1.1 Research Question

Observers of North Korea have been fascinated by the sudden emergence of Kim Jong Un’s daughter, Kim Ju Ae. First seen publicly in late 2022 (O’Carroll and Reddy Reference O’Carroll and Reddy2022), her appearances alongside her father, especially at military-related events, have raised questions among long-time observers about whether Kim Jong Un is preparing for her to be his successor. If something happens to Kim Jong Un, who is only approaching middle age but is of questionable health (Cha and Katz Reference Cha and Katz2023), Kim Ju Ae’s young age (she is believed to have been born in 2013), and possibly her gender, could precipitate challenges if there were efforts to unify North Korea’s top officials behind her leadership.Footnote ¹

Yet, while Kim Jong Un is the undisputed leader and most powerful individual in North Korea (McEachern Reference McEachern2019), it was also not certain that Kim would reach this point. When he succeeded his father in December 2011, long-time observers of North Korea similarly doubted whether he could survive in office, let alone consolidate power. For example, in a December 201New York Time op-ed entitled “China’s Newest Province?,” former Director for Asian Affairs at the National Security Council, Victor Cha (Reference Cha2011), wrote that whether North Korea “comes apart in the next few weeks or over several months, the regime will not be able to hold together after the untimely death of its leader.”Footnote ² As is true for all autocracies, the main threat to Kim at the time of his ascension was perceived to come from within the regime (Joo Reference Joo2012). Elite splits are the biggest danger to any autocrat, either through leading directly to a coup or by motivating the people or foreign actors to challenge the dictator (O’Donnell and Schmitter Reference O’Donnell and Schmitter1986: 19; Svolik Reference Svolik2012: 4). This potential threat to Kim Jong Un was visible at his father’s funeral, when the youthful Kim was accompanied by seven powerful men as he marched alongside his father’s hearse. These individuals included his uncle and the rumored power behind the throne, Jang Song Thaek; the chief of general staff of the military, Ri Yong Ho; and the defense minister, Kim Yong Chun.Footnote ³ All had been key figures under Kim Jong Il, and their experience, connections, and knowledge of the regime’s inner workings predated Kim Jong Un’s.

However, despite these men’s more extensive experience of autocratic elite politics inside North Korea, the succession passed off peacefully. Accounts of the Kim Jong Il to Kim Jong Un succession point to the importance of strategic management of elites in allowing this to happen. For example, Kim Jong Il placed his brother-in-law, Jang Song Thaek, in key positions to help Kim Jong Un govern in his early years (Gause Reference Gause2015: 46–54). Once in power, Kim Jong Un initially relied heavily on Jang to govern, but he discarded him in December 2013 by having him expelled from a Korean Workers’ Party (KWP) politburo meeting and subsequently executed (Mansourov Reference Mansourov2013). The case of how Jang was managed before and after the succession by Kim Jong Il and Kim Jong Un, respectively, highlights the importance of elite management for a succession to occur as planned.

Thus, although there is no guarantee that a transition to Kim Ju Ae would occur peacefully, North Korea has shown that even in the adverse circumstances posed by a young and inexperienced successor, the regime can pull off a peaceful leadership transition. North Korea’s accomplishment is not unique. In post-Cold War Asia for instance, there have been successful autocratic leadership transitions in Cambodia (one), China (three), Indonesia (one), Kazakhstan (one), Lao (three), Malaysia (three), Myanmar (one), Singapore (two), Turkmenistan (two), Uzbekistan (one), and Vietnam (five).

The frequency of succession in autocracies around the world contrasts with academic research on the subject in two ways. First, succession has traditionally been viewed as a destabilizing time for autocracies (Herz Reference Herz1952: 20–1; Huntington Reference Huntington1965: 396; Olson Reference Olson1993: 572). Despite this, many autocracies in Asia and elsewhere have successfully navigated this supposedly challenging time. Second, there is a limited body of empirical research on the subject (Brownlee Reference Brownlee2007; Frantz and Stein Reference Frantz and Stein2017; Kokkonen and Sundell Reference Kokkonen and Sundell2014; Kokkonen et al. Reference Kokkonen, Møller and Sundell2022; Kurrild-Klitgaard Reference Kurrild-Klitgaard2000; Meng Reference Meng2020, Reference Meng2021). We are therefore in the unsatisfactory position where succession is critical to the endurance of authoritarian regimes, but we have limited knowledge about how autocratic regimes carry out leadership transitions.

Empirical research on the subject provides some important evidence about how autocracies carry out successions, but our understanding is incomplete. Prior work mainly focuses on how autocrats use institutions to dissuade elites from challenging the incumbent when he is still in office (Brownlee Reference Brownlee2007; Frantz and Stein Reference Frantz and Stein2017; Kokkonen and Sundell Reference Kokkonen and Sundell2014; Kokkonen et al. Reference Kokkonen, Møller and Sundell2022; Kurrild-Klitgaard Reference Kurrild-Klitgaard2000; Meng Reference Meng2020: chap. 7, Reference Meng2021). However, as the North Korea case shows, dictators also employ noninstitutional strategies to navigate succession. Further, as Kim Jong Il’s and Kim Jong Un’s management of Jang Song Thaek illustrates, elite management beforan after the transition is important in shaping whether a succession occurs as planned. We therefore ask: How do dictators manage elites to facilitate succession?

1.2 Scope Conditions and Argument

In this Element, succession refers to the regular transfer ode fact leadership of an autocratic regime from the incumbent to another individual of the incumbent’s choosing.Footnote ⁴ This captures a subset of successions in autocracies – regimes that lack meaningful contestation and participation (Dahl Reference Dahl1971) – but it is a significant subset. Dictators often attempt to hand power to their preferred individual, sometimes successfully as Kim Jong Il did to Kim Jong Un in 2011, and sometimes unsuccessfully such as when Egypt’s Hosni Mubarak’s attempt to hand power to his son, Gamal Mubarak, was interrupted by the Arab Spring. Other instances of this kind of succession include the transitions from Mao Zedong to Hua Guofeng in China in 1976, Nursultan Nazarbayev to Kassym-Jomart Tokayev in Kazakhstan in 2019, and Chiang Kai-shek to Chiang Ching-kuo in Taiwan in 1972. As these cases illustrate, this kind of succession can occur in any type of autocracy, and it can involve hereditary and nonhereditary transitions. There are other types of succession where leaders do not attempt to hand off power to a preferred successor, either due to a rare lack of interest (e.g., Castelo Branco to Costa e Silva in Brazil in 1967; Skidmore Reference Skidmore1989: 40) or because they lack the power necessary to install their preferred choice (e.g., Hu Jintao to Xi Jinping in China in 2012). However, as noted, leadership transitions from one dictator to an individual of their choosing are a relatively frequent type of succession in autocracies around the world.

There are two periods when dictators pay close attention to how they manage elites to facilitate succession. Prior research on succession highlights the importance of what happens in the period leading to an autocratic leadership transition (Frantz and Stein Reference Frantz and Stein2017; Kokkonen and Sundell Reference Kokkonen and Sundell2014; Kokkonen et al. Reference Kokkonen, Møller and Sundell2022; Meng Reference Meng2020: chap. 7, Reference Meng2021). If incumbents fail to prepare or initiate a strategically flawed succession process, they risk elites attempting to remove them from office prior to their death or intended departure date. Thus, this prior work highlights that succession planning helps protect an incumbent from a coup attempt before he leaves office. But a corollary of these actions is that by deterring or resisting attempts to remove them from office, an incumbent’s actions before a succession also play an important role in allowing their successor to enter office. Then, research not explicitly on succession but on authoritarian survival highlights the importance of elite management for dictators once they enter office (Goldring and Matthews Reference Goldring and Matthews2023, Reference Goldring and Matthews2024; Sudduth Reference Sudduth2017; Svolik Reference Svolik2012: chap. 3). To better understand how autocratic leadership transitions can occur, we therefore focus on (1) how outgoing autocrats and their successors manage elites prior to the transition and (2) how successors manage elites once they enter office.

Prior to a transition, we argue that incumbents prepare for succession by building a power base for the successor among elites who are outside of the incumbent’s inner circle, or by creating the conditions that permit their successor to do this. No dictator rules alone; successors require a power base to stabilize the regime so that they can keep providing the functions of autocratic governance without disruption. Relying on incumbent-era elites for this power base, however, would be risky for the successor. These elites are comparatively strong, having spent significant time proximate to the regime’s center of power. On the other hand, empowering elites who were outside the incumbent’s inner circle negates this ‘sovereign’s dilemma’ (Wang Reference Wang2022) – when a coherent group of elites facilitates a strong state but threatens the leader’s tenure – because the career trajectories of these elites become tied to the successor’s rise. Elites who were outside the incumbent’s inner circle but whose fortunes rise prior to the succession are likely to support the successor.

Then, after the transition, successors manage elites with the aim of consolidating the transition once they enter office. Looking to solidify their hold on power, successors seek to reduce the size of their ruling coalition. However, successors initially retain the elites who were raised in prominence to support the successor; these elites are valuable in helping them govern in the nascent stage of their rule. Successors do not retain these elites indefinitely though. Over time, they also marginalize these individuals to consolidate power and establish themselves as the regime’s undisputed leader. Because our argument relies on an autocrat being able to manipulate the status of elites, it primarily applies to autocracies with personalistic incumbents.

1.3 The Evidence

We test the arguments using the two leadership transitions in North Korea from Kim Il Sung to Kim Jong Il in 1994 and from Kim Jong Il to Kim Jong Un in 2011. North Korea is rarely used to test theories about comparative authoritarianism with broader applicability – in this case, autocracies with personalistic incumbents – so why do we select these transitions?

First, North Korea features two leadership transitions from personalistic incumbents. North Korea therefore represents a typical case that we examine to probe whether the posited causal mechanisms function as expected. Failing to find evidence of the arguments would disconfirm our theory (Seawright and Gerring Reference Seawright and Gerring2008: 299). Second, the transitions in North Korea represent prominent but puzzling cases of succession – prominent because of the value of understanding internal North Korean politics due to the regime’s geopolitical importance and puzzling because many experts did not expect the regime to pull off either succession (Byman and Lind Reference Byman and Lind2010). Third, the two transitions feature clearly identifiable periods before and after the successions where the outgoing and new leaders were attempting to prepare for and secure the succession. We therefore expect to observe evidence of our arguments in these periods. Relatedly, in the Kim Jong Il to Kim Jong Un transition, Kim Jong Il’s stroke in August 2008 shocked him into preparing for succession (Gause Reference Gause2011). This allows us to make causal inferences about how the second Kim’s decision to prepare for succession affected his management of elites. In sum, examining the leadership transitions in North Korea entails explaining two substantively consequential transitions in East Asia and provides inferences about how succession occurs in autocracies with personalistic incumbents across the world.

North Korea has been largely excluded from studies on comparative authoritarianism not just because it is often viewed as atypical but because its opacity makes it hard to study. However, we utilize innovative qualitative and quantitative data to glean insights into elite management surrounding the transitions. For the Kim Il Sung to Kim Jong Il transition, we draw on a range of primarily qualitative literature, including elite North Korean defector memoirs, official North Korean government sources, third-country archival sources, and Korean- and English-language secondary literature. Then, for the Kim Jong Il to Kim Jong Un transition, we primarily rely on an original dataset of 1,953 leadership events in North Korea between 1994 and 2013, combined with novel biographical data on the 653 elites who attended these events. Dictators like the Kims use invitations to events to influence the power of elites; analyzing attendance patterns therefore provides insights into how dictators manipulate intra-elite power relations.

1.4 Contributions

The Element contributes to two bodies of literature. First, it contributes to research on comparative authoritarianism in several ways. Most broadly, we focus on an understudied but consequential topic – succession – and develop a novel set of theoretical arguments about how this process occurs. Research on authoritarian survival features many studies on how dictators avoid falling to irregular types of exits, especially at the hands of insider elites (e.g., De Bruin Reference De Bruin2020) but also from the people (e.g., Chin et al. Reference Chin, Song and Wright2023) and foreign states (e.g., Escribà-Folch Reference Escribà-Folch2012). Yet succession is also an extremely common form of leadership transition in autocracies.Footnote ⁵ Several scholars recently investigated the determinants of peaceful succession (Kokkonen and Sundell Reference Kokkonen and Sundell2014; Kokkonen et al. Reference Kokkonen, Møller and Sundell2022; Meng Reference Meng2020, Reference Meng2021); however, to quote Meng (Reference Meng2020: 206), “[t]hough the politics of succession is considered to be one of the central challenges of autocratic rule, the mechanisms that facilitate peaceful leadership transitions are not well understood for modern dictatorships.” This is partly because prior work focuses on how autocrats use institutions to facilitate succession. We therefore build on this work by highlighting a noninstitutional strategy that outgoing and incoming dictators rely on to stabilize their regime: elite management. Overall, by providing arguments and evidence of how dictators manage elites to facilitate succession, we contribute to knowledge about how autocratic regimes endure.

Then, we develop a new set of theoretical arguments about how autocratic leadership transitions occur by challenging a key assumption in comparative research about succession. This assumption contends that dictators do not care about the regime after their death (Meng Reference Meng2020: 207; Shih Reference Shih2022: 23).Footnote ⁶ However, Kim Il Sung’s and Kim Jong Il’s careful management of elites was not aimed at protecting their tenures from internal challenges but to aid their successors and support their legacies beyond their lifetimes. Case vignettes of elite management surrounding succession elsewhere suggest that this is a broader trend, which gives pause to consider how dictators’ actions are motivated by goals beyond their political survival.

Second, the Element contributes to North Korean studies. It first builds on but also challenges prior accounts of the two transitions. Many accounts of the Kim Il Sung to Kim Jong Il transition suggest that Kim Jong Il made few changes to the ruling elite during his first three years due to the precariousness of his position, with purges not starting until the late 1990s (Jeon Reference Jeon2000: 765–7). The Element challenges this narrative; drawing on new and previously untapped sources about the 1990s, we show that consequential changes to the composition of the ruling elite took place soon after and even before the transition – via the tactics of elite emasculation and purges – and this provided the roots of elite support that Kim Jong Il enjoyed throughout his rule. Regarding the Kim Jong Il to Kim Jong Un transition, observers agree that the party was important prior to the transition (Gause Reference Gause2011), but we know little about how Kim Jong Il elevated party officials or how this aided a peaceful transition. As Gause (Reference Gause2011: 118) writes, “[h]ow Kim Chong-ilsi] secured the initial support for his chosen successor is not well known.” We explaiho Kim Jong Il used party elites to help ensure Kim Jong Un’s position after the former’s death. Then, post-succession, our account diverges from the main accounts of the transition, primarily by journalists or those in the policy community, which emphasize how attention-grabbing events, especially the 2013 execution of Jang Song Thaek, helped Kim Jong Un consolidate power (Fifield Reference Fifield2019). Instead, we contend that Kim consolidated power before purging Jang, through sidelining specific elites at certain times to implement a strategically sequenced process of selective marginalization.

The Element further contributes to North Korean studies by integrating the country more deeply within the study of comparative authoritarianism. North Korea has traditionally been viewed as an idiosyncratic autocracy and has largely been excluded from research on comparative authoritarianism.Footnote ⁷ However, scholars have recently demonstrated the applicability of comparative theory to explain North Korean politics (Koo et al. Reference Koo, Choi and Kim2016; Mahdavi and Ishiyama Reference Mahdavi and Ishiyama2020; McEachern Reference McEachern2018; Song and Wright Reference Song and Wright2018). We take this further, testing a set of comparative arguments using data from North Korea for primarily methodological reasons, and then demonstrating through comparative applications of the argument in other autocracies with personalistic incumbents that studying North Korean politics can help us learn about how autocracy functions in broader contexts.

1.5 Roadmap

The Element proceeds in five remaining sections. Section 2 provides the Element’s theoretical arguments. We set out two main arguments about how dictators manage elites to facilitate succession. These focus on (1) how autocrats manage elites to prepare for the transition to the incumbent’s preferred successor and (2) how successors manage elites to consolidate power and solidify the transition. Since the arguments apply beyond North Korea, we set out these arguments in general comparative terms, with occasional illustrative examples from North Korea and broader personalistic authoritarian contexts.

Sections 3 and 4 test these arguments using the transitions in North Korea from Kim Il Sung to Kim Jong Il and Kim Jong Il to Kim Jong Un, respectively. Our use of empirical evidence in these sections is the inverse of one another. In Section 3, we adopt a primarily qualitative approach, drawing from a range of primary and secondary English- and Korean-language sources, including original North Korean sources, to identify how Kim Il Sung facilitated the construction of a power base for Kim Jong Il, and then how Kim Jong Il managed elites once he came to power. We supplement this qualitative analysis with some descriptive data on which elites attended public leadership events, subject to data availability. Then, in Section 4, we employ a primarily quantitative approach using an original dataset of 1,953 leadership events between 1994 and 2013, combined with novel biographical data on the 653 elites who attended these events. To gain insights into what is driving our quantitative findings, we supplement this quantitative analysis with qualitative insights of how Kim Jong Il and Kim Jong Un managed elites before and after the transition.

Section 5 moves beyond North Korea, providing comparative applications of the Element’s arguments. As mentioned, our arguments primarily apply to autocracies with personalistic incumbents. Beyond this scope condition, however, we provide illustrative qualitative evidence that the arguments apply to autocracies irrespective of ideology, institutional makeup, and whether the succession is hereditary. We provide case vignettes of autocratic elite management surrounding succession in China’s party-based regime, Egypt’s military regime, and monarchical Saudi Arabia.

Section 6 concludes by considering implications of the findings for researchers and practitioners, what our findings suggest about how future succession in North Korea may occur, and questions for future research.

2.1 Introduction

This section contains comparative theoretical arguments about how dictators manage elites to facilitate leadership succession. These arguments are causal in that we explain (a) how preparing for succession affects a personalistic dictator’s management of elites and (b) how a successor attempting to solidify a transition affects elite management. We first posit that incumbents prepare for succession by building, or enabling the successor to build, a power base of elites from outside their inner circle to help the successor stabilize the regime once they come to power. Then, we describe how successors manage elites to solidify the transition, arguing that successors reduce the size of their ruling coalition to consolidate power. The exceptions to this are the elites who the preceding autocrat raised up in prominence to support the successor; successors initially retain these individuals because they can help them govern in the nascent stage of their rule.

2.2 Elite Management before Succession

A growing body of empirical research shows that incumbents use primarily institutional solutions to stabilize their regime against the future shadow of succession (Brownlee Reference Brownlee2007; Frantz and Stein Reference Frantz and Stein2017; Kokkonen and Sundell Reference Kokkonen and Sundell2014; Kokkonen et al. Reference Kokkonen, Møller and Sundell2022; Kurrild-Klitgaard Reference Kurrild-Klitgaard2000; Meng Reference Meng2020: chap. 7, Reference Meng2021). These studies assess evidence from Europe and sub-Saharan Africa as well as a global sample of post-World War II non-monarchical regimes to explore how incumbents solve the thorny issue of succession. For example, examining forty-two European states between 1000 and 1800, Kokkonen and Sundell (Reference Kokkonen and Sundell2014) show that succession based on agnatic primogeniture – the oldest son inheriting power – reduces the likelihood of monarchs being deposed. This argument encapsulates the themes in these studies about how institutionalized rules can facilitate succession: They bring clarity and predictability to an otherwise unstable situation, thereby dissuading elites from challenging the incumbent.

However, incumbents care about succession for reasons beyond stabilizing their rule (Wong and Chan Reference Wong and Chan2021: 595; cf. Meng Reference Meng2020: 207, Reference Meng2021: 958; Shih Reference Shih2022: 23). If the incumbent has not died, a handpicked successor can protect the incumbent from arrest or execution after leaving office (Goemans Reference Goemans2008: 772). Incumbents also often care about protecting their family and allies after their departure, as well as preserving their legacy (Lachapelle et al. Reference Lachapelle, Levitsky, Way and Casey2020: 35). For example, before succeeding Mahathir Mohamad in 2003, knowing it was important to Mahathir, Abdullah Badawi told him that “‘My vision for Malaysia is Vision Reference Lachapelle, Levitsky, Way and Casey2020,’ which was Dr. Mahathir’s blueprint” (Wain Reference Wain2009: 307). For these reasons, incumbents do not just desire regime stability before the succession; they also want stability after they are gone and the successor takes over.

Incumbents therefore often seek to empower the successor for when he takes power, so that he can achieve the goals described in the previous paragraph. If the successor is not empowered, instability may ensue. Meng (Reference Meng2020: 208) observes that power vacuums after a leader’s death can invite coups or civil war as elites fight to become the incumbent.Footnote ⁸ In a less extreme scenario, a successor may hold the position ode jur leader, but if they lacde fact power then challenges from elites are more likely because elites will see the successor as weak. A successor being weak can also precipitate agitation from actors outside the regime – that is, the people or a foreign state – who seek to capitalize on regime weakness (Bueno de Mesquita and Smith Reference Bueno de Mesquita and Smith2017; Kokkonen et al. Reference Kokkonen, Møller and Sundell2022: 19). In short, successors experience instability when they are insufficiently empowered to govern effectively. Incumbents therefore seek to empower the successor so that he can stabilize the regime by continuing to provide the functions of autocratic governance. This subsequently helps the successor deter, or at least overcome, challenges within and outside the regime.

It is a well-known aphorism that no dictator governs alone (Svolik Reference Svolik2012: 79). Dictators require a power base of trusted elites who can help them nullify threats within the regime, control the people, and navigate threats from foreign states; these are the primary sources of threat to an autocrat’s survival (Svolik Reference Svolik2012: 4–5). In North Korea, the regime’s founding leader Kim Il Sung relied especially on elites who had fought alongside him as guerrillas in Manchuria during the Japanese occupation of the Korean Peninsula (Lankov Reference Lankov2013: chap. 1). A power base of trusted elites is essential for all dictators, including successors, to help them wield the ‘infrastructural power’ of the state. This is “the capacity of the state to actually penetrate civil society, and to implement logistically political decisions throughout the realm” (Mann Reference Mann1984: 189). Put differently, dictators require support from elites to enforce their will across the polity, handling functions like repression, surveillance, propaganda, and other forms of political control necessary, thereby stabilizing the regime and ensuring the leader’s continuing rule (Hassan et al. Reference Hassan, Mattingly and Nugent2022).

Incumbents therefore prepare for succession by building a power base for their successor, or by creating the conditions for their successor to do so (Meng Reference Meng2020: 309). This helps the successor, once in office, gain control of the state to govern and subsequently maintain power.Footnote ⁹ To emphasize, this power base is not intended to counterbalance incumbent-era elites or the military once the successor is in power (De Bruin Reference De Bruin2020); although it helps the successor avoid falling to a coup, it does not do this via traditional means of coup-proofing (Quinlivan Reference Quinlivan1999). Rather, a power base helps the successor command the ‘infrastructural power’ of the state, helping the dictator harness the state’s coercive apparatus to maintain control over the people and persuade incumbent-era elites to refrain from obstruction (Gerschewski Reference Gerschewski2013: 21–2). Thus, a power base helps a successor avoid falling to a coup by helping the successor continue to deliver the functions of autocratic governance that are essential to the regime’s stability.

Building a power base for the successor entails the earlier empowerment of elites who will make up this power base. These elites then have the necessary clout within the regime to enact the successor’s directives. To empower elites, dictators often bestow prestigious tasks upon them, either through formal roles or by involving them in public displays of governance. These strategies signal an elite’s importance within the regime and help them develop expertise. In Egypt, for example, before he was assassinated, Anwar Sadat prepared for succession to Hosni Mubarak (Time 1981). Key elites were empowered to establish a support base for Mubarak. This included Abu Ghazala, who gained valuable experience at dealing with the United States as the defense attaché to Washington between 1976 and 1979. Ghazala was then a key figure for Mubarak as defense minister from 1981 to 1989 (Springborg Reference Springborg1987: 6). The idea that dictators can empower elites has been noted in prior research. Kokkonen et al. (Reference Kokkonen, Møller and Sundell2022: 221) write that “[i]t seems as if the dictator is often able to decide how much to empower the elite.” Building on this, not all elites are empowered to prepare for succession; instead, dictators empower specific elites to form a new power base for the successor to help him govern once in office.

In deciding which elites to tap for the successor’s power base, incumbents and their successors must be careful not to trigger the ‘sovereign’s dilemma.’ Yuhua Wang (Reference Wang2022: 13) describes how a “coherent elite helps the ruler strengthen the state, but threatens his survival.” Wang’s argument suggests that building a power base to strengthen the successor’s grip over the state’s ‘infrastructural power’ could strengthen the state but threaten the successor’s survival. The power base for the successor therefore cannot be constructed from incumbent-era elites. These elites have garnered status and prominence via their presence in the regime’s upper echelons of power and through their interactions with the incumbent. Incumbent-era elites have had the chance to develop important relationships with other prominent elites, as well as acquire significant material resources based on their proximity to the regime’s center of power. Successors are generally not impotent when they arrive in office (Kokkonen et al. Reference Kokkonen, Møller and Sundell2022: 114; Meng Reference Meng2020: 209), but if they have to rely on incumbent-era elites, then they are vulnerable to being ‘captured,’ where elites force their preferences on the successor, or they may even be overthrown by these elites. When Park Chung-hee was assassinated in South Korea in 1979, his prime minister, Choi Kyu-ha, became president. Despite his former position, Choi was a career bureaucrat with no power base (Lee Reference Lee1980: 70). In less than seven weeks, the powerful Park-era elite and head of the Military Security Command, Chun Doo-hwan, instigated a military coup to become South Korea’de fact leader (Kim and Larson Reference Kim and Larson1988: 87).Footnote ¹⁰

Incumbents and their successors therefore alleviate the ‘sovereign’s dilemma’ by building the successor’s power base from elites who are outside their inner circle. These elites are not nobodies; they could be deputy ministers in a party regime or princes in a monarchy who do not sit on the primary governing council. In Kim Jong Il’s North Korea in 2008, examples include Kim’s brother-in-law Jang Song Thaek, who was then not in Kim’s good graces having been purged in 2004 (Ra Reference Ra2019: 121–6), or the KWP official Pak To Chun, who haonl been the party secretary for Chagang Province in the northwest since 2005; both became prominent once Kim started preparing for succession.Footnote ¹¹ Regardless of their professional history or demographic characteristics, individuals chosen for the successor’s power base are not members of the incumbent’s inner circle when preparations for succession commence.

Building a power base of elites from outside of the incumbent’s inner circle alleviates the ‘sovereign’s dilemma’ for the successor (Wang Reference Wang2022). The rapid elevation of these elites motivates them to support the successor for two reasons. First, elites brought into a power base from outside the incumbent’s inner circle are motivated to help the successor govern and solidify his position because their rise in fortunes is indelibly linked to the successor’s continued primacy (Weeks Reference Weeks2008: 41; York Reference York2024). The removal or weakening of the successor would threaten the newfound fortunes of these elites.Footnote ¹² Second, any challenge from these elites to the successor stands a low chance of success initially. These elites’ newfound status in the regime’s center of power means they have had little chance to develop strong vertical relationships with junior officials or strong horizontal relationships with fellow elites who might support them in a power grab (Goldring and Matthews Reference Goldring and Matthews2023). Elites are unlikely to challenge the successor if they believe there is a low likelihood of success since the costs of a failed challenge are so high (Finer Reference Finer1962). Thus, building a power base of elites from outside the incumbent’s inner circle helps the successor wield the ‘infrastructural power’ of the state while not threatening the successor’s personal survival (Mann Reference Mann1984), thereby circumventing the ‘sovereign’s dilemma’ (Wang Reference Wang2022).

This strategy is not without risks. First, empowering a successor who need not be the incumbent’s son could trigger the ‘crown-prince’ problem, where the successor has the “motive and opportunity to mount a coup” (Kokkonen and Sundell Reference Kokkonen and Sundell2014: 440; see also Herz Reference Herz1952: 30). Second, increasing the size of the ruling coalition – a side effect of building a power base for the incumbent while not necessarily purging incumbent-era elites – could be dangerous since it decreases the value of membership of the ruling coalition; this is risky in autocracies because leaders engender loyalty through providing private goods (Bueno de Mesquita et al. Reference Bueno de Mesquita, Smith, Siverson and Morrow2003). Third, incumbent-era elites may react negatively if they believe they are being marginalized by the elevation of an alternative power base for the successor (Svolik Reference Svolik2012: 59). Elite infighting could destabilize the regime as a whole, potentially morphing into open conflict, even civil war (Roessler Reference Roessler2011).

However, there are logical reasons that explain why these possible costs are outweighed by the benefits of this strategy. Taking these points in turn, empowering a successocoul trigger the ‘crown-prince’ problem, but this is unlikely for the same reason that certain institutional approaches tend not to. For example, Meng (Reference Meng2021: 952) describes how identifying a successor via constitutional rules mitigates the ‘crown-prince’ problem because, “once named, the designated successor has a strong incentive to protect the existing regime.” The noninstitutional strategy of developing the successor’s power base is similarly unlikely to trigger the ‘crown-prince’ problem because the successor knows that the incumbent is working to solidify their future. Then, contrary to the possible risks of a larger ruling coalition highlighted by selectorate theory, coup attempts are in fact less likely with larger ruling coalitions (Marcum and Brown Reference Marcum and Brown2016). This brings us to the related point that incumbent-era elites may react negatively to the emergence of an alternative power base. However, structural features of regimes with personalistic incumbents – the type of autocracy to which our argument applies – make this unlikely. Elites in these regimes have imperfect information about the dictator’s goals and behaviors, including whether he is altering the intra-regime balance of power. This, combined with the risks of mounting a challenge, makes it costly for elites to try to stop the dictator. Personalistic dictators and their successors can therefore manipulate elites’ statuses with a low risk of a challenge by incumbent-era elites (Svolik Reference Svolik2012: 55–60).

Incumbents and their successors sometimes also marginalize prominent elites while developing a power base for their successor, but whether they do this is conditional on perceptions of the dictator’s strength among elites. Marginalizing elites through purges may seem appealing, for instance, because it removes potential rivals to the successor; in Syria, Hafez al-Assad sought to secure Bashir al-Assad’s succession in part by purging military elites who were opposed to the succession or had poor relations with Bashir (Leverett Reference Leverett2005: 62). However, purging elites can precipitate a backlash (Sudduth Reference Sudduth2017). For example, Zimbabwe’s Robert Mugabe was removed from office when he attempted to purge powerful elites to strengthen the position of his preferred successor (Beardsworth et al. Reference Beardsworth, Cheeseman and Tinhu2019). Mugabe’s old age and ill health signaled his impending departure and reduced the costs of challenging him (Bueno de Mesquita and Smith Reference Bueno de Mesquita and Smith2017). Thus, whether incumbents and their successors prepare for succession by also purging incumbent-era elites depends on the perceived strength of the dictator among elites. But even for dictators like Kim Jong Il, who was visibly weakened when preparing for succession, the structural conditions of personalistic autocracy mean that they can elevate a power base with a low risk of incumbent-era elites reacting negatively. For incumbent-era elites, being compelled to share the spotlight within the regime is considerably more palatable than being jailed or executed, the likely outcomes of a failed attempt to challenge the incumbent.

The argument that personalistic autocrats and their successors prepare for a transition by creating a power base for the successor builds on recent research on the determinants of succession. First, our claim that incumbents empower a successor echoes arguments about how institutions facilitate succession. Meng (Reference Meng2021: 953) argues that constitutional rules that identify the vice president as the successor increase stability before the transition “by empowering the designated successor.” Second, prior work describes how successors undergo a grooming process (Kokkonen et al. Reference Kokkonen, Møller and Sundell2022: 29, 114; Meng Reference Meng2021). Analogously, by becoming more prominent and attaining greater status during preparations for succession, we argue that certain elites are also groomed to prepare them to become members of the successor’s power base. Third, our claim that ruling coalition membership is likely to change prior to succession complements arguments about elite management around the time of succession. Scholars have described how the ruling coalition is likely to change once a new ruler succeeds (Kokkonen et al. Reference Kokkonen, Møller and Sundell2022: 113; Svolik Reference Svolik2012: 198). We suggest that consequential changes in elite management also occur before the transition.

2.3 Elite Management Following Succession

We assume that successors wish to remain in power once they enter office, based on the broader assumption that all dictators wish to remain in power (Wintrobe Reference Wintrobe1998: 5). There are exceptions to this claim; Brazil’s Castelo Branco, for instance, had no interest in being a dictator, but such cases are rare (Skidmore Reference Skidmore1989: 40). The assumption also does not preclude the possibility that dictators have other goals, relating to specific policies or broader ideologies (Lachapelle et al. Reference Lachapelle, Levitsky, Way and Casey2020). However, no leader, including autocrats, can accomplish wider aims without first maintaining power. We therefore explicitly assume that successors wish to remain in office.

The main threat to a successor’s position once in office comes from elites who are part of the regime (Svolik Reference Svolik2012: 4–5). We assume that these elites wish to preserve their status within the regime that entitles them to private goods, whether through prestigious positions or direct financial payoffs (Bueno de Mesquita et al. Reference Bueno de Mesquita, Smith, Siverson and Morrow2003). The start of a successor’s tenure is therefore a risky time for elites within the regime. Successors are generally inexperienced and untested at leading an autocratic polity (Leber et al. Reference Leber, Carothers and Reichert2023). If their decision-making is poor or they convey an image of weakness, this could make the regime vulnerable to challenges. This could be via a regime change coup that overhauls the state’s existing autocratic institutions or through pressure from the people or foreign states, which could herald the arrival of democracy. In any of these scenarios, elites in favorable positions under the existing autocratic regime would likely find themselves on the outside looking in. For instance, in the Ivory Coast, when Félix Houphouët-Boigny’s successor, Konan Bédié, lost power to a regime change coup in 1999, his prime minister, Daniel Kablan Duncan, was not part of Robert Guéï’s new regime and instead fled to France (CGTN 2023). Thus, elites who observe successors governing in a manner that they believe threatens the autocratic regime’s continued existence might challenge the leader, calculating that the risks of perpetrating a leader reshuffling coup are worth the benefits of preserving their access to the regime’s spoils.

Faced with potentially restive elites, successors need to solidify the transition, which entails establishing themselves as the unquestioned ruler in the eyes of elites within the regime. There are different ways that successors can achieve this. One seemingly appealing option is to institutionalize their regime. This means a successor accepting constraints on their actions and sharing power with elites, which generates a stable equilibrium (Meng Reference Meng2020).Footnote ¹³ However, while institutionalization is an effective way for dictators to extend their tenure, doing so curtails their future freedom of action. Elites should not oppose institutionalization since it bestows greater power upon them, but they may oppose future deinstitutionalization, which is tricky for dictators to pursue thereafter because in the intervening time, elites will have been empowered by the preceding institutionalization. Further, dictators who are strong upon entering office tend not to institutionalize, which implies that dictators prefer not to institutionalize when they have a choice (Meng Reference Meng2020).

An alternative option for successors attempting to solidify the transition is they can seek to consolidate power. Consolidation refers to a leader having a significant share of power relative to elites within the regime (Gandhi and Sumner Reference Gandhi and Sumner2020). Consolidating power helps successors negate intra-regime perceptions of inexperience, unpreparedness, or weakness and reduce the likelihood of an internal challenge (Chin et al. Reference Chin, Escribà-Folch, Song and Wright2022). Successorma arrive in office too weak to consolidate power (Leber et al. Reference Leber, Carothers and Reichert2023), and so are compelled to pursue an institutionalization strategy to survive (Meng Reference Meng2020). However, because successors are often favorably positioned by their predecessor in the manner described in Section 2.2 – that is, through the construction of a power base to help them govern as soon as they come to power – successors in fact have significant leeway to pursue alternative strategies, including attempting to consolidate power.

How, then, can successors consolidate power? Prior research highlights several institutional and structural factors that facilitate autocratic consolidation of power (Geddes et al. Reference Geddes, Wright and Frantz2018: chap. 4; Leber et al. Reference Leber, Carothers and Reichert2023; Timoneda Reference Timoneda2020).Footnote ¹⁴ For instance, Fails (Reference Fails2020) shows that natural resources can aid consolidation by giving a dictator access to cash flows.Footnote ¹⁵ Dictators can buy the loyalty of elites and fund a surveillance state that deters potential elite conspiracies, shifting the balance of power in the dictator’s favor. Alternatively, Leber et al. (Reference Leber, Carothers and Reichert2023) highlight the absence of oversight as critical to whether an autocrat can consolidate power. Specifically, the absence of old guard elites like the former leader increases the likelihood of consolidation. In sum, prior work highlights how structural and institutional characteristics, which dictators are passive experiencers of, affect whether they can consolidate power.

In addition to these factors, dictators have agency and can also proactively try to consolidate power. They can appoint officials to positions of high office, purge elites by jailing or executing them, and utilize various actions in between such as nonviolently dismissing or marginalizing an elite by excluding them from certain events (Goldring and Matthews Reference Goldring and Matthews2023; Shih et al. Reference Shih, Adolph and Liu2012). These tactics fall under the banner of elite management. Tools of elite management speak to the heart of autocratic consolidation since they directly affect the status of individual elites within the regime, which in turn affects the share of power that a dictator possesses relative to elites (Gandhi and Sumner Reference Gandhi and Sumner2020).

The question then, is how do successors manage elites in their regime to consolidate power? First, we contend that they shrink the size of their ruling coalition. Like autocrats in general, successors engender loyalty among key officials by providing private goods, which are more likely to encourage loyalty when there is a smaller number of officials because then finite resources are divided among a smaller number of recipients (Bueno de Mesquita et al. Reference Bueno de Mesquita, Smith, Siverson and Morrow2003). Engineering a smaller ruling coalition through marginalizing once-prominent elites allows successors to step out of the shadow of the former leader and other prominent individuals within the regime (Jiang et al. Reference Jiang, Xi and Xie2024; Leber et al. Reference Leber, Carothers and Reichert2023). Decreasing the size of the ruling coalition therefore helps successors solidify their position as the accepted leader of the regime.

There are risks to this strategy, but they can be mitigated. First, larger ruling coalitions make coup attempts less likely, so diminishing the size of the ruling coalition could be risky (Marcum and Brown Reference Marcum and Brown2016). However, at this point a successor is focused on consolidating power rather than necessarily deterring coup attempts. Although coup attempts can be fatal for any dictator, they can counterintuitively help a successor because they provide an opportunity for the successor to identify disloyal elites and further consolidate power (Timoneda et al. Reference Timoneda, Escribà-Folch and Chin2023; Woldense Reference Woldense2022). Second, a smaller ruling coalition could also make a coup attempt more likely by leaving a small number of powerful officials remaining in the ruling coalition who could later perpetrate a leader reshuffling coup (Chin et al. Reference Chin, Carter and Wright2021). But this risk can be alleviated if these elites can be trusted, which a significant proportion can be because they have been groomed for the successor’s power base. Third, reducing the size of the ruling coalition could remove valuable expertise from the regime and make it more vulnerable to popular or foreign threats (Talmadge Reference Talmadge2015). However, these threats are insignificant in contrast to elite threats (Svolik Reference Svolik2012: 4–5), so consolidation is a more pressing priority. Overall, the risks of reducing the size of the ruling coalition are surmountable and the benefits of consolidation and establishing themselves as the regime’s accepted ruler make this strategy worth pursing for successors.

However, as noted, no dictator governs alone (Svolik Reference Svolik2012: 79). While successors reduce the size of the ruling coalition to consolidate power, they retain some officials to help them govern. They otherwise risk experiencing the fate of dictators like Francisco Macías Nguema in Equatorial Guinea who purged his regime to such an extent that there was barely anyone left to govern or protect him when his nephew, Teodoro Obiang Nguema, overthrew him (Decalo Reference Decalo1989). Since successors are trying to reduce the size of their ruling coalition, it is important that the officials they do retain are loyal. Private goods are intended to induce loyalty (Bueno de Mesquita et al. Reference Bueno de Mesquita, Smith, Siverson and Morrow2003), but successors are also strategic in who they select to retain so that these individuals do not take the resources that come with being a member of the ruling coalition and use them against the dictator.

Retaining loyal supporters is their primary concern but this consideration is challenging since loyalty is a latent trait and cannot be directly measured (Aaskoven and Nyrup Reference Aaskoven and Nyrup2021). Successors can use heuristics, however, to gauge the likelihood of an elite being loyal. This is a common strategy of dictators; for instance, in more ethnically diverse contexts than Korea, dictators sometimes use ethnicity to gauge expected loyalty (Hassan Reference Hassan2020). In the context of succession, recall how in the prior subsection we argued that preparations for succession involve the emergence of a power base for the successor. A successor reasonably expects that elites in this power base are more likely than other elites to be loyal. The sudden rise in prominence of these elites during preparations for succession intrinsically links these elites’ statuses to the successor’s continued primacy (Weeks Reference Weeks2008: 41; York Reference York2024). Assuming they wish to retain this status, these elites are likely to follow the successor’s orders, and they may even defend him if he is challenged. Thus, while successors work to generally reduce the size of their ruling coalition to consolidate power, they tend to retain the aforementioned elites in their power base to help them govern as they attempt to consolidate power.

However, successors do not retain the individuals from their initial power base indefinitely. As successors consolidate power, they eventually also marginalize these individuals. Although they are more likely to be loyal because they were raised up in prominence to support the successor, the timing of their ascent potentially threatens the dictator’s status as the unequivocally most powerful individual in the regime. Since these elites’ experience of top-level politics within the regime is similar to the successor’s, there is a risk that these individuals may use their horizontal relationships with one another, their vertical relationships with junior officials, or their expertise at navigating the bureaucracy – which have all been enhanced by this stage – to outwit the successor (Goldring and Matthews Reference Goldring and Matthews2023). These elites may not seek to remove the successor, but they can use their similar knowledge of the inner workings of the polity to influence policy, build personal fiefdoms, or even marginalize the successor. If he is to consolidate power, the successor cannot risk being seen as the ‘first among equals’ alongside these individuals; he needs to establish himself as the clear leader. Successors therefore need to eventually marginalize these individuals so that they can step out of their shadow (Jiang et al. Reference Jiang, Xi and Xie2024; Leber et al. Reference Leber, Carothers and Reichert2023). Elites appointed to replace them, by contrast, lack these networks and independent power centers, being new in the job. What is more, witnessing what happens to their predecessor makes them more cautious and loyal. In North Korea, this risk was reportedly foreseen by Kim Jong Il. Kim Jong Un’s uncle, Jang Song Thaek, was a key part of the power base developed by Kim Jong Il to help Kim Jong Un govern. However, Kim Jong Il supposedly also warned his sister and Jang’s wife, Kim Kyong Hui, that “Jang’s continued presence at the pinnacle of power would eventually threaten Kim family rule … Thus, from the onset of the Kim Jong-un era, Jang Song-taek’ssi] fate was not a matter of if, but a matter of when” (Gause Reference Gause2015: 54). Overall, successors initially retain the elites raised up by predecessors to help them govern, but over time, as they consolidate power, they marginalize these individuals too so that they can become the undisputed leader of the regime.

This argument about how successors manage elites to solidify a transition makes several contributions to research on elite politics in dictatorships. First, our overall argument builds on recent work that highlights autocratic agency relating to power consolidation. Prior work largely highlights how structural and institutional factors affect whether dictators can consolidate power (Fails Reference Fails2020; Geddes et al. Reference Geddes, Wright and Frantz2018: chap. 4; Leber et al. Reference Leber, Carothers and Reichert2023); instead, like Timoneda et al. (Reference Timoneda, Escribà-Folch and Chin2023), we highlight that there are strategies that dictators can deploy to attempt to consolidate power. Second, our research helps further distinguish exactly why dictators reduce the size of their ruling coalition. Contrary to Bueno de Mesquita et al. (Reference Bueno de Mesquita, Smith, Siverson and Morrow2003), Marcum and Brown (Reference Marcum and Brown2016) show that a smaller coalition could heighten coup risk. On the other hand, we contend that if consolidation is the aim, a smaller ruling coalition is helpful and coup risk can be managed through careful selection of which elites are retained. Further, if dictators are confident that they can defeat a coup attempt, these are counterintuitively helpful for dictators trying to consolidate power (Timoneda et al. Reference Timoneda, Escribà-Folch and Chin2023). Third, while prior work raises the importance of elite management for autocratic power consolidation (Bueno de Mesquita et al. Reference Bueno de Mesquita, Smith, Siverson and Morrow2003; Geddes et al. Reference Geddes, Wright and Frantz2018; Svolik Reference Svolik2012), our argument highlights the importance of target selection and the sequencing of when officials are marginalized. These are important factors to study to better comprehend the micro-dynamics of autocratic consolidation.

2.4 Conclusion

In this section, we provided theoretical arguments about how dictators manage elites to facilitate a leadership transition. This entailed first theorizing how personalistic incumbents and their successors manage elites to prepare for a transition, and then how successors manage elites to solidify a transition. We contended that preparations for succession involve building a power base of elites from outside the incumbent’s inner circle to help the successor stabilize the regime once they come to power. We then argued that successors reduce the size of their ruling coalition to consolidate power. However, successors initially retain elites who their predecessor raised up in prominence to help them govern, but over time they also marginalize these elites so that they can step out of their shadow and become the undisputed most powerful individual within the regime. The next two sections test these arguments using the two leadership transitions in North Korea.

3.1 Introduction

In this section, we assess the argument against the succession from Kim Il Sung to Kim Jong Il. We show that, per the theoretical expectations in Section 2, Kim Il Sung gave his son the scope and capacity to build a power base. Then, after Kim Il Sung’s death, Kim Jong Il reduced the overall size of his support group, thinning it through emasculating officials. However, he did not marginalize everyone; he initially relied on members of his power base, who had risen in prominence between Kim being named the successor and coming to power, to govern.

We take a predominantly qualitative approach, using a wide range of sources, including North Korean defector memoirs, North Korean official sources, and secondary literature. We supplement this with quantitative descriptive data on leadership events, which provide another window into how elites were managed within the regime. Taken together, these diverse sources point to a new narrative that emphasizes how proactive elite management strategies were crucial to Kim Jong Il’s rise and his consolidation of power.Footnote ¹⁷

3.2 Research Design

3.3 How Kim Il Sung Managed Elites to Prepare for Succession

Kim Il Sung was installed by the Soviet Union as North Korea’s founding ruler, but it took time for him to transition the regime from an oligarchy to a personalist autocracy (Lankov Reference Lankov2005: chap. 1). Between 1953 and 1958, Kim Il Sung purged opponents in the party and military (Lankov Reference Lankov2005; Song and Wright Reference Song and Wright2018). The First Party Conference of March 1958 marks when all of Kim’s major factional opponents were purged from the party and military (Buzo Reference Buzo2018: 46; Ward Reference Ward2019). The last remnants of non-state farming and industry were also nationalized by Reference Ward1958, which consolidated the party-state’s grip on society (Scalapino and Lee Reference Scalapino and Lee1972: 1060; Ward Reference Ward and Buzo2020: 79–80).

Kim thereafter built a state in which infrastructural power was concentrated in the party, as is the norm for communist states. Power was centralized in Kim’s hands, through his position as general secretary of the KWP. The positions of minister of defense, minister of internal affairs, secretaries in the party secretariat, and major director positions in the CC apparatus were the principal conduits through which power was ultimately wielded. At the same time, Kim Il Sung was premier and head of the cabinet until 1972, hence the cabinet office was also significant until then in the power structure (Scalapino and Lee Reference Scalapino and Lee1972). After 1958, Kim Il Sung ruled alongside his own faction (former guerrilla partisans from his time in Manchuria) and with the support of a growing technocratic elite (Hyun Reference Hyun2006: 60).

Succession was almost never mentioned in open-access North Korean press before 1980 (Tertitskiy Reference Tertitskiy2022b), but the factional struggles of the mid 1950s were rooted in this issue. Specifically, North Korean elites observed Stalin’s successors initiating a process of de-Stalinization. This led factional opponents of Kim to question his growing power (Lankov Reference Lankov2005; Szalontai Reference Szalontai2005). However, Kim Il Sung did not seriously consider appointing a successor until after 1970 (Tertitskiy Reference Tertitskiy2022b: 34).

Elite North Korean defector accounts indicate that Kim Jong Il was not designated as the successor in his youth, but he started being groomed for high office at university. He received bespoke lessons from members of his father’s personal secretariat and from other members of the elite (Jung Reference Jung2000: 90). Kim Yong Ju, his uncle and Kim Il Sung’s brother, was particularly important in his education and acted as something akin to a guardian (Jung Reference Jung2000: 91; Son Reference Son2004: 40).Footnote ¹⁹ Kim Jong Il also got the opportunity to listen in on cabinet, as well as meetings of the party CC, politburo, and military officials (Jung Reference Jung2000: 93; Son Reference Son2004: 45). With his uncle’s support, he also became actively involved in the management of Kim Il Sung University through its party committee (Son Reference Son2004: 45–6). These activities were early apprenticeships for high office, and thereafter, Kim Jong Il was given positions in the Organization and Guidance Department (OGD) and then the Agitation and Propaganda Department (APD) of the central party. Appointments to official positions and media messages are controlled by the OGD and the APD, respectively (Scalapino and Lee Reference Scalapino and Lee1972: 740).Footnote ²⁰ These roles therefore allowed him to build networks in two of the most powerful departments in the central party bureaucracy and of the entire party-state: (1) its human resources department (OGD) and (2) its public relations department (APD).

Kim Jong Il was confirmed as successor in 1974; his father then began openly supporting his policy initiatives and standing within the party state.Footnote ²¹ The organizational positions that Kim Il Sung bestowed on Kim Jong Il also allowed the latter to build up substantial personal power and a power base of supporters. Some of these individuals had been friends or allies for far longer. Unlike Kim Jong Un after him, Kim Jong Il attended high school and university in the country and had a large network, which was used to staff the upper echelons of the security sector and the party. Kim utilized the powers of appointment he had through his control of the OGD, and his father largely deferred to his son’s choices (Jung Reference Jung2000: 123, 163).

Consequently, when preparations for succession were underway in the second half of the 1970s as well as the 1980s, a power base for Kim Jong Il began to form. Kim Jong Il cultivated an entire generation of future high-level party, state, and military officials. Some were close confidants he picked up post-1964, after he started working in the central party apparatus, others rose to prominence later. North Korea did not just become a hereditary dictatorship at the level of the leader but also among those who surrounded Kim Il Sung in the 1930s and 1940s. This second-generation group of elites emerged in the 1970s, and by the early 1990s, much, but not all, of Kim Il Sung’s original group had died or retired (Hyun Reference Hyun2006: chap. 4; Lee Reference Lee1992, Reference Lee1995, Reference Lee2000; Jung Reference Jung2000, 240–52). Many of these elites achieved elevated status through connections to Kim Jong Il, while some were already at the lower echelons of the party-state bureaucracy when Kim Jong Il became the successor. Kim Jong Il did not pluck these officials from nowhere, but they were relative outsiders, junior officials with positions in the bureaucracy, which accorded them some power and resources that they could call on in his service.

With the support of his father, having created a group of loyal lieutenants in the OGD and the APD, Kim Jong Il then built a financial infrastructure of foreign trade enterprises (with large industrial and extractive affiliates) and banks under a new office, Office 39, staffed by loyalists to fund his emerging leadership (Kim Reference Kim2008: 53–67). He also created his own secretariat that helped him coordinate and concentrate power (Hyun Reference Hyun2006: 142–3), and he established the Ministry of State Security (MSS), the country’s secret police, as a new and separate state agency. The MSS fast became the most feared of agencies for its power to monitor and surveil the population, party officials, and elites for anti-system sentiments or conspiracies (Kwak Reference Kwak2018: 19–21). Kim made a close confident from the time, Ri Jin Su, the minister of state security. Another close elite, Sim Chang Hwal, was given the head position at the Ministry of Public Security’s Political Bureau, the country’s ordinary police, further enhancing Kim’s control over society (Hyun Reference Hyun2006: 150).

Kim Jong Il’s succession was publicly unveiled at the Sixth Party Congress in 1980.Footnote ²² The 1980s then saw substantial changes to the composition of the ruling elite; the CC was increasingly filled with people personally loyal to Kim Jong Il rather than those associated with his father’s rule (Kim and Lee Reference Kim and Lee2009). For instance, So Yun Sok, party chief for Pyongyang in the 1980s, joined the CC in 1980 alongside O Kuk Ryol (a key military elite), while Jon Pyong Ho (a munitions expert) did so in 1982. The 1980 CC was more than 70 percent new, though turnover was usually high in the Kim Il Sung era (Lee Reference Lee1982: 443). Thus, the changes to the composition of North Korea’s elite that were underway behind the scenes from 1974, which had increasingly populated North Korea’s key institutions of governance in the party-state with Kim Jong Il’s supporters, spread upward to the public-facing CC in the 1980s.

After the succession, the North Korean elite was increasingly dominated by the military; indeed, much of Kim Jong Il’s public retinue were senior military personnel. Yet it was only in the early 1990s that Kim Jong Il began building an elite power base in the military, after his father appointed him as head of the Korean People’s Army (KPA). Much of Kim’s power base in the military was composed of close confidants who he worked with more closely after 1991 (Hyun Reference Hyun2006: 150, 156n247). But it is worth stressing that Kim Jong Il’s power base was initially primarily more in the party, although in the early 1990s he built closer ties with members of the military beyond some of his father’s more senior military confidants.

The evidence of how preparations occurred for the Kim Il Sung to Kim Jong Il transition highlights the difference between our argument and previous research about how autocrats prepare for succession. Prior comparative research highlights how incumbent autocrats use institutions, such as formal rules regarding succession, to prevent coup attempts against them once it becomes apparent that they will depart from office (Kokkonen and Sundell Reference Kokkonen and Sundell2014; Meng Reference Meng2021; see also Hummel Reference Hummel2020). The probability of a coup attempt against Kim Il Sung after he identified Kim Jong Il as successor in 1974 was low, but this does not mean Kim Il Sung was unconcerned by the possibility of one (Yi Reference Yi2003). In fact, the opposite was true; the likelihood of a coup attempt against Kim Il Sung was low preciselbecaus he was concerned by this possibility, a fear that was enhanced by watching elite fractures in China related to Mao’s equivocation over succession (Jin Reference Jin1999: 54–61). However, the empowerment of Kim Jong Il through the key positions in the APD and the OGD, and the cultivation of a key group of supporters within the party-state, was not designed to directly protect Kim Il Sung from a coup attempt. This was done to provide Kim Jong Il with the support to govern once he came to power.

Overall, with the support of his father, Kim built multiple pillars of a power base that encompassed the party, security services, and the military. In a country where the ‘party decides,’ having a tight grip on party personnel and Agitation and Propaganda was essential. Further, in a consolidated authoritarian state, secret police are essential to maintaining the state’s infrastructural power over society. The same is the case for the military, a major threat to authoritarian regimes generally. Kim Jong Il was given the space, time, and support from his father to build a power base over decades, drawing upon networks he created within the major organizations of the party-state.

3.4 How Kim Jong Il Managed Elites to Consolidate the Transition

Kim Jong Il became North Korea’s leader, after Kim Il Sung’s death in 1994, amid domestic and foreign upheaval. On the domestic front, the country’s food production and distribution system was collapsing, a huge issue by itself but one that also posed internal security problems for Kim Jong Il. With a famine about to engulf the country, Kim was forced to ask the outside world for aid, having refused to reform the country’s economic system (Haggard and Noland Reference Haggard and Noland2007). At the same time, the country’s major political and economic sponsor, the Soviet Union, had recently collapsed, and much of the country’s imports subsequently disappeared (Eberstadt et al. Reference Eberstadt, Rubin and Tretyakova1995). The famine led to a breakdown in public order, with Kim complaining to officials in late 1996 that the country was reduced to “anarchy” because of the food crisis (Kim Reference Kim1997: 311). This led to illegal internal migration within North Korea, as people moved around the country, leaving their homes and places of work without permission in search of food, straining systems of state surveillance (Kim Reference Kim1997).

On the international scene, Kim came to power as the country was backing away from the brink of war with the United States. North Korea began pursuing nuclear technology in the 1950s, the country’s nuclear scientists experimented with separating plutonium in the mid 1970s, and they undertook conventional explosive tests in the early 1980s to perfect trigger technology for use in nuclear weapons (Panda Reference Panda2020). The North became a signatory to the Nuclear Non-Proliferation Treaty (NPT) in 1985 under pressure from Moscow but continued pursuing nuclear weapons to the point where the United States was close to attacking the country in June 1994 to prevent it from going nuclear (Wit et al. Reference Wit, Poneman and Gallucci2004). Ultimately, the two sides retreated from the brink and in late 1994 they signed the Agreed Framework, which committed North Korea to dismantling its plutonium production complex in return for light water reactors and political normalization with the United States (Lawrence Reference Lawrence2020).

Amid this tumultuous backdrop of famine and a tense international situation without its former and principal patron, Kim Jong Il sought to consolidate power. He did not do this through explosive moments of internal upheaval; comparative research highlights how dictators use a victory over real or imagined coup attempts as an opportune moment to consolidate power (Timoneda et al. Reference Timoneda, Escribà-Folch and Chin2023). Though there was an alleged coup plot prior to the succession in 1992 – the Frunze academy incident – Kim consolidated power after 1994 through subtler but systematic tactics of elite management.Footnote ²³

First, Kim Jong Il reduced the size of North Korea’s ruling coalition. Unlike his father, who appreciated official, large-scale meetings of senior officials to facilitate decision-making, Kim Jong Il preferred to rule through a small coterie of officials whom he frequently invited to on-the-spot guidance trips and secret parties (Hyun Reference Hyun2006; Park Reference Park2017; Son Reference Son2004).Footnote ²⁴ This is visualized with descriptive data on the average number of attendees who attended leadership events in North Korea under Kim Jong Il.Footnote ²⁵ Figure 1 shows that as soon as Kim Jong Il took power, the number of officials invited to leadership events shrunk drastically, going from more than ten officials on average at an event in July 1994 to less than five by 2000.

Figure 1 Lowess smoothed average number of elites at leadership events under Kim Jong Il, 1994–2011.

In practice, Kim Jong Il shrunk his ruling coalition through two means. The first, generational replacement, is a tactic that previous accounts of how Kim Jong Il consolidated power highlight (Jeon Reference Jeon2000; Kim Reference Kim2009; Park et al. Reference Park, Lee, Jung and Lee2004). According to these narratives, Kim Jong Il took limited steps to try to consolidate power because he was wary of the power wielded by influential allies of his father, whose status stemmed from having fought alongside Kim Il Sung in Manchuria. For instance, O Jin U was part of Kim Il Sung’s anti-Japanese guerrilla activities in Manchuria. He was Minister of the People’s Armed Forces from 1976 and was able to remain so until his death in 1995 (Kim Reference Kim2009: 128). Overall, this body of scholarship describes how Kim Jong Il was only able to reactively shrink his ruling coalition to consolidate power when powerful elites died from natural causes.

However, Kim Jong Il systematically employed another subtler tool of elite management to effectively shrink his ruling coalition – that is, reduce the number of elites with notable power whose support he needed to rely on to maintain office – to consolidate power for himself. Specifically, Kim emasculated elites from the Kim Il Sung era, allowing them to retain official positions and roles that appeared to denote them with status, but concurrently he removed any substantial independent power from them by reallocating or diluting their responsibilities. Kim Jong Il applied this strategy to many incumbent elites from Kim Il Sung’s inner circle. Elites like Pak Song Chol, Ri Jong Ok, Kim Chol Man, and Kim Yong Ju retained apparent high formal standing as members of the CC and sometimes the politburo but without actual power or real influence. These institutions increasingly became ‘retirement homes’ for old officials. What this meant in practice was that these institutions continued to exist on paper, and membership conferred privileges and prestige, but they ceased meeting and stopped being the institutions through which Kim Jong Il issued instructions or sought advice from (Lee Reference Lee2023: 188).

However, Kim Jong Il did not emasculate all elites. As our theory predicts, he retained important figures in the power base that he built up with Kim Il Sung’s support between 1974 and 1994 to help him govern in his early years. This included younger members of the Manchurian guerrilla faction, party officials from the departments that Kim Jong Il came up through (the APD and the OGD), and members of the security state. Among the Manchurian faction, several members of the so-called ‘1.5 generation’ of former young guerrillas played key roles in helping Kim govern. Ri Ul Sol headed the KPA Bodyguard Command, a key counterbalancing institution, while Jo Myong Rok led the KPA Airforce and then the General Political Bureau under Kim Jong Il. Ri and Jo became members of Kim Jong Il’s power base after he ascended to Supreme Commander of the Armed Forces in 1991 and became head of the NDC in 1993 (Hyun Reference Hyun2006: 150).Footnote ²⁶ Many children of Manchurian guerrillas also helped Kim Jong Il gain a stranglehold over the regions, parts of the central party apparatus not under his direct control (such as the security organs, key heavy industries, and the youth league), and the military. Individuals who came up through the OGD and the APD with Kim held important positions across the state under his leadership, especially in party finance and espionage (Hyun Reference Hyun2006: 138–51), while Won Ung Hee headed the KPA’s security command (its counterespionage branch) from 1989 to shortly before his death (Hwang, I. Reference Hwang2006); like Jo Myong Rok and Ri Ul Sol, Won became close to Kim Jong Il after Kim’s elevation to head of the armed forces. Finally, several of Kim’s family and relatives who gained key positions between 1974 and 1994 also become prominent figures after Kim’s ascension. Most notably, Kim Kyong Hui and Jang Song Thaek, his sister and her husband, respectively, held various economic briefs in the party and key roles in the OGD in Kim Jong Il’s early years. In short, Kim did not take a scythe to the entire North Korean elite after taking power. Instead, he strategically emasculated certain figures whose power or status derived solely from their association with Kim Il Sung, and to govern he relied on the individuals who had been cultivated into his power based once he was named successor in 1974.

Even these officials though were not safe indefinitely. Our theory suggests that once the successor has stabilized the regime, he should marginalize members of his power base since their prominence in the regime predates the successor taking power. Consistent with this logic, this is what happened in the late 1990s, once Kim Jong Il navigated the worst of the famine and his father’s mourning period. Many members of the power base – younger members of the Manchurian guerrilla faction, princelings, party officials from the OGD and APD, and members of the security state – were purged or marginalized and replaced by new individuals particularly from lower echelons of the military.Footnote ²⁷ A significant portion of Manchurian guerrilla-related elites were purged in the late 1990s, including So Yun Sok and Ri Pong Won (Ri Reference Ri2012: chap. 2; Yi Reference Yi2003: 142). Choe Ryong Hae, the son of a key Manchurian guerrilla (Choe Hyon), was also purged on charges of corruption in 1998, after emerging as a key youth leader in Kim’s early years (Jeong Reference Jeong2019).Footnote ²⁸ From the ‘1.5 generation,’ Lee Du Ik was also marginalized after 1998. Among the officials in Kim’s power base who originated in the OGD and APD, many who subsequently held positions in the crucial security sector under Kim Jong Il were purged in the late 1990s. Kim Yong Ryong, the vice minister of state security, was executed in late 1997 (Paek Reference Paek2005).Footnote ²⁹ Kwon Hui Gyong, a senior official in the MSS and a key figure in the early stages of the nuclear program, latterly managed espionage directed toward South Korea, but he was also purged in early 1998 (Kim and Paek Reference Kim and Paek2009). Even in Kim’s family, no one was safe if they posed a possible threat to Kim’s consolidation of power. In 2004, Kim purged his own brother-in-law, Jang Song Thaek, for allegedly undermining his authority (Ra Reference Ra2019: 124).Footnote ³⁰

One alternative explanation for these purges is that they were driven by idiosyncratic factors. Each purge featured allegations of misconduct or the target being scapegoated: So Yun Sok and many others supposedly sabotaged the country’s agricultural system trigging the famine of the 1990s; Ri Pong Won allegedly conspired against other elites; Choe Ryong Hae reportedly engaged in sexual impropriety and financial misconduct; Kim Yong Ryong was said to be in contact with South Korean spies; while Kwon Hui Gyong was allegedly insubordinate and corrupt (Jeong Reference Jeong2019; Kim and Paek Reference Kim and Paek2009; Kwak Reference Kwak2018: 23; Ri Reference Ri2012: chap. 2; Yi Reference Yi2003: 142). Each case, however, highlights how power can be misused by elites in ways that undermine the dictator, thereby showing that Kim Jong Il needed to act to eliminate the perception of there being alternative independent power centers within the regime. Purges can serve multiple purposes, but overall, purges of elites in Kim Jong Il’s power base helped him further consolidate power after having already emasculated his father’s key allies.

3.5 Conclusion

In this section, we assessed our arguments using the Kim Il Sung to Kim Jong Il transition in North Korea. Utilizing a wide range of sources, including North Korean defector accounts, North Korean official sources, data from North Korean media, and secondary literature we have shown, first, how Kim Il Sung gave his son crucial positions and opportunities to develop a network of key contacts in the core organizations of the North Korean state, which gave him a power base that helped him govern in his early years. Second, the evidence shows how, once in office, Kim Jong Il shrunk the size of North Korea’s ruling coalition through emasculating officials. To govern, Kim initially leaned heavily on the officials in the power base that he developed between 1974 and the early 1990s, but many of these officials were cast aside in the late 1990s as Kim further consolidated power.

4.1 Introduction

We now test our arguments on the transition from Kim Jong Il to Kim Jong Un. Per Section 2, we expect that once Kim Jong Il started preparing for succession, he should have facilitated the development of a power base for Kim Jong Un. Then, in Kim Jong Un’s early years, while we should observe him reducing the overall size of his ruling coalition, he should initially retain the officials who Kim Jong Il raised up in prominence to help Kim Jong Un govern in his early years.

We adopt a primarily quantitative approach in this section, taking advantage of available data on leadership events in North Korea as well as biographical data on hundreds of North Korean elites in this more recent period. We present descriptive statistics and substantive findings from regression models in the main text but relegate much of the technical discussion to the Supplementary Materials to aid readability.Footnote ³² We also supplement the quantitative analyses with qualitative data on North Korean elite politics to illustrate the processes behind the statistical results.

4.2 Research Design

4.3 How Kim Jong Il Managed Elites to Prepare for Succession

In North Korea under Kim Jong Il prior to 2008, military politics were dominant. The evidence in Section 3 showed that over time Kim began to purge or marginalize members of the power base that Kim Il Sung facilitated the creation of, including younger members of the Manchurian guerrilla faction, princelings, and party officials from the OGD and APD. Many of these were replaced by individuals from the lower echelons of the military. Thus, once Kim started preparing for succession after his stroke, military elites were more central in his inner circle. Wright’s (Reference Wright2021) research on the structures of autocratic rule show how different latent features vary over time within a regime; in North Korea before 2008, this meant a tilt toward the military. This shift toward military elites in the ruling group occurred post-1991 as Kim Jong Il shifted his base of support as the country entered a period of crisis (famine). This was given an ideological basis with the ‘Military First’Songu, 선군) ideologySongu meant that the NDC’s power outstripped its purview in the North’s 1998 revised constitution, and it became “the de facto general administration” (Gause Reference Gause2011: 119). This was also reflected in the relative status of military elites within Kim’s inner circle. For instance, elites in the KPA were Kim’s “favored bodyguard” (Jeon Reference Jeon2009: 197).

Consistent with the prominence of military elites prior to Kim’s stroke, once Kim started preparing for succession after August 2008, it initially appeared as though the military, especially the KPA, would be the foundation for the succession. Indeed, actions were taken after Kim’s stroke to boost Kim Jong Un’s standing within the military. A propaganda campaign in the military from December 2008 preceded Kim Jong Un being named a four-star general by April Reference Kim and Lee2009 (Cheong Reference Cheong2010: 169; H. Kim Reference Kim2015: 179; Thae Reference Thae2018: 277). Members of the security services, including the military, were required to swear allegiance to Kim Jong Un in early Reference Kim and Lee2009. Then, in the latter half of Reference Kim and Lee2009, heads of the military’s General Political Bureau and other security institutions began reporting to Kim Jong Un (Cheong Reference Cheong2010: 171). Kim Jong Un is also believed to have overseen several military actions to test whether the military would respond to him. The sinking in March 2010 of thCheona, a South Korean warship, and the November 2010 shelling of South Korea’s Yeonpyeong Island are thought to have been ordered by Kim Jong Un in part to quiet voices of discontent about the succession and ensure that the military would obey him (Gause Reference Gause2011: 167, Reference Gause2015: 165n357; Haggard et al. Reference Haggard, Herman and Ryu2014: 776; H. Kim Reference Kim2015: 176; Shin Reference Shin2018: 200; Shin Reference Shin2020: 57–66).

However, Kim Jong Il’s efforts to prepare for succession are closely linked to the resurgence of KWP officials. The KWP subsequently became the main institutional vehicle through which Kim Jong Un’s legitimacy flowed. KWP cadres increasingly occupied prominent positions in important KPA and KWP bodies including the Central Military Commission, NDC, and the politburo. This helped extend Kim Jong Un’s influence over the country’s then primary institutions (Gause Reference Gause2011: chap. 6; Haggard et al. Reference Haggard, Herman and Ryu2014).

In contrast to military elites, party officials were relatively outside of Kim Jong Il’s inner circle before his stroke.Footnote ³⁷ Kim relied on some KWP elites prior to August 2008, including his sister Kim Kyong Hui, but KWP elites were generally not as close as military elites to the regime’s center of power. For example, all the personnel sitting on the NDC in 2003 were connected to the military. After Kim’s stroke, however, several KWP officials became members of the NDC (Haggard et al. Reference Haggard, Herman and Ryu2014: 783–4). Our theory suggests that when preparing for succession, Kim should have integrated elites from outside his inner circle into the ruling coalition to build a power base for Kim Jong Un. This is precisely what we observe in the leadership event data, based on identifying party elites as generally being outside of Kim’s inner circle prior to August 2008. Table 1 shows differences in the means of the proportion of events that different elites attended before and after Kim’s stroke. Consistent with the theory, KWP elites attended 2.77 percent of events on average before Kim’s stroke, and 7.19 percent after. This change is significant at the 99 percent confidence level. Further analysis that accounts for various alternative explanations suggests that party elites were 70.58 percent more likely to attend a leadership event after Kim’s stroke than they were before.Footnote ³⁸

The increased presence of party elites at leadership events reflects systematic changes in the substantive importance of these elites’ roles who, in general, were previously outside of Kim Jong Il’s inner circle. Senior officials entrusted with managing party finances, party control of the military, party control of the military industrial complex (and nuclear program specifically), party work in the capital, human resource issues for both the elite and lower-level cadres, and officials tasked with political oversight of the security services became especially prominent. For example, Han Kwang Sang, who oversaw KWP finances, a crucial role as the party started to take greater control over KPA finances during this period, was regularly invited by Kim Jong Il to events in 2010 (four) and 2011 (eight), despite having never previously been invited. Another example is the rise of Kim Chang Sop, who was head of the political bureau within the Ministry of State Security (the political police). Kim Jong Il also suddenly started inviting Kim Chang Sop to events in 2010 (three) and 2011 (twenty-one), despite his previous absence. These elites are illustrative of the broader trend of party elites, who were previously outside of Kim Jong Il’s inner circle, being raised up in prominence after Kim Jong Il’s stroke and going on, as we describe in Section 4.4, to play key roles in Kim Jong Un’s inner circle in his early years.

Our theory also contains an implication for how Kim Jong Il should have managed military elites after his stroke. We posited that incumbents preparing for succession do not purge elites in their inner circle if the incumbent is visibly weakened, generally due to ill health or old age, when preparing for succession. Kim was visibly weakened by his stroke in August 2008. He therefore should not have attempted to overtly undermine the status of military elites, who we identified as the key elites in his inner circle when he started preparing for succession. Consistent with this, the difference in means (in Table 1) for military (and cabinet) officials is smaller and not statistically significant. Further analysis via regression models provides additional support for this finding.Footnote ³⁹

Thus, alongside the rise of KWP officials, KPA officers remained prominent in elite circles. Many combat field officers tasked with leadership roles, as well as the political officer corps, remained a key element in the ruling coalition. As such, many KPA officers continued to be invited to military and nonmilitary events, likely to provide feedback on policies or ideas for new initiatives. These included senior military men like Kim Yong Chun (defense minister from February 2009 to April Reference Kim2012), Hyon Chol Hae (head of the Political Bureau of the National Defense Commission from May Reference Kim2010 to April Reference Kim2012), and Kim Jong Gak (first deputy director of the KPA General Political Bureau from February 2009 to April Reference Kim2012). There were also some prominent new faces like Ri Yong Ho (the former head of Pyongyang’s capital defense command who became chief of general staff (February 2009 to July 2012), and Kim Won Hong (head of the KPA security command from September 2003 to February 2010 and deputy director of the KPA General Political Bureau from February 2010 to April Reference Kim2012). Nonetheless, these officers were associated with Kim Jong Il’s base of power and were less important in providing a power base to help Kim Jong Un stabilize the regime.

One alternative explanation for the rise and enduring prominence of KWP elites between August 2008 and December 2013 is that Kim Jong Il wanted to focus on policies that KWP elites could help with before his death,Footnote ⁴⁰ and then there was relative stasis in elite turnover while Kim Jong Un found his feet early in his tenure. However, KWP elites were not just more likely to attend events in general after Kim Jong Il’s stroke; they were also more likely to attend military events specifically.Footnote ⁴¹ The argument that Kim Jong Il increasingly invited KWP elites to events after his stroke because he wanted to focus on policies that KWP elites could help with does not explain the increasingly likely attendance of KWP elites at military events after Kim’s stroke. Further, Kim Jong Un purged a high-profile military elite – Ri Yong Ho – in July 2012 (Gause Reference Gause2015: 28). It is generally easier to purge civilian than military elites (Goldring and Matthews Reference Goldring and Matthews2024); thus, if Kim Jong Un was willing to purge military elites, then it is unlikely that Kim Jong Un kept around party elites, whose prominence increased after Kim Jong Il’s stroke, only because he was less sure of his position in his early days.

Another alternative possibility is that KWP elites became more prominent between 2008 and 2011 because Kim Jong Il was guarding against a military coup when he was visibly vulnerable after his stroke (Bueno de Mesquita and Smith Reference Bueno de Mesquita and Smith2017). However, there are several reasons to doubt this explanation. First, Kim Jong Il did not invite party elites more frequently after his stroke to expand the size of the ruling coalition and enhance the visibility of a formal institution – in this case, the KWP – to signal strong support (Timoneda Reference Timoneda2020); the findings are robust to accounting for the number of elites invited to an event.Footnote ⁴² Second, the type of elites more frequently invited did not increase coordination costs for would-be conspirators (De Bruin Reference De Bruin2020: 20–2; Matthews Reference Matthews2022: 667). Inviting KWP elites to events more frequently would not affect coordination costs since military elites could plot at other locations away from the prying eyes of KWP officials. Existing coup-proofing measures also already make a coup attempt unlikely. The military has a split chain of command, which inhibits large troop movements without consent coming from multiple sources. Leadership of the defense of Pyongyang and the leader’s bodyguard are also both run separately and distinctly from the regular military chain of command (Ko Reference Ko2007: 123; Oh Reference Oh2012: 129–30). The embedding of political officers and security command officers (from North Korea’s internal military secret police) within the military also makes a coup attempt unlikely (Tertitskiy Reference Tertitskiy2022a). Kim purged some civilian elites after his stroke – notably the party finance director Pak Nam Gi – but there was little reshuffling of military personnel or military elite purges, both of which we would expect if Kim had been coup-proofing (Park Reference Park2011).

Overall, elites who were outside Kim Jong Il’s inner circle before his stroke – KWP elites – became more prominent once Kim started preparing for succession. This evidence is consistent with our argument’s logic – that this was to facilitate the construction of a power base for Kim Jong Un to help the younger Kim stabilize the regime once in power – and there are reasons to doubt the plausible alternative explanations for the quantitative findings. In the following section, we provide additional qualitative evidence from Kim Jong Un’s tenure that is consistent with the idea that Kim Jong Il was facilitating the construction of a power base for Kim Jong Un.

4.4 How Kim Jong Un Managed Elites to Consolidate the Transition

Kim Jong Un’s tenure as North Korea’s leader began on December 17, 2011, when Kim Jong Il died. The global context was not as dangerous as when his father came to power in 1994, following the collapse of the country’s major sponsor (the Soviet Union). However, Kim Jong Un still had to reckon with a moribund economy, extensive sanctions, and a thriving South Korea that all too many North Koreans were increasingly aware was not the capitalist failure that North Korean propaganda claimed. Additionally, Kim Jong Un came to power at the tender age of twenty-seven, after going through an accelerated grooming process since August 2008 to prepare him for succession. As we have seen, many outside observers doubted that he possessed the leadership skills or experience to be accepted as the leader by North Korea’s elites (Bennett and Lind Reference Bennett and Lind2011: 84; Byman and Lind Reference Byman and Lind2010: 72; Cha Reference Cha2011; Gause Reference Gause2011: 113).

Against this context, Kim Jong Un sought to consolidate power. The December 2013 purge of his uncle, Jang Song Thaek, is often held up as evidence of Kim consolidating power. But, as discussed, Kim’s ability to do this indicated that he had already made significant progress toward this goal. Kim sought to consolidate power not just through the headline-grabbing tactic of violent purges but through the more subtle tactic of gradually reducing the size of his ruling coalition. This was not immediately evident because Kim Jong Un, unlike his father, sought to restore the primary decision-making bodies of the KWP, including the politburo, CC, and even the party congresses (Park Reference Park2017: 113–21).Footnote ⁴³ However, over Kim Jong Un’s first two years, there was a substantial thinning of the ranks. In a system where, as discussed, face time with the leader confers power, elites were invited less often to events in the build-up to Jang’s purge, and Kim relied increasingly on a small coterie of advisers from the KPA – Hyon Chol Hae (First Vice Minister of People’s Armed Forces from April 2012); Ri Myong Su (Minister of Public Security from April 2011 to February 2013); Kim Jong Gak (Minister of People’s Armed Forces up to November 2012); and Jang Jong Nam (KPA First Division Chief in December 2012, Minister of People’s Armed Forces from May 2013 to June 2014) – and the KWP – Hwang Pyong So (OGD Vice Director); Pak Thae Song (Party Central Committee Department Vice Chair); and Ma Won Chun (Party Central Committee Department Vice Chair).

Kim Jong Un’s efforts to reduce the size of his ruling coalition and consolidate power are reflected in attendance patterns at leadership events between December 2011 and December 2013. Figure 2 shows the probabilities of all elites, as well as different types of elites, being invited to a leadership event over Kim Jong Un’s first two years in power.Footnote ⁴⁴ Figure 2 shows that in Kim’s first month in office (December 2011), an elite had a probability of 0.6 of being invited to a leadership event. This was just over 0.1 a year later, and one further year later it was less than 0.05. This reflects the emergence of Kim Jong Un’s ‘kitchen cabinet’ of close advisors, like that which Kim Jong Il relied on (Park Reference Park2017: 108–13).

Figure 2 The probabilities of different types of elites attending an event under Kim Jong Un, December 2011 to December 2013.

Findings from regression models suggest that the consolidation of power and corresponding shrinking of Kim Jong Un’s ruling coalition in his first two years applies to cabinet and party elites, but possibly not military elites.Footnote ⁴⁵ However, it is worth noting that, although data from South Korea’s Ministry of Unification (MOU) shows that between the end of 2011 and the end of 2013 only 1.34 percent (21 out of 1,564) of one- to four-star generals lost their position, there were important changes to the composition of military elites (MOU 2011: 215–25, 2013: 227–36). Ri Yong Ho, the chief of the general staff, was purged (reportedly executed) in July 2012, the minister of defense was changed multiple times, a small number of other generals were demoted or lost their jobs, some major positions were subject to frequent change, and some senior, old military officials retired (Kim, T. Reference Kim2015: 151, 153, 170). Thus, Kim may have been careful to keep the military on side, but he also sought to reduce the military’s influence, including over foreign trade assets.

Like his father though, Kim Jong Un did not marginalize all elites in his first two years. Consistent with the theory in Section 2, Kim Jong Un initially leaned on the elites who Kim Jong Il raised up to help Kim Jong Un govern. More specifically, the majority of KWP elites who Kim Jong Il raised in prominence played important roles in helping Kim Jong Un govern between 2011 and 2013, unlike KWP elites who Kim Jong Il did not raise up in prominence.Footnote ⁴⁶ They carried out key tasks to help Kim Jong Un monitor elites, control the army, and subdue the people. Table 2 summarizes the primary roles that KWP elites, who were raised in prominence by Kim Jong Il after his stroke, held in Kim Jong Un’s early years.Footnote ⁴⁷ To monitor other elites, Kim Kyong Ok was the first vice director of the OGD from 2008, a position he continued to hold in Kim Jong Un’s early years. The OGD manages KWP human resource issues, including appointments, promotions, and disciplinary procedures (see Section 3). To quell potential threats from elites more broadly, Choe Ryong Hae and O Kuk Ryol helped provide control of the army, through their respective roles as chief of the KPA’s General Political Bureau and vice chair of the NDC. Several elites also played key roles in the development of Kim Jong Un’s personality cult, notably Kim Ki Nam as KWP Secretariat Agitation and Propaganda Secretary. In terms of helping Kim govern and wield the power of the state, one task was notably absent among these elites; none of the elites that Kim Jong Il raised in prominence were involved in the Ministry of Public Security, essentially North Korea’s police (Gause Reference Gause2015: 266), although as KWP Administrative Department Director, Jang Song Thaek managed human resources in the security apparatus. Further, Kim Chang Sop was the Political Bureau Director of the Ministry of State Security from 2009, which is North Korea’s secret police and operates the concentration camps. Thus, between Kim Chang Sop, Kim Ki Nam, and Jang Song Thaek, Kim Jong Un had supportive allies who could provide control over the people. Overall, elites who were raised in prominence by Kim Jong Il played important roles in helping Kim Jong Un gain control over the ‘infrastructural power’ of the state (Mann Reference Mann1984).

In contrast, as summarized in Table 3, only a few KWP elites who were not prominent in the post-stroke period were significant under Kim Jong Un pre-December 2013. Ri Thae Chol held the important position of First Vice Minister of the Ministry of Public Security, Kang Tong Yun held a significant role in the OGD, while several of these KWP elites moved to notable posts within the military. However, these were exceptions. The largest portion of KWP elites who were not raised up in prominence by Kim Jong Il played comparatively minor roles in Kim Jong Un’s early years. Many held regional positions away from Pyongyang, others’ primary roles were members of bodies that sound important but contain hundreds of members (the KWP Central Committee or the Supreme People’s Assembly), while several held relatively insignificant roles (e.g., Choe Chil Nam as the Journalists’ Union Central Committee Chair, Pak Kyong Son as North Korea-Laos Friend Association Chair, and Kim Yong Dae as the Korean Social Democratic Party Central Committee Chair).

As time passed though, Kim Jong Un even dispensed with KWP elites who Kim Jong Il had raised up in prominence for the former’s power base. This included people like Choe Thae Bok (KWP education secretary) and Kim Kyong Hui (his aunt and the KWP light industry chief), neither of whom were purged, but both of whom became far less prominent post-succession. Thae Jong Su (KWP General Services), Ju Yong Sik (Chagang province chief secretary), and Han Kwang Sang (party finances) were all demoted in 2012 or 2013. While some very high-profile and important purges occurred post-succession, like that of Jang Song Thaek and his close associates, much of the power base was just gradually pushed out of the limelight and then eventually retired off or died in office.

How Kim Jong Un initially leaned on the KWP elites brought into a power base during his father’s final years before dispensing with them as he consolidated power is reflected in the leadership event data. Figure 3 shows the probabilities of a KWP elite attending an event in Kim Jong Un’s first two years, conditional on whether they were prominent in the period between Kim Jong Il’s stroke (August 2008) and his death (December 17, 2011).Footnote ⁴⁸ Figure 3 shows that, despite the earlier evidence that party elites were less likely over time to attend an event under Kim Jong Un, this was not initially the case for party elites who were prominent under Kim Jong Il after his stroke. In Kim Jong Un’s first month, the probability of a party elite who had been prominent under Kim Jong Il after his stroke attending an event was about 0.95. This decreases marginally to just above 0.8 during Kim Jong Un’s first year, but then rapidly decreases after that to about 0.45 by the time Kim Jong Un purges Jang Song Thaek. In sum, consistent with our theory, party elites who became increasingly prominent under Kim Jong Il when he was preparing for succession played key roles under Kim Jong Un after he initially took power, before he later marginalized them too.

Figure 3 The probability of party elites attending an event, December 2011 to December 2013.

4.5 Conclusion

In this section, we tested our arguments against the Kim Jong Il to Kim Jong Un transition in North Korea. Using novel data on thousands of leadership events in North Korea between 1994 and 2013, combined with original biographical data, we found evidence to support our arguments that (1) the outgoing dictator (Kim Jong Il) built up a power base of elites from outside their inner circle to help the successor (Kim Jong Un) govern and (2) although the successor reduced the size of his ruling coalition upon taking office, he initially refrained from marginalizing elites in his power base who could help him govern in his early years.

5.1 Introduction

The evidence from North Korea provides a comprehensive evidence base for the internal validity of how dictators manage elites to facilitate transition. But do these findings apply to other autocracies where personalist incumbents attempt to hand power to their chosen successor? We assess the external validity of our argument with qualitative case vignettes of elite management surrounding leadership transitions in party-based China, military Egypt, and monarchical Saudi Arabia. We select these cases to assess whether the arguments apply irrespective of ideology and institutional makeup, and whether the succession is hereditary.

5.2 Case Vignettes

5.3 Conclusion

The evidence of how personalist autocrats and their successors manage elites to facilitate succession in these party-based, military, and monarchical regimes largely supports our argument. There are some discrepancies. King Salman and MbS have treated existing elites more aggressively than our theory suggests they should, although this may be a misstep on their part. It may also be harder for successors to marginalize elites to consolidate power when the predecessor is still alive (per the case of Jiang Zemin). Beyond these points though, the case vignettes are consistent with our theory, suggesting it explains how dictators manage elites to facilitate succession in autocracies with personalistic incumbents from across the world.

6.1 Implications

Autocrats manage elites in a distinct manner when they prepare for succession. More broadly then, do an autocrat’s actions vary based on the leader’s time horizon? This points to important questions for further research, which we return to in Section 6.3, but it also suggests that analyzing autocratic elite management can yield insights about the likelihood of future policy shifts and even the probability of leadership turnover occurring as well as forecasting the identity of the successor.

The evidence also highlights the importance of understanding the mechanisms behind observable behaviors in dictatorships to avoid making mistaken inferences about how these opaque regimes operate. Timoneda (Reference Timoneda2020) and Woldense (Reference Woldense2022) provide compelling evidence that dictators increase the size of their ruling coalition at times of vulnerability as a coup-proofing measure. However, this Element shows that this is not the only reason why dictators might increase the size of their ruling coalition. Autocrats may also enlarge their ruling coalition to build a power base for the successor. The multiple mechanisms that drive this possible outcome illustrate the importance of gathering insights from insider sources, where possible, to better understand what strategies drive behavior within autocracies.

Finally, the Element shows how systematic analysis of leadership event data can be instructive for practitioners. The prominence or marginalization of officials is sometimes evident at leadership events before institutional changes are publicly known. In North Korea, one recent example of this is the military elite, Pak Jong Chon, who received a big promotion to vice marshal in May 2020. Pak was known to analysts of North Korea before May 2020, but his importance was not fully realized. However, his rising status was visible prior to May 2020 in the leadership event data as he was the most prominent elite in North Korea between January 1, 2020, and April 12, 2020 (Kim Jong Un’s last event prior to his disappearance that year; see Section 6.2).Footnote ⁵⁷ Elsewhere, Qin Gang’s removal as China’s foreign minister in July 2023 was precipitated by his disappearance for a few weeks from public events (McDonell et al. Reference McDonell, Fraser and Ng2023). These kinds of insights have practical applications. For example, the prominence of elites at public events, including diplomats, may correspond to them having the favor, trust, and authority of the leader when conducting negotiations. Hence, diplomats consistently prominent at leadership events should be the most fruitful counterparts in negotiations. Analysts should of course monitor other sources, including formal membership of institutions, to glean insights into elites’ roles and status. But our work shows the value for policymakers, including those involved in Track One and Track Two diplomacy with North Korea, of systematically analyzing leadership event data.

6.2 Future Succession

The rise of Kim Jong Un and his consolidation of power surprised many observers. However, rumors have swirled since the Covid-19 pandemic about Kim’s health and the long-term viability of his rule. His extended disappearance from public in 2020 prompted speculation about his health. Analysts speculated that internal party changes in 2021, including the creation of a deputy party leadership position (a new First Secretary position), are mechanisms for emergency succession; these added to the perception that Kim Jong Un’s leadership was more fragile than it had hitherto appeared (Lankov Reference Lankov2021; Zwirko Reference Zwirko2021). Concurrently, the growing prominence of Kim Jong Un’s younger sister, Kim Yo Jong, led some to suggest she might be thde fact deputy or that she may become the interim leader if Kim Jong Un suffers an untimely and sudden demise (Cha and Katz Reference Cha and Katz2023).

Perhaps most dramatic, however, has been the sudden emergence of Kim Jong Un’s daughter, Kim Ju Ae, at public events alongside her father from late 2022 (O’Carroll and Reddy Reference O’Carroll and Reddy2022). It is not yet clear what role Kim Ju Ae will play, but her appearances raise questions about the regime’s leadership planning. Indeed, there are signs that she is being groomed for a leadership role, with the North Korean press describing her as one of two “great persons of guidance” alongside her father, using terminology generally only associated with her father, grandfather, and great-grandfather (Cheong Reference Cheong2024; Rodong Sinmun Reference Sinmun2024). Yet she is believed to have been born in 2013, and so will not legally become an adult until around 2031. Even then, she would almost certainly need further education before she could be anointed successor. However, her grandfather’s and father’s ascents to power indicate that it might not be that long before she could theoretically be anointed as successor.

Kim Jong Il was thirty-two when he was secretly appointed successor and thirty-nine when he was publicly unveiled. He then had fourteen years supporting his father’s leadership before Kim Il Sung died in 1994 (Cheong Reference Cheong2011: chap. 2). Kim Jong Il received training and gained experience in politics and administrative decision-making while in university, and then his father helped him build a power base in the central party apparatus. He also had valuable childhood experiences in the country, being schooled alongside other members of the top elite, helping to build valuable friendships from a young age (see Section 3). By contrast, Kim Jong Un was partially educated overseas and was not well integrated into elite social networks as a child. He was hurriedly designated successor at twenty-four, unveiled at twenty-six, and began running the country at twenty-seven.

Neither Kim Ju Ae’s father nor her grandfather made public appearances prior to being publicly announced as successor. Hence, her appearance, if she is to become the successor, places her in uncharted territory. Kim Il Sung was not afraid to appoint direct blood relatives like his brother (Kim Yong Ju) and his uncle (Kang Ryang Uk) to positions of power. Similarly, Kim Jong Il gave his sister (Kim Kyong Hui) a prominent position in the central party, as has Kim Jong Un (Kim Yo Jong). Thus, it may be that Kim Ju Ae will become another family member handed positions and responsibilities within the regime, rather than being the future leader. Another important factor to consider is the prominent role that Kim Jong Un has thus far accorded his wife Ri Sol Ju (Tertitskiy Reference Tertitskiy2018). Kim Jong Un appears to be interested in cultivating the image of a loving husband and father. The emergence of his daughter may therefore also be part of such image building, but it appears that this is not his sole intention.

Coverage of her appearances in the North Korean press is instructive on this point. Analysis of photographs and accompanying descriptions of her appearances in the press indicate that Kim Ju Ae is being vested with a higher status than her aunt, for instance, who is a high-level party official (Cheong Reference Cheong2023; Tertitskiy Reference Tertitskiy2023). The decision to present her meeting with senior military officers and show them deferentially bowing to her also appears to signal special status, unmatched by other family members who have held important positions.

Despite these signs, however, it remains to be seen whether Kim Jong Un is planning a succession, and whether his daughter will become his preferred candidate. If the last two successions are a guide, the identity of the successor was not set in stone until a candidate has been designated internally. However, if Kim Ju Ae is being groomed to take the reins, and what has come before is a guide, there are several patterns of elite management that our research suggests we might observe. First, Kim Ju Ae would need a power base of loyal officials to help her wield the power of the state as she embarks on succeeding her father. This would likely include officials working in the party’s human resources department (the OGD), communications specialists (the APD), the security services, and munitions experts (the Military Industry Department). Today, the OGD is headed by one of her father’s most trusted lieutenants, Jo Yong Won, while her aunt is deputy director of the APD. Second, Kim Ju Ae’s associations with the people charged with overseeing these departments will likely deepen in the coming decade(s), and she may receive practical experience working in these departments, perhaps under the stewardship of her aunt or other close confidants of her father. Third, succession would likely not lead to a significant loss of power for the military. Just as Kim Jong Il refrained from marginalizing military officials when preparing for succession (see Section 4), Kim Jong Un would also likely be similarly wary of attempting to sideline military elites given their currently prominent status in the regime as well as Kim’s possible health issues.

A fourth generation of a ruling family would be almost unprecedented for modern dictatorships. However, North Korea’s Kim regime has well-established practices of elite management that could allow Kim Ju Ae to become North Korea’s next and first female leader.

6.3 Further Research

This Element’s findings point to several fruitful areas for further research. First, future work should examine whether the elite management strategies that we identify are effective at facilitating succession. While recent prior research on autocratic leadership succession has examined the effectiveness of institutionalized rules on the likelihood of succession occurring (Kokkonen and Sundell Reference Kokkonen and Sundell2014; Kokkonen et al. Reference Kokkonen, Møller and Sundell2022; Meng Reference Meng2020, Reference Meng2021), our goal was more circumspect. We simply sought to identify how incumbents manage elites to prepare for succession, and how successors manage elites to secure the transition once they come to office. Evidence from North Korea suggests that these tactics are effective, or at least not ineffective, but such claims are currently speculative and should be tested. Collecting additional elite-level data across dictatorships can facilitate testing of these relationships alongside alternative explanations of how autocratic regimes pull off succession.

More generally, scholars have only scratched the surface of succession in autocracies. A second important question is to examine how less personalistic autocrats prepare for succession. Our argument applies exclusively to personalistic incumbents because it relies on autocrats being able to manipulate the status of elites. But non-personalist autocrats also often try to influence the identity of their successor. In heavily institutionalized Vietnam, for instance, the less personalistic Nong Duc Manh helped position Nguyen Phu Trong as his successor (Koh Reference Oh2012: 366). Understanding the noninstitutional strategies that such dictators employ can lead to a fuller picture of the strategies that autocrats employ to prepare for succession.

Finally, future work should examine policy effects related to succession. For instance, how is policy affected by impending succession as autocrats near the end of their expected tenure, either due to their age or because they are adhering to formalized rules that imply that they should leave office? Do autocrats become more or less risk-averse? Are there types of policies or goals that they are more likely to pursue before leaving office? Related work examines how external actors’ views of an autocrat’s position affect their policies toward autocracies (e.g., Bak Reference Bak2016), but autocrats are also influential actors so understanding how their time horizon affects their policy decisions is worth studying. As Pepinsky (Reference Pepinsky2014: 650) writes, “authoritarian regimes do many things besides grow/stagnate and survive/collapse.” Studying succession’s effects on policy can improve our understanding of how these regular events affect the lives of the billions of people who live under autocracy.