflict countries during the period 1992-2011. The collapse of the Soviet Union in 1991 significantly altered the world’s political landscape, and, since the end of Cold War, the majority of conflicts have occurred at the intrastate level. As such, the focus of this study is the post-Cold War period. To be considered a post-conflict state, a country must have experienced a termination of all civil conflicts during the period of observation or within five years of the onset of said period. Thus, countries that experienced intrastate conflict ending in 1987 or later are included in this sample. Civil conflict data is drawn from the UCDP/PRIO Armed Conflict Dataset. To be coded as an intrastate conflict, cases in the UCDP/PRIO dataset must involve a confrontation over government and/or territory where armed force is deployed and at least 25 battle-related deaths result. The unit of analysis is country-years.

    Dependent Variable

    The dependent variable in this study is peace durability, measured by duration in years since the cessation of all civil conflict within a country. The onset of any new civil conflict is considered the end, or failure, of the peace period. The variable for the incidence of civil conflict was created using the UCDP/PRIO Intrastate Armed Conflict Dataset, with years in which conflict occurred coded as 1 and years in which no conflict occurred coded as 0.

    This study examines the relationship between the nature of the state and the general durability of peace in that state. As such, this study is not concerned with the history of any given conflict dyad and will thus follow the model of prior studies in considering the outbreak of any new civil conflict as the end of the peace period (Collier, Hoeffler, and Soderbom 2008; Doyle and Sambanis 2000; Hartzell and Hoddie 2003; Mason et al. 2011). Conditions of multiple sovereignty may allow any number of groups, not merely former rebel groups, to mobilize and challenge the state. As this study is concerned with the conditions arising from the nature of the state that allow for the development of multiple sovereignty and possibly renewed civil conflict, any form of relapse into conflict, regardless of the conflict dyad involved, is counted as a failure of the state to establish lasting peace.

    Independent Variables

    To test the hypotheses that the degree of personalization of power and the degree of institutionalization of power affect the likelihood of conflict recurrence, a measure of the nature of the state must be established. Since no satisfactory operationalized measure of the modality of power in the state exists in the current literature, this study incorporates an original five-point ordinal scale of measurement for the nature of the state in terms of modalities of power.

    Most prior literature on the nature of the state is limited to theoretical discussions of the neo-patrimonialism, neosultanism, or corporatism, or only examines a limited number of case studies related to one of these concepts (Adams 2004; Bratton and Van De Walle 1994, 1997; Chehabi and Linz 1998; Eisenstadt 1973; Linz and Stepan 1996; O’Donnell 1979; O’Donnell 1977; O’Donnell, Schmitter and Whitehead, 1986; Schmitter 1979; Wiarda 2004). Erdmann and Engel (2007) note that the nature of the state remains very difficult to operationalize, as there exist no empirical indicators for clientilism, patronage, or neo-patrimonial regimes as a whole. Thus, as an admittedly imprecise operationalization of the nature of the state, the author has chosen to create a five-point ordinal scale of measurement for the modality of power of a regime.

    Regimes in which power is deemed highly personalized due to extreme centralization of power in the hands of one individual with few constraints on executive authority are coded as 1. Regimes in which power is somewhat personalized are coded as 2. These regimes display centralization of power around an individual but executive authority is either somewhat limited or the state possesses strong institutions of mediation between the state and society. Regimes in which power is deemed mixed are coded as 3. The strong presence of both personalized and institutionalized aspects of power in the form of individual-dominated politics characterized by frequent turnovers of power preclude these regimes from being classified as more institutionalized or more personalized. Regimes in which power is somewhat institutionalized are coded as 4. These regimes demonstrate regular turnover of power and clear limits on executive authority, but Transparency International corruption scores remain high, showing a reliance on patronage within the state apparatus. Regimes in which power is highly institutionalized, evidenced by regular peaceful turnover of power and low corruption scores, are coded as 5.

    This coding system is based on annual political reports compiled by the Economist Intelligence Unit and contemporary political history summaries from Europa World Plus, as well as the Corruption Perceptions Index from Transparency International. The Modality of Power (MoP) score is recognized to be a very imperfect measure of the nature of the state, but in the absence of better measures and proxy variables that could indicate the modality of power exercised within the state, the author has deemed this scoring system to be a plausible measure for the modality of power of a regime.

    To test the impact of personalization of power on the durability of peace, a dichotomous variable for personalization of power was created, with 1 representing a personalized regime (coding 1 or 2 on the Modality of Power scale), and 0 representing all other more institutionalized regimes. A dichotomous variable was chosen in order to avoid use of an ordinal variable as a dependent variable and to test the assumption that any significant degree of personalization of power increases a state’s vulnerability to conflict recurrence. In accordance with hypothesis 1 and 2, this study expects to find a negative relationship between personalization and peace duration.

    Control Variables

    In previous studies, a number of factors have been found to affect peace durability, and these factors will need to be controlled for in order to isolate the relationship between the nature of the state and peace duration. The level of economic development has been established as a significant factor in peace durability and conflict onset (Collier and Hoeffler 2004; Elbadawi and Sambanis 2002; Fearon and Laitin 2003). Following Walter (2004) and Mason et al. (2011), the level of development is controlled for using countries’ infant mortality rate and gross domestic product per capita. These variables were created using data from the World Bank development indicators dataset, and the positive relationship between levels of development and peace duration observed in prior studies is expected to be supported. Relative regime type of democracy or autocracyhas also been shown to affect peace duration. Hegre et al. (2001) and Mason et al. (2011) found support for an inverted-u relationship between regime type and conflict onset, in which strong autocracies and strong democracies are least vulnerable to conflict while regimes falling in the middle range (deemed “semi-democracies,” “weak authoritarian regimes,” and “anocracies”) remain most vulnerable. Following Mason et al. (2011), in order to control for this effect, the POLITY IV democracy-autocracy score (Polity2) was used to determine regime type and the square of the this score was created as a control variable. This study expects to similarly observe the inverted-u relationship between Polity scores and peace duration.

    Fearon and Laitin (2003) observed a relationship between ethnic fractionalization and peace failure so their ethnic fractionalization index was included as a control variable. As Fearon and Laitin noted, ethnic fractionalization carries much less weight than factors such as economic development; thus, there is no expectation to see a significant relationship between ethnic fractionalization and peace duration. Walter (2004) and Doyle and Sambanis (2000) found that the type of incompatibility influences peace duration; specifically, peace-building may be more difficult following territorial conflicts. Three dichotomous control variables for government, territorial, and both government and territorial conflicts were created using the type of incompatibility variable from the UCDP Conflict Termination Dataset. The expectation is for territorial incompatibilities to have a significant impact on peace duration.

    Mason et al. (2011) and Doyle and Sambanis (2000) find that the cost of the previous conflict may impact peace duration. To measure previous conflict cost, previous conflict duration and levels of battle-related deaths are included as two continuous control variables. The previous conflict duration was drawn from the UCDP Conflict Termination Dataset measured in years. For years in which multiple conflicts ended, the duration of the longest conflict was retained as the measure of previous conflict duration for that year and the proceeding years until the end of the next conflict. Conflicts lasting less than one year were rounded up to one year, and for all other conflict durations, rounding up occurred at the six-month mark. The battle-related deaths variable was created using the number of battle-related deaths as recorded in the UCDP Conflict Encyclopedia. For years in which multiple conflicts ended, the highest number of battle-related deaths was maintained as the measure of battle-related deaths for that year and proceeding years until the end of that conflict and the start of a new conflict. The expectation is that a positive relationship between previous conflict cost and peace duration will be found. Mason et al. (2011) also find that the civil conflict outcome impacted peace duration. To control for this effect, two dichotomous variables for settlement and military victory were created using the outcome data from the UCDP Conflict Termination Dataset. Victory is expected to positively influence peace duration while settlement negatively impacts peace duration.

    Finally, as in many studies of post-Cold War civil conflict, many of the cases included in this dataset occurred on the African continent. To control for possible regional anomalies, a dichotomous variable was created, with 1 representing Africa and 0 representing all other regions. Region is not expected to have a significant effect on peace duration.

    Model Specifications

    A high degree of collinearity was found between the squared Polity2 score variable and the personalization measure, suggesting that both measures capture a highly similar portion of the variance in the dependent variable. For this reason, two separate models are used, one including the Polity2 variables and one including the personalization measure, with all controls included in each. Two restricted models are also run using only one variable from each category of control factors to avoid potential collinearity that could skew results. Thus, the duration variable is included as a measure of conflict cost, infant mortality rate for economic development, settlement for conflict outcome, the territorial variable for type of incompatibility, and the ethnic fractionalization variable, with Polity2 variables and the personalization measure again being run separately. Finally two restricted models are conducted again including the Africa variable in order to determine potential regional effects on the analysis results.

    To test the hypotheses regarding the effect of modality of power on peace durability, a Cox Proportional Hazard model was used to assess the potential duration of peace. The Cox model considers the duration of a time-variant dependent variable by estimating the hazard of failure relative to a baseline hazard rate. Unlike other duration models, the Cox model does not assume a particular form for the baseline hazard rate over time, offering a flexible method of analyzing the relationships between the covariates of interest and the dependent variable of peace duration (Box-Steffensmeier and Jones 2004). This model has been previously established as an effective tool for assessing the influence of particular factors on peace duration in previous literature (Fortna 2004; Hartzell and Hoddie 2003; Mason et al. 2011). The use of a partial likelihood method in the Cox model also allows adjustment for censored data without skewing results (Box-Steffensmeier and Jones, 2004).

    The Cox model additionally provides an effective method of handling tied failures, situations where failures occur conterminously. In this study, this refers to situation in which peace fails simultaneously due to multiple causes. To manage tied failures, this study employs the Efron method, a more computationally complex method of approximation than the simpler but widely used Breslow method. The Efron method allows for greater precision when working with a large number of tied events by accounting for a change in the risk set dependent upon the order in which tied events occur, unlike the Breslow, which assumes a static risk set despite variance in sequencing. In this study, the Cox model was employed to evaluate the duration of peace. Failure is thus the recurrence of conflict that ends the duration of peace, and the baseline hazard ratio and subsequent hazard ratios generated in each model refer to the hazard of peace failure.


    This study includes 112 separate peace spells out of 769 total peace-years. Of these, 81 peace spells end in renewed conflict. The results of the hazard analysis for all models are presented in Table 1. The Cox model generates hazard ratios rather than coefficients, allowing one to interpret these ratios as the percent change in the likelihood of failure, in this case conflict recurrence. A hazard ratio greater than 1 suggests a percent increase in the hazard of peace failure from the baseline hazard rate, while a hazard ratio less than 1 suggests a percent decrease in the hazard of peace failure relative to the baseline. Thus a hazard ratio of 1.2 implies a twenty percent increase in the hazard of peace failure, and a hazard ratio of 0.7 implies a thirty percent decrease in the hazard of peace failure. Factors that extend peace duration will thus possess hazard ratios less than one, while factors that shorten peace duration will hold hazard ratios greater than one. This analysis was run as a two-tailed significance test, with α levels set at 0.001, 0.01, and 0.05.

    Table 1 shows the effects of personalization of power and other variables on the hazard of peace failure across all models. In all models, personalization achieves significance at the p<0.01 level, with personalization reaching significance at the 0.001 level in both restricted models. Since personalized regimes were coded as 1 and all other regimes coded as 0, the 0.179 hazard ratio in the full model indicates that the presence of a personalized regime decreases the hazard of peace failure by approximately 82 percent. Similarly, in the restricted and restricted Africa models, hazard ratios of 0.127 and 0.114 suggest a decrease in the hazard of peace failure by 87 and 89 percent, respectively. Thus a significant positive relationship between personalization of power and peace duration can be inferred.

    Table 2 demonstrates the effects of relative regime type (Polity2) and other variables on the hazard of peace failure across all models. The raw Polity2 score, indicating degree of democracy or autocracy, fails to achieve significance at any level in any model; however, the square of Polity2, which measures relative strength of a regime (democratic or autocratic) with higher scores representing stronger regimes, reaches significance at the p<0.05 level in both the full model and the restricted regional model. The hazard ratio of 0.961 in the full model suggests a 4 percent decrease in the hazard of peace failure at higher levels of regime strength. Similarly, the hazard ratio of 0.973 in the restricted regional model indicates a 3 percent decrease in the hazard of peace failure at higher levels of regime strength. This result provides support for the findings of Mason et al. (2011) and Hegre et al. (2001) of an inverted-u relationship between relative regime type and peace duration.

    As shown in Table 1 and Table 2, previous conflict duration achieves significance in all models run with both personalization and regime type measures. Hazard ratios less than 1 indicate a percent decrease in the hazard of peace failure at higher levels of conflict duration, suggesting that, consistent with expectations, more costly conflicts reduce the risk of peace failure. Similarly, hazard ratios of economic development variables are in the expected direction, with higher levels of economic development associated with lower hazards of peace failure. However, GDP per capita and infant mortality rates prove to be highly collinear, precluding the analysis of both in the same model, and infant mortality rate achieves only a low level of significance in certain models. Settlement as a conflict outcome achieves significance in most models, and very small hazard ratios suggest a large percent increase in the hazard of peace failure in the presence of a settlement outcome, providing support for Mason et al.’s (2011) finding. Counter to expectations, territorial compatibility does not achieve significance in any model and provides conflicting hazard ratios depending on the model specification. As expected, ethnic fractionalization fails to achieve significance in all models, mirroring Fearon and Laitin’s (2003) findings. The regional control variable for Africa proves significant in the personalization model at the p<0.05 significance level, an unsurprising result given the high number of personalist regimes in this region.

    Table 3 provides results of an analysis carried out with both Polity2 measures and the personalization measure. In this model, the squared polity score achieves significance at the p<0.05 significance level, while the personalization variable fails to reach significance at any level. The implications of this result will be discussed below. All other variables provide results similar to those mentioned above, with the exception of ethnic fractionalization, which achieves significance at the p<0.05 level with a very high hazard level indicating a sizeable percent increase in hazard of peace failure at higher levels of fractionalization. However, as mentioned before, the personalization measure and the Polity2 variables share a high degree of collinearity which likely affected the results of this model.

    Discussion of Results

    These results show that, counter to both hypotheses, personalization of state power decreases the likelihood of peace failure. Personalized regimes are more likely to experience more durable peace than their more institutionalized counterparts. Additionally considering the interaction of the Polity2 variables with the personalization measure creates a fascinating puzzle. The simple answer might be that the personalization measure merely captures the autocratic end of the Polity2 score spectrum. However, an examination of the data proves that this is not the case, and the failure of the raw Polity2 score to achieve significance in any model while personalization achieves high significance suggests that this explanation does not stand. Rather, one must turn to the squared Polity2 score, which provides a measure of regime strength apart from levels of autocracy or democracy. The significance of this variable across all models again provides evidence for the argument that “anocracies,” regimes that are neither strong democracies nor strong autocracies, remain most vulnerable to civil conflict. Based on the model in Table 3, relative regime strength appears to more significantly impact the hazard of peace failure than personalization of power. The significance of the personalization measure when run independently from the Polity2 measures may hold a key to understanding potential variance in the vulnerabilities of “anocratic” or relatively weak states, a category into which many post-conflict states fall. In the realm of “anocracies,” personalization of power may greatly influence a regime’s vulnerability to renewed conflict.

    Personalization of power may decrease the hazard of peace failure within a post-conflict state for several reasons. First, the centralization of power around one individual may indicate state strength. The ability of a personalist leader to suppress rivals implies an ability to crush potential challenges to state authority. In a post-conflict situation where the state is already weakened, such a regime may be better equipped to manage potential challengers to the state than a state in which power is distributed among multiple individuals, none of whom can muster enough political clout to centralize power within his or her own hands. This situation may breed more inter-elite conflict that could potentially devolve into civil conflict, especially in a post-conflict situation where former rebel fighters may not be completely disarmed, and thus cripple the state’s ability to respond quickly to threats to its sovereignty. By contrast, the centralization of power within the personalist state may allow leaders to act quickly without concerns over accountability to crush potential revolts and budding conflicts.

    Second, the clientilistic nature of the personalist state and the dependence of the leader on a system of patronage to maintain legitimacy may prove beneficial for the preservation of peace because it creates a group of powerful elites who are vested in the survival of this particular state. Clients of personalist regimes hold a significant stake in the current regime’s survival because the benefits they reap from the current system of governance are greater than the risks of allowing the overthrow of said regime for the potential benefit of gaining more power. For this reason, said clients may actively cooperate with the state to suppress potential challengers and prevent the outbreak of conflict.

    Third, in the chaos of a post-conflict state, a leader who personalizes power creates stability, potentially proving more beneficial for peace than a more inclusive or institutional regime that cannot provide stability. As mentioned above, the centralization of power in the hands of one individual grants the state the ability to act quickly and decisively to address potential challengers to state authority. Along with this, the stability of leadership in a personalist state, despite its potential political exclusivity, may prove more conducive to post-conflict economic recovery and rebuilding than a more institutionalized regime where leadership changes frequently. Such stability may also create a greater sense of security within the state that, when combined with potential economic recovery, may reduce the incentives for citizens to support armed challenges to the state. In weak institutionalized regimes, however, the inability of the state to act decisively may create and exacerbate grievances, increasing incentives for support of an armed challenge. Thus, similar to the squared Polity2 score, the personalization measure may serve as another proxy for a measure of state strength. Unlike Polity2, this measure demonstrates the potential to differentiate among “anocratic” states, providing further insight into the factors impacting the risk of conflict recurrence in post-conflict states.

    Conclusion: The Implications of the Nature of the State

    Current literature in the field of peace and conflict research has focused attention on the influence of regime type, political inclusion, and state strength on the durability of peace and conflict recurrence, but largely ignored the potential impact of the nature of the state on these outcomes. This study has attempted, albeit imperfectly, to develop a clear typology of the modality of power exercised within the state apparatus in order to facilitate the development of an operationalized measure of the nature of state. These endeavors have allowed for a quantitative analysis of post-conflict states in the post-Cold War period and produced evidence of a clear relationship between the nature of the state and peace duration. The results of this analysis suggest that personalization of power has a positive and significant impact on the duration of peace in post-conflict states.

    The relationship between personalization of power and duration of peace observed in this study may provide part of the explanation for the persistence of semi-authoritarian and autocratic regimes twenty years beyond the Third Wave of democratization following the collapse of the Soviet Union. While remaining an impediment to potential democratic transitions, personalist leaders may serve to insulate the states they lead from falling into the “conflict trap” of recurring violence. The assumption that the rise of a “strongman” leader in a post-conflict state is a negative development may need to be reconsidered. Rather, the rise of a personalist leader or the increasing personalization of power in a state, such as in Russia under Putin, may prove beneficial in terms of stability and peace. The findings of this study suggest rather that multiple forms of governance may prove able to maintain peace following civil conflict, including nondemocratic forms of governance.

    At minimum, the results of this study provide highly compelling grounds for further scholarly inquiry into the nature of the state and its implications for peace and conflict. The development of a more precise operationalization of modalities of power could more accurately illuminate the impact of the nature of the state on matters of peace and conflict. Beyond the invention of better measures, these results raise many questions about the further implications of modalities of power on peace and conflict, such as the impact of personalization of power on the vulnerability of states to the initial onset of civil conflict. Are states with personalist regimes less prone to the onset of civil conflict? Further studies of post-conflict states over greater periods of time could further illustrate other processes that might be at work. What happens as personalist leaders age and die? Could institutionalization of power provide greater long-term stability as mechanisms for peaceful turnovers of power become entrenched? The author of this paper hopes that this study provides sufficient grounds to spark further investigations beyond the institutional structure of the state, and into the manner of power exercised within the state apparatus and the consequences of these modalities of power.


    • Adams, Paul S. 2004. “Corporatism in Latin America and Europe: Origins, Developments, and Challenges in Comparative Perspective.” In Authoritarianism and Corporatism in Latin America-Revisited, ed. Howard J Wiarda. Gainesville: University Press of Florida, pp. 58-90.
    • Albrecht, Holger, and Oliver Schlumberger. 2004. “‘Waiting for Godot’: Regime Change Without Democratization in the Middle East.” International Political Science Review/Revue Internationale de Science Politique 25(4): 371-392.
    • Box-Steffensmeier, Janet M., and Bradford S. Jones. 2004. “The Cox Proportional Hazards Model.” In Event History Modeling: A Guide for Social Scientists, Cambridge, U.K.; New York, NY, USA: Cambridge University Press, pp. 47-67.
    • Bratton, Michael, and Nicolas Van De Walle. 1994. “Neopatrimonial Regimes and Political Transitions in Africa.” World Politics 46(4): 453-489.
    • Bratton, Michael, and Nicolas Van de Walle. 1997. Democratic Experiments in Africa: Regime Transitions in Comparative Perspective. Cambridge, U.K.; New York, NY, USA: Cambridge University Press.
    • Chehabi, H, and Juan J Linz. 1998. Sultanistic Regimes. Baltimore: Johns Hopkins University Press.
    • Collier, Paul, and Anke Hoeffler. 2004. “Greed and Grievance in Civil War.” Oxford Economic Papers 56(4): 563-595.
    • Collier, Paul, Anke Hoeffler, and Mans Soderbom. 2008. “Post-Conflict Risks.” Journal of Peace Research 45(4): 461-478.
    • Collier, Paul, and Nicholas Sambanis. 2002. “Understanding Civil War: A New Agenda.” Journal of Conflict Resolution 46(1): 3-12.
    • Corruption Perceptions Index. 2011. Transparency International. (Accessed July 5, 2012).
    • Doyle, Michael W, and Nicholas Sambanis. 2000. “International Peacebuilding : A Theoretical and Quantitative Analysis.” The American Political Science Review 94(4): 779-801.
    • Economist Intelligence Unit. 2012. “Country Analysis.” (Accessed July 5, 2012).
    • Eisenstadt, S. 1973. Traditional Patrimonialism and Modern Neopatrimonialism. Beverly Hills [Calif.: Sage Publications].
    • Elbadawi, Ibrahim, and Nicholas Sambanis. 2002. “How Much War Will We See? Explaining the Prevalence of Civil War.” The Journal of Conflict Resolution 46(3): 307-334.
    • Erdmann, Gero, and Ulf Engel. 2007. “Neopatrimonialism Reconsidered: Critical Review and Elaboration of an Elusive Concept.” Commonwealth and Comparative Politics 45(1): 95-119.
    • Europa World Plus. 2012. (Accessed July 5, 2012).
    • Fearon, James D, and David D Laitin. 2003. “Ethnicity, Insurgency, and Civil War.” The American Political Science Review 97(1): 75-90.
    • Fortna, Virginia Page. 2004. “Does Peacekeeping Keep Peace? International Intervention and the Duration of Peace After Civil War.” International Studies Quarterly 48(2): 269-292.
    • ———. 2003. “Inside and Out: Peacekeeping and the Duration of Peace after Civil and Interstate Wars.” International Studies Review 5(4): 97-114.
    • Geddes, B. 1999. “What Do We Know about Democratization after Twenty Years?” Annual Review of Political Science 2(1): 115-144.
    • Gleditsch, Nils Petter, Peter Wallenstein, Mikael Eriksson, Margareta Sollenberg, and Håvard Strand. 2002. “Armed Conflict 1946-2001: A New Dataset.” Journal of Peace Research 39(5): 615-637.
    • Goodwin, J. 2001. No Other Way Out: States and Revolutionary Movements, 1945-1991.
    • Gurr, Ted Robert. 1974. “Persistence and Change in Political Systems, 1800-1971.” The American Political Science Review 68(4): 1482-1504.
    • Hadenius, Axel, and Jan. Teorell. 2007. “Pathways from Authoritarianism.” Journal of Democracy 18(1): 143-157.
    • Harbom, Lotta & Peter Wallenstein, 2010. ”Armed Conflict, 1946-2009.” Journal of Peace Research 47(4): 501-509.
    • Hartzell, Caroline A, and Matthew Hoddie. 2003. “Institutionalizing Peace: Power Sharing and Post-Civil War Conflict Management.” American Journal of Political Science 47(2): 318-332.
    • Hegre, Havard et al. 2001. “Toward a Democratic Civil Peace? Democracy, Political Change, and Civil War, 1816-1992.” The American Political Science Review 95(1): 33-48.
    • Huntington, Samuel P. 1965. “Political Development and Political Decay.” World Politics 17(3): 386-430.
    • Ishiyama, John. 2002. “Neopatrimonialism and the Prospects for Democratization in the Central Asian Republics.” In Power and Change in Central Asia, ed. Sally N Cummings. London; New York: Routledge, p. 42-58.
    • Kang, Seonjou, and James Meernik. 2005. “Civil War Destruction and the Prospects for Economic Growth.” The Journal of Politics 67(1): 88-109.
    • Kaufmann, Chaim. 1996. “Possible and Impossible Solutions to Ethnic Civil Wars.” International Security 20(4): 136-175.
    • Kreutz, Joakim. 2010. “How and When Armed Conflicts End: Introducing the UCDP Conflict Termination Dataset,” Journal of Peace Research 47(2): 243-250.
    • Lemarchand, Rene, and Keith Legg. 1972. “Political And Clientelism Development.” Comparative Politics 4(2): 149-178.
    • Levitsky, Stepan, and Lucas Way. 2002. “The Rise of Competitive Authoritarianism.” Journal of Democracy 13(2): 51-65.
    • Linz, Juan J, and Alfred Stepan. 1996. Problems of Democratic Transition and Consolidation: Southern Europe, South America, and Post-Communist Europe.Johns Hopkins University Press.
    • Luttwak, Edward N. 1999. “Give War a Chance.” Foreign Affairs 78(4): 36-44.
    • Mason, T. David, Mehmet Gurses, Patrick T Brandt, Jason Michael Quinn. 2011. “When Civil Wars Recur: Conditions for Durable Peace after Civil Wars.” International Studies Perspectives 12(2): 171-189.
    • Mason, T David, and Patrick J Fett. 1996. “How Civil Wars End: A Rational Choice Approach.” The Journal of Conflict Resolution 40(4): 546-568.
    • Mason, T David, and Dale A Krane. 1989. “The Political Economy of Death Squads : Toward a Theory of the Impact of Terror.” International Studies Quarterly 33(2): 175-198.
    • Mesquita, Bruce Bueno De, James D. Morrow, Randolph Siverson and Alastair Smith. 2000. “Political Institutions, Political Survival, and Policy Success.” In Governing for Prosperity, ed. Bruce Bueno de Mesquita and Hilton L. Root. New Haven: Yale University Press, 2000, pp. 59-84.
    • ———. 2003. The Logic of Political Survival. Cambridge, Mass.: MIT Press.
    • O’Donnell, G. 1979. “Tensions in the Bureaucratic-Authoritarian State and the Question of Democracy.” In The New Authoritarianism in Latin America, ed. David Collier. Princeton: Princeton University Press, pp. 285-318.
    • O’Donnell, Guillermo. 1977. “Corporatism and the Question of the State.” In Authoritarianism and Corporatism in Latin America, ed. James M. Malloy. Pittsburgh: University of Pittsburgh Press, pp. 47-89.
    • O’Donnell, Guillermo, P Schmitter, and L Whitehead. 1986. Transitions from Authoritarian Rule: Tentative Conclusions about Uncertain Democracies (Volume 4). Baltimore; London: Johns Hopkins University Press.
    • Paris, R. 2004. At War’s End: Building Peace After Civil Conflict. Cambridge, U.K.; New York, NY, USA: Cambridge University Press.
    • Polity IV Project, 2011. Political Regime Characteristics and Transitions, 1800-2010, Centre for Systemic Peace. (Accessed July 5, 2012)
    • Schmitter, PC. 1979. “Still the Century of Corporatism?” In Trends Toward Corporatist Intermediation, eds. Philippe C Schmitter and Gerhard Lehmbruch. Beverly Hills, London: Sage Publications, pp. 7-52.
    • Smith, Alistair, and Allan C. Stam. 2004. “Bargaining and the Nature of War.” Journal of Conflict Resolution 48(6): 783-813.
    • Snyder, Richard. 1992. “Explaining Transitions from Neopatrimonial Dictatorships.” Comparative Politics 24(4): 379-399.
    • Stedman, Stephen John. 1997. “Spoiler problems in Peace Processes.” International Security 22(2): 5-53.
    • Theobald, Robin. 1982. “Patrimonialism.” World Politics 34(4): 548-559.
    • ———. 1999. “So what really is the problem about corruption?” Third World Quarterly 20(3): 491-502.
    • Themnér, Lotta & Peter Wallenstein, 2012. “Armed Conflict, 1946-2011.” Journal of Peace Research 49(4).
    • Tilly, Charles. 1978. From Mobilization to Revolution. Reading, Mass.: Addison-Wesley Pub. Co.
    • Uppsala Conflict Data Program (Date of retrieval: 12/07/05) UCDP Conflict Encyclopedia:, Uppsala University.
    • Walter, Barbara F. 2006. “Building Reputation: Why Governments Fight Some Separatists but Not Others.” American Journal of Political Science 50(2): 313-330.
    • ———. 2004. “Does Conflict Beget Conflict? Explaining Recurring Civil War.” Journal of Peace Research 41(3): 371-388.
    • Wiarda, Howard J. 2004. “Introduction: Whatever Happened to Corporatism and Authoritarianism in Latin America?” In Authoritarianism and corporatism in Latin America-revisited, ed. Howard J Wiarda. Gainesville: University Press of Florida, p. 1-28.
    • World Bank. 2012. World Development Indicators. (Accessed July 5, 2012).

    Appendix: Coding Rules Regarding Modalities of Power

    Regimes in which power is deemed highly personalized are coded as 1. These regimes are characterized by clear centralization of power around one individual or family, as evidenced by the lack of turnovers of power except within a family and the appointment of family members to high office. Little to no limit on chief executive’s power within the state is present in the state constitution or applied in practice. The state apparatus is dominated by patronage and loyalty based appointments, generally indicated by high levels of corruption, and there is often evidence of purges conducted by the chief executive within the bureaucratic and military apparatus. If elections are held, returns in favor of the executive are usually above ninety percent and conducted in a manner that is not deemed free or fair by outside observers. Regimes of this type include Azerbaijan, where the Aliyev family has retained a firm grip on power since the country gained independence in 1992, and the Democratic Republic of Congo (Zaire) under the thirty-two year rule of Mobutu Sese Seko.

    Regimes in which power is somewhat personalized are coded as 2. These regimes are characterized by some centralization around one or several individuals, again evidenced by appointment of family members or longtime supporters to high office and a lack of turnovers of power. High levels of corruption again provide some evidence of loyalty-based appointment. However, in these regimes, evidence of bureaucratic process or mechanisms of state-society mediation is present and some limits on chief executive’s power exist, indicated by successful pressuring of the chief executive by the military or other political bodies or figures. Regimes of this type include the Mubarak regime in Egypt, where Hosni Mubarak retained power for thirty years, but the state apparatus possessed clear, well-developed corporatist institutions of mediation between the state and society. Morocco under Hassan II and Mohammed VI, where the monarch clearly holds the highest level of power and can overrule all other political bodies, but generally submits to the decisions of the elected parliament, also falls into this category.

    Regimes in which power is deemed mixed are coded as 3. These regimes are characterized by centralization of power around one or several individuals and clear patronage-based appointment while simultaneously displaying signs of institutionalization through frequent turnovers of power and clear, exercised limits on executive authority. The two most exemplary cases of this categorization are Bangladesh and Papua New Guinea, both possessing very weak and extremely corrupt democratic regimes. In these regimes, power is centralized around two or three individuals who maintain a firm grasp on their political parties through total patronage and loyalty-based appointments. However, these individuals regularly cycle through office at each election, regularly turning over power peacefully. The strong presence of both personalized and institutionalized aspects of power precluded these regimes from being classified as more institutionalized or more personalized.

    Regimes in which power is somewhat institutionalized are coded as 4. These regimes are characterized by regular (two or more in the period of observation) turnovers of power and clear limits on executive authority in the state constitution and practice, demonstrated by successful challenge of executive authority by political bodies such as a legislature, judiciary, or the military without overthrow of said executive. Institutions are at least somewhat developed, demonstrated by their possession of some measures to combat corruption and checks and balances to prevent capture of the state by a single group. However, these regimes are generally characterized by high levels of corruption, revealing a dependency on loyalty-based appointments within the political structure. Countries possessing regimes of this type include the Philippines, where the regime continues to sustain high levels of corruption, but there is frequent turnover of power, presidents obey constitution, and parliament possesses a demonstrated ability to remove a president from power.

    Regimes in which power is highly institutionalized are coded as 5. These regimes are characterized by regular (more than two in the period of observation) turnovers of power and possess clearly developed institutions, demonstrated by stability of officials despite changes of executive leadership. Clear limits on executive authority are evidenced within the constitution and in practice through successful challenges of executive authority by other political bodies such as the legislature and judiciary. Low corruption scores indicate merit-based appointment within the state apparatus. The Transparency International Corruption Perception Index (CPI) provided one of the primary delineating factors between highly institutionalized and somewhat institutionalized regimes as a proxy measure of institutional strength and merit-based appointments. In order to be classified a highly institutionalized regime, the country’s CPI score must be sustained above a 4. Countries possessing regimes of this type include South Africa, in which power has turned over between four presidents since the beginning of the observation period, and Israel, in which the office of Prime Minister has been turned over seven times since 1992. In countries that possess a hereditary monarchy, the monarch must retain only symbolic power (thus precluding familial inheritance of high political power), such as in Spain or the United Kingdom, to be considered highly institutionalized.

    This coding is based on annual political reports compiled by the Economist Intelligence Unit and contemporary political history summaries from Europa World Plus, as well as the Corruption Perceptions Index from Transparency International. The Modality of Power score is recognized to be a very imperfect measure of the nature of the state, but in the absence of better measures and proxy variables that could indicate the nature of the state, the author has deemed this scoring system to be a plausible measure for the modality of power of a regime.

    Table 1: Cox Proportional Hazard Analysis of Peace Duration following Civil Conflict Personalization of Power Models

      Full Model   Restricted Model   Restricted with Africa  
    Variable: Hazard Ratio
    (Std. Error)
    P>|z| Hazard Ratio
    (Std. Error)
    P>|z| Hazard Ratio
    (Std. Error)
    Polity2 Score  1.125



    Square of Polity2 0.961



    Previous Conflict Duration 0.813



    Total Battle-Related Deaths 1.000

    Settlement 0.060



    Victory 0.367

    Territorial Incompatibility 0.944



    Ethnic Fractionalization 25.532



    Infant Mortality Rate 0.961



    GDP Per Capita 1.000

    Africa         0.420


    *p<0.05, **p<0.01, ***p<0.001 (two-tailed tests).

    Table 2: Cox Proportional Hazard Analysis of Peace Duration following Civil Conflict Polity2 Models

      Full Model   Restricted Model   Restricted with Africa  
    Variable: Hazard Ratio
    (Std. Error)
    P>|z| Hazard Ratio
    (Std. Error)
    P>|z| Hazard Ratio
    (Std. Error)
    Personalization 0.179



    Previous Conflict Duration 0.861



    Total Battle-Related Deaths 1.000

    Settlement 0.114



    Victory 0.877

    Territorial Incompatibility 1.036



    Ethnic Fractionalization 6.325



    Infant Mortality Rate 0.998



    GDP Per Capita 1.000

    Africa          0.271


    *p<0.05, **p<0.01, ***p<0.001 (two-tailed tests).

    Table 3: Cox Proportional Hazard Analysis of Peace Duration following Civil Conflict Personalization and Polity2 Model

    Variable: Hazard Ratio
    (Std. Error)
     Personalization 0.182

     Polity2 Score  1.036

     Square of Polity2  0.971

     Previous Conflict Duration  0.907

     Settlement  0.072

     Territorial Incompatibility  1.572

     Ethnic Fractionalization  57.511

     Infant Mortality Rate  0.979

     Africa  0.298


    *p<0.05, **p<0.01, ***p<0.001 (two-tailed tests).