‘No longer what you used to be’: Renegotiating relations between de facto states and their patrons

Abstract

The dependence of de facto states/breakaway-territories on external patrons is well-established, but we lack detailed knowledge of the dynamics of these hierarchical relationships, including opportunities for client agency. Through a comparative study of two patron–client relationships that had reached a crisis point, this article examines how the weaker client responds when at odds with its patron. Drawing on original data, including archival sources and interviews, it provides a bottom-up analysis of the relationship between Serbia and Republika Srpska after the signing of the Dayton Peace Agreement and between Armenia and Nagorno Karabakh immediately after the 2020 war. The out-of-equilibrium behaviour presents an opportunity for analysing the underlying mechanisms of patron–client relationships. In both cases, the crisis resulted in deep mistrust of the patron, but the de facto states were unable to sever links. They were constrained by a lack of reliable alternative supporters, continued patron dependence and powerful ethno-nationalist narratives. Non-recognition added to the client’s weakness, yet we still find clear examples of agency: external supporters were courted, and patrons were pressured through obstruction or appeals to its domestic forces. What resulted was a renegotiated, less hierarchical relationship. The article adds to our understanding of patron–client relationships, including those involving non-sovereign entities where an effective patronage monopoly still exists. It contributes to the existing literature with typologies of patron–client crises and client strategies for reclaiming autonomy, and with detailed empirical evidence of client agency and constraining factors, and by demonstrating the potential leverage of some of the weakest international actors.

Keywords

de facto states patron–client hierarchy state sovereignty conflict international relations

What happens when a patron fails to deliver on key promises, or if a patron and its client diverge on existential issues? How much autonomy does the less powerful client have in this hierarchical relationship? Moreover, how are these dynamics affected by a context of non-recognition? Patron–client relationships are commonplace in international relations, but the framework is rarely applied beyond a state-centric focus. However, such relationships are of significant importance in conflicts involving de facto states (Berg and Vits, 2018; Kolstø, 2020; Kosienkowski, 2019). Nearly all breakaway territories owe their continued survival to the support of a powerful patron. This patron, usually a neighbouring state, provides economic, diplomatic and, most importantly, military support that sustains their de facto independence. Examples include Russia’s support for Abkhazia and South Ossetia (Georgia), Armenia’s support for Nagorno Karabakh (Azerbaijan) and Turkey’s support for Northern Cyprus. Despite the importance of this relationship, its dynamics and underlying mechanisms remain under-analysed.

The relationship between de facto states and their patron is akin to a contractual relationship in which the contested entities, and their leaders, receive vital resources in return for relinquishing a degree of autonomy. The underlying dynamics share many similarities with relationships between small- and medium-sized states and their hegemonic powers. In both cases, the fluid nature of the relationship reflects, and affects, the contestation and re-ordering of regional and global hierarchies (see, for example, Adler-Nissen and Zarakol, 2021). Due to the covert nature of some forms of patron support for the de facto states, the relationship comes closest to what has been termed an informal hierarchy (Hobson and Sharman, 2005; Lake, 1996) although there may be more formalised elements, especially if the patron recognises its client’s independence. These relationships are defined by the power imbalance between the patron and the client, but they are also mutually beneficial, and this reciprocity sustains the patron–client relationship (see, for example, Carney, 1989; Shoemaker and Spanier, 1984). However, behind the shared interests, there is often ‘an informal reality of tensions, disagreements and shifting directions of influence’ (Bercovitch, 1991). Such tensions can reach a crisis point (Kosienkowski, 2019), which presents a moment of significant danger for de facto states, but also an opportunity for change. As Schoemaker and Spanier (1984: 8) point out, crises are the key mechanism for change in patron–client relations as they ‘dispel the shrouds that cloak the extent of commitment of both patron and client’. Moreover, as Lake (1996: 9) argues, an informal hierarchical relationship, becomes more evident in ‘out-of-equilibrium’ behaviour: the patron controls become manifest only once the client tests its dependence.

In this article, we examine how de facto states respond when they are at odds with their patrons: if they no longer share the same goals, or the patron is unwilling or unable to provide the resources they need to survive. Do they find another more reliable external supporter, do they have the leverage to effect a change in the patron’s position, or are they able to renegotiate the relationship in other ways? We examine these questions through an in-depth comparative study of two cases of patron–client relations that had reached a crisis point, two examples of ‘out-of-equilibrium’ behaviour: (1) the relationship between Armenia and Nagorno Karabakh following the significant territorial losses suffered in the 2020 war, and (2) between the Federal Republic of Yugoslavia (Serbia) and Republika Srpska following the signing of the 1995 Dayton Peace agreement. Following the 2020 war, Nagorno Karabakh found itself in a much weaker position: facing severe threats to its security and suspecting its patron of being willing to accept Azerbaijani control over the entity. The Dayton agreement, which led to the reintegration of Republika Srpska into Bosnia-Herzegovina, was negotiated by the patron against the wishes of the entity’s leaders and fell short of the independence they had demanded. The empirical analysis draws on a range of primary and secondary sources, including interviews and archival sources.

Most of the existing literature on de facto states and their external patrons has drawn on single-case studies (see, for example, Kanol and Köprülü, 2017) and stressed the client’s significant dependence (Kolstø and Blakkisrud, 2008). The agency of the de facto state is sometimes noted: they actively seek support from the patron (Kolstø, 2020) and they have various strategies for countering the patron’s influence (Caspersen, 2008). However, there is a gap in the literature when it comes to more detailed analyses of the relationship and its underlying mechanisms, including opportunities for the client to increase its autonomy, as well as constraints on such agency and catalysts for change. The patron–client framework has predominantly been used for analysing Cold War relationships between the competing superpowers and their client states (see, for example, Ladwig, 2017; Westad, 2005). These clients were constrained by their dependence on the patron (see, for example, Carney, 1989; Rodman, 1997), yet the potential leverage of the weaker states is frequently mentioned in this literature: the potential to defect to the other superpower gave the client useful bargaining power (see, for example, Wong, 2020). More recent applications of the framework, to hierarchical power relations between small- and medium-sized states and global powers, also emphasise the client’s leverage, again with a focus on alternative patrons and possible goods substitution (see, for example, Adler-Nissen et al., 2021; Cooley and Nexon, 2020; Veenendaal, 2017). We expect the context of non-recognition to result in even more constrained agency for the de facto states. It leaves them with fewer alternative external supporters (Caspersen, 2012), and loss of support would threaten their survival as de facto independent entities (Berg and Vits, 2018).

Analysing the impact of a profound challenge to the patron–client relationship helps us uncover the dynamics of this relationship and the options available, or not, to de facto states. By moving beyond the dominant state-centric focus of the patron–client framework (Biermann, 2024), the article adds to our understanding of hierarchical relationships in the international system, including those involving non-sovereign entities, and the agency of some of its weakest actors. The article examines a range of strategies deployed by clients and the constraints faced. This includes but also goes beyond the usual focus on the availability of alternative patrons. The cases of Republika Srpska (1995–1997) and Nagorno Karabakh (2020–2022) are well-suited for a comparative analysis of how de facto states respond when at odds with their patrons. In both cases, the de facto authorities had previously demonstrated considerable agency, sometimes openly defying their patron (Caspersen, 2008). The patron states were both kinstates, and the shared ethnicity had been argued to give the client more leverage (Caspersen, 2008; Waterbury, 2021). However, the two cases differ on variables that we would expect to affect their responses to the patron, including the type of patron–client crisis experienced, the severity of the security threat faced, the degree of economic dependence on the patron and the availability of alternative external supporters.

The article proceeds as follows: we first develop a theoretical framework for analysing the agency of de facto states, including a typology of patron–client crises and possible client responses, and we identify factors expected to affect the agency of the weaker party. We then present our case selection and the methods used. The empirical analysis falls in the following three parts: (1) reactions in the de facto states; (2) strategies for protecting autonomy or ensuring greater patron compliance, and the constraints faced; and (3) the dynamics of the renegotiated relationship. Despite differences between the two de facto states, we find striking similarities in their responses. In both Nagorno Karabakh and Republika Srpska, the perceived failure of the patron resulted in deep mistrust, yet the de facto states were constrained in their ability to defy or break from their benefactor. This was a result of their continued dependence on the patron, when it came to security, the economy or the political survival of individual leaders. We show how these constraints are compounded by non-recognition. We also demonstrate the importance of factors not usually explored in the patron–client literature, including the constraining influence of ethno-national ties. Although the agency of the client is constrained, it is not without significance, and we point to important drivers of change in the relationship. Despite the weakness of the client, what we see in both cases is a renegotiated, less hierarchical patron–client relationship.

Hierarchy, dependence and agency in Patron–client relationships

Most authors define patron–client relationships according to four interlinked characteristics: reciprocity, asymmetry, compliance and affectivity (Veenendaal, 2017; see also Kaufman, 1974). The relationship is based on a degree of reciprocal exchange ‘both immediate and in the coming future’ (Hicken, 2011: 292). It is a form of a ‘vertically structured pact’ (Holden, 2004: 20) in which the patron provides resources the client lacks. Most patron–client literature has focused on how the patron guarantees the security of its weaker client in an unsafe environment (Shoemaker and Spanier, 1984: 14). However, patrons also provide diplomatic and economic support, thereby helping the client build internal and external legitimacy (see, for example, Veenendaal, 2017). Due to their contested status, de facto states face a persistent threat of forceful reintegration and external military support is vital for their survival. The patrons also supply other resources needed for the entity to survive (see, for example, De Waal, 2018; Kolstø, 2020). This includes helping to fill the resource gap resulting from relative international isolation, for example, by providing diplomatic and consular networks, mobility of goods and people, telecommunications and international banking (Blakkisrud and Kolstø, 2012; Blakkisrud et al., 2021). The benefits the patron gets in return are more varied but may include strategic advantage and ideological convergence (Shoemaker and Spanier, 1984: 18–19; Carney, 1989: 49).

Despite this reciprocity, the relationship is hierarchical and highly asymmetrical. Patrons have far greater capacity than their clients do, and the benefits derived from the relationship are also unbalanced: for the client, the support from the patron is often of crucial importance, whereas for the patron, the client is often merely an asset (Veenendaal, 2017: 567). For de facto states, and their leaders, patron state support is a question of survival (Berg and Vits, 2018), while the patron will weigh support for its client against other demands and counter-pressures (Caspersen, 2008). The hierarchical nature of the relationship is reinforced by the de facto state’s lack of international recognition. Sovereignty as a norm prohibits relations of authority between states (Lake, 2017), but de facto states are not even equal to their patron in a formal sense.

Both asymmetry and reciprocity are linked to the third characteristic of patron–client relations: compliance. The client willingly puts aside some aspects of its agency, without coercion from the patron, based on the expected benefits of the relationship. This makes the patron–client relationship distinct from one of dependency (Carney, 1989: 43). Although highly unequal, it is a ‘bargaining relationship in which each state tries to extract from the other valuable concessions at a minimal cost’ (Shoemaker and Spanier, 1984: 24). Like other clients, the de facto state will court the patron state to ensure continued support, while also trying to reduce the costs associated with the relationship. The de facto authorities will often publicly express their gratitude to the patron. For example, after Russia recognised Abkhazia’s independence in 2008, the entity proclaimed President Dimitry Medvedev a hero (Bullough, 2008). However, this gratitude is often mixed with wariness of the patron’s influence. Too much compliance is costly; it could confirm general perceptions of de facto states as puppets and undermine both internal and external legitimacy (Caspersen, 2015; Kolstø, 2020).

Finally, patron–client relationships have an element of affectivity. A form of loyalty develops between the patron and the client, which enhances the stability of the relationship (Veenendaal, 2017: 565). Hierarchical relationships are not simply based on power and coercion, but also rely on authority: the subordinate state accepts the legitimacy of the unequal situation and the resulting obligation to comply (Hobson and Sharman, 2005: 68–69). For de facto states, this affectivity is sometimes based on shared ethnicity (Caspersen, 2008).

Tensions, conflict and client responses

Even though the relationship is marked by both reciprocity and affectivity, we rarely find complete consensus between the patron and the client (Caspersen, 2008), and tensions and disagreements are commonplace (Bercovitch, 1991). Moreover, goals and priorities may change over time, for example, as a result of a leadership change or a shift in resource base. The relationship can also be put under strain by changes in the wider geopolitical context (Shoemaker and Spanier, 1984: 48ff). Open support for a breakaway territory can be associated with significant costs for the patron. For example, international sanctions were imposed against the Federal Republic of Yugoslavia, due to its involvement in the wars in Croatia and Bosnia (Caspersen, 2008). The patron may decide to alter its support for the client if the relationship becomes increasingly costly, if the patron faces a reduction in capacity, or if other objectives take priority. Similar dynamics have been observed in relations between rebel groups and their external supporters, if the conflict escalates out of control and jeopardises the patron’s position (Salehyan, 2010: 507).

Any conflict with the patron presents a risk to the client (see, for example, Caspersen, 2010; Gueudet, 2018). However, the severity of the threat depends on the type of patron–client conflict, as well as the ability of the client to respond. Below we first outline three different patron actions, which present different levels of threat to the client, and we then examine strategies for the client to protect its autonomy or ensure greater patron compliance.

At the extreme end of the spectrum, the patron may choose to completely cut-off its support for the de facto state. Such abandonment would, in many cases, pose an immediate existential threat to the client. For example, one of the main causes of the violent downfall of Republika Srpska Krajina, in wartime Croatia, is argued to be its loss of patron state support (Kolstø and Pauković, 2014: 9). A more common response would be a reduction in economic or military support. For example, both the Georgian breakaway regions, Abkhazia and South Ossetia, claim that the level of economic support they receive falls short of what they were promised when Russia recognised their independence (Berg and Vits, 2018). Such a reduction in support may not pose an immediate threat to the survival of the de facto state, but it could raise doubts about its future sustainability. Even more damaging, the patron could choose to reduce its military support or simply not have the capacity to protect the entity effectively.

Alternatively, the patron may try to make the relationship less costly by forcing its client to moderate its position. The asymmetric nature of the relationship gives the patron significant leverage, and it could use the threat of reduced support to ensure compliance, for example with a negotiated settlement that was otherwise rejected by the de facto leadership (Karlén, 2017; see also Lake, 1996). This would threaten the separatist goals of the client and constrain the autonomy of its leaders. The ability of the patron to force its client to comply is aided by the frequent inclusion of external patrons in peace talks, on behalf of or in addition to their unrecognised clients. For example, Armenia negotiated on behalf of Nagorno Karabakh and Turkey is involved in the Cyprus peace talks as one of the original guarantor states.

In all three types of patron–client conflict, the patron alters the terms of the pact or contract, and we would expect a reaction from the client. The reciprocity of the patron–client relationship is reduced, and this could lead to a rethink of the relationship and trigger the de facto state to assert its independence. Although there is a lack of detailed analysis of such strategies in the existing patron–client literature, the following can be identified.

Most of the literature has focused on the strategy of defecting or seeking alternative sources of support. If the client is unhappy with its external benefactor, it could abandon the patron entirely and replace it with another, more supportive state (Veenendaal, 2017: 567). Courting alternative patrons may be enough to protect its autonomy, as even the possibility of defection provides the client with leverage (Cooley and Nexon, 2020; see also Wong, 2020). Veenendaal (2017) has shown how micro-states use their vote in international organisations as a bargaining chip that can be awarded to the highest bidder. Sharman (2017) describes this as ‘à la carte hierarchy’ and notes the unpresented latitude weak actors enjoy as a result. In the context of de facto states, Kosienkowski (2019) argues that Transnistria, in response to reduced Russian support, had some success pretending to seek an alternative patron. An alternative strategy is for the client to take advantage of domestic dynamics within the patron state to protect its autonomy, for example, by getting supportive officials, activists, politicians or military officers to lobby the patron government (Bercovitch, 1991; Kosienkowski, 2019). They could also make use of the affective bond between the two states by appealing directly to the government and to public opinion in the patron state (Bercovitch, 1991). A third and final option would be to make it more costly for the patron to get the client to comply. This strategy is not generally explored in the patron–client literature but follows from the overall argument and is also supported by the emphasis on the cost of governance, or control, in the literature on hierarchical relations (e.g. Lake, 1996). The client could do this either by reducing its dependence on the patron, for example, by seeking alternative forms of external support, or by using existing resources to obstruct efforts to enforce compliance. If successful, the result would be a less hierarchical patron–client relationship (Lake, 1996).

The choice between these responses is likely to depend on the type of conflict as well as the client’s perception of the patron’s intentions. Did the patron lack the capacity to protect them, and their demand for independence, or did they deliberately choose not to? The possible reactions would range from disappointment and a sense of abandonment, to anger and accusations of betrayal. Finally, it will depend on the client’s bargaining power and the degree to which its agency is constrained. This is discussed below.

Factors constraining or enabling client agency

Asserting its autonomy comes with significant risks for the client: the de facto authorities may enhance their internal legitimacy but could simultaneously undermine their chance of retaining independent statehood (Caspersen, 2015). De facto states must navigate the tension between dependence and agency (see also Comai, 2018), and the authorities tend to know in which areas they must toe the line (Caspersen, 2023; see also Ladwig, 2017). However, some de facto states are more constrained than others when it comes to bargaining with their patrons.

In the patron–client literature, the following inter-related factors are argued to be of importance. First, the degree of client dependence matters: the more dependent the client is on trade, aid or protection, the more responsive it will be to patron state pressure (Ladwig, 2017: 57). Significant security threats increase the client’s dependence. As Rodman (1997: 97) points out, a client faced with the threat of war will seek to tighten its relationship with its patron, as it needs additional capabilities, and will be willing to cede autonomy in return. Moreover, the threat of war also increases the patron’s strategic interests and makes it more willing to enforce compliance (Rodman, 1997: 35). The flip side of this is the value of the relationship to the patron: the more valuable the client, the more likely it is to be able to sway the patron (Ladwig, 2017: 212; see also Jacobsen, 2020). A second factor is the level of control the patron has over the client. This includes the patron’s ability to influence who governs the weaker actor (Ladwig, 2017: 290). Client agency is constrained if the local leaders depend on the patron for their political survival, for example, in cases of ‘double clientelism’ in which domestic patron–client networks are financed with resources supplied by the external patron (Veenendaal, 2017: 567). Third, and finally, the availability of alternative patrons is significant: the more potential patrons, the greater the client’s leverage as they can play them off against each other (Bercovitch, 1991; Veenendaal, 2017). Cooley and Nexon (2020) argue that the West enjoyed a de facto patronage monopoly in the 1990s, which meant that developing countries could not exert leverage by threatening defection as they had done during the Cold War. However, the decline in US hegemony has ended this monopoly and the agency of clients has increased as a result (see also Adler-Nissen et al., 2021). If clients are unhappy with the benefits provided by the patron, or the autonomy they must give up in return, they can seek alternative patrons as ‘geopolitical hedges’ (Cooley, 2019).

What is the likely effect of non-recognition on the client’s agency? De facto states all enjoy territorial control and have access to other resources that can be used to resist enforced compliance. Nevertheless, we would expect their agency to be more constrained than recognised states. First, the significant security threat faced by many de facto states makes them highly dependent on patron support: as Berg and Vits argue it is often a case of ‘do or die’ (Berg and Vits, 2018: 393). Second, their lack of recognition and resulting international isolation significantly limits the availability of alternative patrons, or other sources of external support (Kolstø, 2020; Kosienkowski, 2019). That said, their international isolation is relative, not absolute. Many de facto states benefit from diaspora support (Caspersen, 2012) and may also be able to achieve forms of engagement, without recognition, from other recognised states (Berg and Vits, 2018: 398; Ker-Lindsay, 2013; see also Dembińska and Mérand, 2019). Moreover, they may have leverage that is not available to other weak actors, due to the benefits the patron derives from the relationship. Although the de facto states cannot use their sovereignty as a bargaining chip, their strategic value is often significant (see, for example, Graham and Horne, 2012: 10) and affectivity is likely to be important if the patron is a kinstate. The shared ethnicity gives the client potential leverage, since the patron government could face a domestic backlash if seen to abandon the co-ethnics it had vowed to protect (Caspersen, 2008; Waterbury, 2021).

Even though we expect de facto states to have some bargaining power, de facto state must tread carefully not to lose vital support. The following in-depth comparative case study of Republika Srpska and Nagorno Karabakh allows us to examine these dynamics in depth: how to de facto states react to a perceived failure of the patron to honour its side of the bargain, what constraints do they face and what are the dynamics of the resulting relationship?

Case studies: Republika Srpska and Nagorno Karabakh

Both de facto states saw significant patron state involvement from the start, but also tensions and occasional open conflict, culminating in a crisis in the patron–client relationship. These changes in the relationship had many sources, including shifting geopolitical circumstances which constrained the two patrons. However, the focus in this article is the client and their responses to the crisis. The two cases share many similarities but also some important differences, which make them well suited for a comparative study of client agency and leverage.

Republika Srpska (RS) was proclaimed on 9 January 1992, but patron state involvement predated this declaration. The Republic of Serbia (hereafter Serbia) – a component of the Socialist Federal Republic of Yugoslavia and, from 1992, the Federal Republic of Yugoslavia – played a crucial role in the initial ethno-nationalist mobilisation and arming of the local hardliners (Caspersen, 2010). The President of Serbia, Slobodan Milošević, had full control over the Yugoslav institutions, including the security apparatus (Ramet, 2006),¹ and the patron provided vital military support to its client (Caspersen, 2010; Gueudet, 2022: 31), although always maintaining ‘plausible deniability’ (Vukušić, 2022). Dependence continued during the war, but the relationship became increasingly conflictual from May 1993 when Serbia began using coercive measures to make the RS government accept successive peace plans. These plans would guarantee the lifting of international sanctions against the patron, but for the client it would mean significant territorial losses and an outcome that fell short of its demand for independence (Caspersen, 2010). Following pressure from the US envoy, Richard Holbrooke to control the RS leadership, Milošević brokered the so-called Patriarch Agreement, under the aegis of the Serbian Orthodox Church. This agreement, signed on 30 August 1995, endorsed Milošević as the lead negotiator in any upcoming peace conference (Donia, 2014: 476; Gueudet, 2018). The resulting Serb delegation to the Dayton peace talks was joint in name only (see also Chollet, 2007: 75). The top military and political leaders of RS, Ratko Mladić and Radovan Karadžić, were already excluded by their ICTY indictments, and Milošević also excluded the RS President and Chief of Staff from the delegation. In any case, the RS delegation had few opportunities for influencing the outcome as negotiations were carried out behind closed doors between Milošević and Holbrooke (Donia, 2014: 476). One member of the RS delegation, Nikola Koljević (2014: 544), recalls their consternation that the patron leader negotiated and signed a peace agreement that entailed the reintegration of RS into the Bosnian state without their approval. Other firsthand testimonies similarly refer to frustration, discontent and chaos on the sidelines (Chollet, 2007: 180).

We see similarly strong patron state involvement in the other case study. The Nagorno Karabakh (NK) conflict originated in the contested territory, but the local forces received sustained military support from their kinstate Armenia. The two leaderships did not always see eye to eye (Caspersen, 2008), but the strong patron support enabled the entity to achieve de facto independence from Azerbaijan (Caspersen, 2008). Following the 1994 ceasefire, Armenia supported NK’s call for full (de jure) independence and described itself as a ‘guarantor and supporter of security’ for the Karabakh population. It provided both weaponry and training for the local forces and deployed conscripts to the entity (Smolnik, 2016: 223–235). NK also received an estimated 60% of its budget from the patron state (Ó Beacháin et al., 2016: 443), and the two economies were increasingly integrated (Grono, 2016). The catastrophic defeat in the autumn of 2020 came as a shock. The Armenian defence crumbled in the face of a well-equipped Azerbaijani army supported by Turkey. The entity was on the brink of complete defeat when a Russian-brokered ceasefire was signed. However, this peace came at a high price: NK had lost more than two-thirds of the territory it controlled, and Armenia had to agree to withdraw its armed forces from the entity. Instead, a Russian peacekeeping force was to guarantee the entity’s security (see, for example, Crisis Group, 2021). Patron state support had proved insufficient, and the de facto state found itself in a much weaker position, with little prospect for independence.

As noted earlier, the two cases share important similarities: In both cases, the patron state is also a kinstate, and ethno-nationalist narratives were central to legitimising both the territorial claims made by the de facto state and the patron’s support. At the time of the defeat/reintegration, both Nagorno Karabakh and Republika Srpska enjoyed territorial control and had built state-like institutions (see, for example, Caspersen, 2012; Gueudet, 2022; Kolstø and Pauković, 2014). Afterwards, we continue to find elements of de facto statehood in both cases, despite the differing outcomes. RS maintained a high degree of autonomy within Bosnia-Herzegovina, including its own armed forces (until 2005) and the right to sign parallel agreements with neighbouring states, that is, Serbia. In NK, the de facto authorities were in control of what remained of the entity and its institutions were still functioning. In both cases, the patron continued to represent its client in peace talks.

The two cases differ when it comes to the type of conflict that led to a crisis in the relationship, including the patron’s apparent intentions. Armenia tried to defend its client, but was unable to do so effectively, while Serbia accepted a peace agreement that went against the wishes of its client. Consequently, the leaders of RS could argue that they were betrayed by their patron, whereas the leaders of NK were more likely to blame lack of preparedness. However, in the aftermath of the 2020 war, the conflict between Armenia and NK came to resemble the other case study more closely. In a landmark speech in April 2022, the Armenian Prime Minister, Nikol Pashinyan, stated that the International Community was asking Armenia to ‘lower the bar’ on the territory’s status. This was interpreted by many as laying the groundwork for significant concessions (Isayev et al., 2022). In May 2023, this was finally made explicit when Pashinyan stated that Armenia was ready to recognise NK as part of Azerbaijan (Shahverdyan, 2023). Yet in the immediate aftermath of the war, the patron’s intentions were less clear. We would therefore expect a stronger reaction in Republika Srpska than in Nagorno Karabakh. However, the two clients also differ when it comes to two factors expected to affect their agency and leverage. First, there is a notable difference in the level of security threat. NK’s statehood remained contested and fragile: it maintained its claim to full independence and the threat of forceful reintegration was very real. This threat was borne out by the successful September 2023 offensive by the Azerbaijani army, which led to the fall of the entity. As a federal entity, RS has been legalised and its situation is a lot less precarious. This could increase its leverage against the patron. However, a second difference points in another direction: RS had few options when it came to alternative sources of support, whereas NK enjoyed significant support from the sizable Armenian diaspora and could court Russia as a potential new patron.

The following analysis draws on primary sources, including government documents from the National Archives of Republika Srpska, statements and documents from the official websites of the Nagorno Karabakh authorities, and interviews conducted with officials and analysts. These interviews were used as sources of information for the NK case study, rather than as point of analysis, due to the relative lack of availability of other primary sources.² We also draw on an extensive collection of secondary sources, including articles from newspapers, online news sites and the official press agencies of both entities. In both cases, the analysis covers the final stages of the war and its aftermath: 1995–1997 in Republika Srpska and 2020–2022 in Nagorno Karabakh. In the case of NK, we only analyse the immediate effects of the 2020 defeat on relations with the patron and do not include the fall of the entity, whereas in RS we also briefly examine the longer term reconfiguration of the patron–client relationship (1998–2001). In the next section, we examine the reactions in the de facto states. We then analyse the strategies employed by the de facto states, as well as the constraint they faced. Finally, we examine the transformed patron–client relationship that resulted from the crisis.

Reaction in the de facto states

In the immediate aftermath, we observe strong reactions in both de facto states. However, the target of the criticism is not just the patron, but also their own government. In Republika Srpska, the disappointment and frustration of the Dayton delegation was matched by the resentment they faced at home (Koljević, 2014: 541–544). Both opposition parties and the powerful military held the RS government responsible for the humiliation inflicted in Dayton and openly questioned their legitimacy to govern (Koljević, 2014: 548–549). Their blind obedience to the patron was seen as a sign of weakness. For example, Milorad Dodik, then leader of the parliamentary opposition, demanded that the RS leaders ‘face parliament and take a stand on the agreement’. He disputed their legitimacy to implement the agreement and called for elections (Svarm, 1995). However, the patron leader himself was not exposed to much direct criticism (Lekic, 1995). Only months later, in April 1996, did Karadžić publicly express his disappointment with Milošević. In front of the RS national assembly, he lamented, ‘you’re no longer what you used to be’ (Donia, 2008). At that point, Karadžić attacked the legitimacy of Milošević as the leader and voice of the Serbs. However, he did not question the ethno-nationalist narrative of the matica [motherland] and the objective of unifying all Serbs (Maksić, 2017). The ‘material, moral and spiritual cooperation’ with the patron (Koljević, 2014: 550) had been central to the mobilisation of the Bosnian Serbs and was not questioned.

In the case of Nagorno Karabakh, the patron leader faced more intense pressure. The military defeat and the deeply unpopular ceasefire agreement ushered in violent protests in Armenia’s capital, senior officials resigned from their posts and 40 high-ranking military officers publicly called for the resignation of Prime Minister Nikol Pashinyan (Azatutyun, 2021; Crisis Group, 2020: 12). In Nagorno Karabakh itself, the response was more subdued. The defeat led to criticism of the Armenian government, but not of the patron–client relationship, although some implicitly questioned if it was enough to ensure the protection of the remaining territory (Socor, 2021a). A ‘government of national accord’ had been formed in response to the war and the new head of the security council, Vitaly Balasanyan, openly blamed both the former NK government and the patron government for the defeat, condemning their ‘defeatist attitude’, weakness and incompetence (Socor, 2021a). However, the Karabakh president adopted a more conciliatory approach towards the patron and offered to mediate between the Yerevan protesters and Pashinyan (Socor, 2021b). Pashinyan eventually weathered the domestic storm and the NK president, and his more conciliatory approach, also emerged victorious from the post-war power struggle in Karabakh (Hauer, 2021). Nevertheless, the patron–client relationship had undergone a change and following Pashinyan’s speech in April 2022, the NK authorities took the unusual step of criticising the patron publicly. By signalling willingness to accept Azerbaijani control of the entity, a red line had been crossed (Interview B, 2022). The NK president, foreign minister and parliament all issued strongly worded statements that called on the Armenian government to ‘abandon the current catastrophic position’ of lowering the bar on NK’s status ‘under the pretext of “peace”’ (Mejlumyan, 2022; NK parliament, 2022a).

Reclaiming autonomy or seeking greater patron compliance

In both cases, we saw expressions of disappointment and sometimes even anger with the patron. However, it was the patron leader, rather than the patron–client relationship that was criticised. In addition to these few examples of public criticism of the patron, the de facto authorities employed various strategies to enhance their autonomy and seek greater compliance from the patron. However, these responses are constrained. Not only does the client risk losing vital patron support, as highlighted in the literature, distancing themselves from the patron could also undermine the ethno-nationalist narrative, which is important for both internal and external legitimacy.

Courting alternative external supporters

If the patron is no longer seen as a reliable or capable protector of the entity, one option is to seek alternative supporters: to replace the patron, gain greater leverage or increase the cost of enforced compliance. In this strategy, the clients used their control of local resources, including the state-like institutions they had created.

Following the signing of the Dayton Peace Agreement, the RS Ministry of Foreign Affairs prioritised the development of its international representations. A telegram from the minister of foreign affairs, Aleksa Buha, to the RS government (7 December 1995) points to the importance of this strategy as he requests an increased budget for the RS representations already opened in Germany, the Russian Federation, the United Kingdom and the United States (MFA Republike Srpske, 1995a). These offices provided spaces for advocacy and enabled the client to present a narrative that differed from the patron’s. In a letter addressed to the Russian Minister of Foreign Affairs, Buha thanked him for the support given to the Bosnian Serb representatives in Dayton and for supporting, the right of RS to open its own institutions, despite protests from other international actors. He also invited his counterpart to open a Russian representation in RS, which would support the growth of diplomatic relations independent from Belgrade and from Sarajevo (MFA Republike Srpske, 1995b). Buha stressed the importance of RS international representations as a means of ‘overseeing humanitarian aid, information-propaganda activities, organising and improving economic relations, and encouraging economic investments’. However, these attempts to court alternative external supporters were constrained by a lack of financial resources, to the point that Buha launched a fund-raising campaign among leading RS companies to keep the strategy afloat (MFA Republike Srpske, 1995c). The effect appears to have been limited. RS was isolated internationally and failed to attract new external supporters or increase support from existing ones. Even the appeals to Russia were futile at the time; the long-discussed opening of a Russian representation in Banja Luka was only agreed in May 2022 (Glas srpski, 2022).

Compared to Republika Srpska, Nagorno Karabakh had more opportunities when it came to courting alternative external supporters. The de facto authorities already benefitted from significant support from the sizable Armenian diaspora, and the diaspora remained central to the entity’s international strategy. It provided vital resources and was important for attempts to seek further international support (see, for example, MFA, Nagorno Karabakh 2022b and 2022c). The Karabakh MFA and its offices abroad, which were located in countries with large diaspora populations (Washington DC, Paris and Sydney), continued their campaign for international recognition, with the difference that the need for protection against military offences became a central argument for recognition (Interview B, 2022; see also O’Connor, 2020). However, ensuring support from the diaspora became more difficult following the war, as it had led some to see the entity’s defeat as inevitable (Interview B, 2022; Interview D, 2022). The scope for additional support was limited.

However, the aftermath of the 2020 war saw the emergence of a potential new patron. With the stationing of 2000 peacekeepers, Russia replaced Armenia as the entity’s security guarantor. Russia already acted as Armenia’s patron, through the Collective Security Treaty Organization, thereby rendering the patron support a form of nested hegemonic relationship (see Adler-Nissen et al., 2021). Some Stepanakert officials, including the head of the powerful security council and the speaker of parliament, publicly acknowledged this. They openly identified Russia’s peacekeepers along with Karabakh’s own armed forces as security guarantors, omitting Armenia from the equation (Socor, 2021a). A 2022 tweet from the NK army announced that all defence programmes would be ‘synchronised with the functions of the Russian peacekeeping contingent’ (Mejlumyan, 2022). In the immediate aftermath of the war, the peacekeepers also provided important humanitarian support and helped with infrastructure repairs, such as the entity’s water supply (Mnatsakanyan, 2021). This ‘social guarantor’ part of the patron role was subsequently reduced (Interview B, 2022), but the continued security guarantee was crucial to the entity’s survival. Nagorno Karabakh needed Russia’s support, and this dependence was reflected in attempts by the authorities to curry favour with the new patron, for example, by making Russian an official language in the entity. The Karabakh president stated, ‘giving the Russian language an official status would deepen Nagorno-Karabakh’s history of “cultural, military, and economic links” with Russia’ (RFE/RL, 2021). Unlike Armenia, the NK authorities were presenting the entity as part of the Russian World (Russkiy Mir; Socor, 2021a). Following Pashinyan’s April 2022 speech, a Karabakh politician even called for a referendum on joining Russia, and the entity’s president did not rule this out:

To continue living here, having de facto independence, striving for further de jure changes, it can be independence, it can be a union with Armenia, which I consider unlikely, but it also can be some relations with Russia in a direct vertical framework. (Mejlumyan, 2022)

However, the strategy of seeking a new patron was associated with costs and significant risks for Nagorno Karabakh. First, Russia’s support translated into a degree of control over the entity, including over who was allowed to enter the territory (Mejlumyan, 2022). Although the NK government still functioned, the situation was ‘somewhere between strong Russian influence and de facto control’ (Hauer, 2021). This loss of autonomy was significant since the de facto state lacked any leverage over its new patron. Second, and even more importantly, there were serious doubts about the reliability of the new patron. The Russian invasion of Ukraine and increasing tensions between Moscow and Armenia heightened fears in Karabakh that continued protection was not guaranteed (Broers, 2022; Interview B, 2022; Kucera, 2022). Local analysts argued that calls for joining Russia could be explained by international isolation (Aghavalyan, 2022) and panic (Interview A, 2022). In the end, these fears were realised as changing geopolitical priorities and lack of capacity led the new patron to look the other way when Azerbaijan launched its successful offensive in September 2023. The Russian peacekeepers did not intervene, leading to the exodus of the Armenian population. This underlines the lack of alternatives available to de facto states.

Even before September 2023, it had been clear to the de facto authorities that severing links with the old patron would leave Karabakh without a reliable external supporter. Given the weakness of the entity, and its continued economic dependence on Armenia, this would be a high-risk gamble which the authorities could not take. We discuss these constraints further in the following sections.

Attempts to increase autonomy

If replacing the patron is not an option, and alternative external supporters do not provide the de facto state with enough leverage, the client may instead try to increase its autonomy, for example, by insisting on direct representation in peace talks. This would reduce their reliance on a patron they no longer fully trust.

Following the signing of the Dayton Peace Agreement, several multilateral conferences were held to agree on the specifics of its implementation. In these talks, the patron still acted as the plenipotentiary representative of the Bosnian Serbs and the RS leadership fought hard to gain even minimal representation. Prior to the first Donors Conference for the reconstruction of Bosnia-Herzegovina in December 1995, the RS government sent a telegram to the MFA of Yugoslavia demanding their inclusion in the delegation (MFA Republike Srpske, 1995d). Although the request was granted, it did not end the client’s diplomatic isolation. Following the conference, the RS foreign minister publicly denounced, the effective exclusion of the RS delegation, arguing that the patron’s intention had been to force the entity into accepting a ‘deceitful revision’ of the peace agreement (MFA Republike Srpske, 1995e). As will be shown below, this drove the RS government to adopt a more confrontational strategy.

The Karabakh authorities also long insisted on their inclusion in the peace talks, but this demand was always fiercely rejected by Azerbaijan (see, for example, Avetisyan, 2009). Instead, close consultations took place between the leaderships of Yerevan and Stepanakert to coordinate their positions. However, Pashinyan reportedly negotiated with little, if any, input from Karabakh (Mejlumyan, 2022). This led to concerns of being ‘sold out’ by the patron and when the NK parliament called on the Armenian government to renounce its position, they also reminded Yerevan that ‘no government has the right’ to accept a status ‘unacceptable for Artsakh [NK]’ or negotiate away its right to self-determination (NK parliament, 2022a).

Seeking patron compliance: through obstruction or through its domestic politics

If de facto states fail to increase their autonomy, and direct appeals to the patron leader fall on deaf ears, they may resort to other strategies for ensuring greater patron compliance with their preferences. One such strategy is direct obstruction. Another approach is to rally the patron’s political opposition to their cause and try to force a change in position.

From the winter of 1996, the relationship between RS and Serbia entered a confrontational phase. The RS leadership continued to be excluded from multilateral conferences dedicated to key issues in the peace process, and Milošević thereby retained top-down control. However, the patron did not control the organs of local government, which were firmly in the hands of Karadžić’s Serb Democratic Party (Srpska Demokratska Stranka, SDS). The peace agreement could not be implemented without their collaboration, and the potential non-compliance by local RS authorities provided the client with leverage. The so-called evacuation of Sarajevo provides a good example of this obstructionist strategy. The London Peace Implementation Conference (December 1995) had prioritised the transfer of Serb-controlled parts of the city to the Bosniak-Croat entity (Peace Implementation Council, 1995). The practicalities were agreed at the Rome conference where the only RS representative was Prime Minister Rajko Kasagić, a Milošević-loyalist (Peace Implementation Council, 1995). However, thanks to Karadžić’s control over the entity’s media, he was able to circulate alarming information about the fate of the Serb inhabitants in the first area to be transferred (Glas srpski, 1996a; Novinska Agencija Republike Srpske, 1996). This disinformation campaign created favourable grounds for the RS authorities to organise the ‘evacuation’ of the remaining Serbs in a climate of panic, with the help of SDS local offices, police stations under the authority of the Ministry of Internal Affairs, and most of the SDS-controlled local council (Glas srpski, 1996b). By sabotaging the transfer, the RS authorities were able to reclaim agency in a bid to preserve their remaining legitimacy.

Although the RS leaders were able to increase the cost of patron control, their obstructionist strategy quickly backfired. The last straw for the patron was Karadžić’s open challenge to Milošević in front of the RS National Assembly: ‘a child cannot be submissive to his mother without dignity and contrary to his own interest’ (Donia, 2008, 80). This sparked concerns that the RS government would oppose the implementation of the peace agreement in even more radical ways. In response, international officials called for Karadžić’s dismissal and for Milošević to deliver it. Following a meeting between the RS vice-president and Milošević in Belgrade, the replacement of Karadžić was announced (Donia, 2008, 81). The patron had demonstrated its influence over the client’s internal politics, and its agency was severely constrained as a result.

In response, the RS leadership appears to have attempted to influence the patron leader by aligning themselves with other forces in Serbian politics. An archive document from the RS Ministry of Foreign Affairs compiled the stances of the Serbian opposition, noting an emerging consensus that the regime had failed the Bosnian and Croatian Serbs, and given up on the objective of unifying all Serbs in a single state (MFA Republike Srpske, n.d.). However, given the autocratic nature of Milošević’s regime and the absence of a unified opposition, this strategy for altering the patron’s policy never paid off.

In the case of Nagorno Karabakh, we find no examples of direct obstruction, but we do see attempts to change the position of the patron through its domestic politics. The contested entity is central to Armenia’s ethno-nationalist narrative (Malkasian, 1996) and this provided the facto authorities with leverage. They had previously put this to effective use, most notably in 1998 when the NK leadership managed to oust the patron leader following his willingness to consider a settlement proposal they flatly refused (Caspersen, 2008). Following the 2020 war, they NK leadership started to publicly air their differences with the patron, if red lines were crossed (Interview B, 2022). Ahead of peace talks in Sochi in October 2022, and following renewed signals that the patron would be willing to compromise on Karabakh’s status, the NK authorities again issued public statements against the patron’s position (Mejlumyan et al., 2022). They backed this up with popular mobilisation by organising a mass rally in the entity, attended by tens of thousands (Mejlumyan et al., 2022). Their statements reminded the patron of its obligation to the client: it had no right to refuse its ‘mission’ to protect the security of the people of Karabakh, or to accept a document questioning the entity’s sovereignty (NK parliament, 2022b). They also urged the patron to recall the centrality of Karabakh to Armenia’s ethno-national narrative, arguing that the ‘struggle . . . expresses the collective will of Armenians worldwide’ (NK parliament, 2022b) The warning by foreign minister, David Babayan, that ‘after the destruction of Artsakh [NK] there would be no Republic of Armenia’ (Mejlumyan, 2022) further stressed this point.

This strategy had some effect on a patron leader who knew he was constrained in his ability to compromise on NK’s status. Following the strong statements in April 2022, the NK president reportedly met with Pashinyan and received assurances (Interview A, 2022), which resulted in improved relations (Interview B, 2022). Pashinyan also pushed back publicly against accusations that he was sacrificing the Karabakh Armenians and stressed the need to ensure their ‘rights, freedoms, and . . . status’ (Mejlumyan, 2022). However, the focus on rights, rather than territory, marked an important change in the patron’s rhetoric. As a result, relations were characterised by a profound mistrust (Interview A, 2022; Interview B, 2022; Interview C, 2022), but the de facto authorities were constrained in their ability to respond. The constraints were two-fold. First, the weakness of the entity and its continued economic dependence on the patron meant that the NK leaders had to tread carefully. Pashinyan made sure that the Karabakh members of parliament were mindful of this dependence when, in response to their criticism, he reminded them that Armenia pays their salaries (Interview C, 2022). Second, not just the patron but also the client was constrained by the ethno-nationalist narrative. The de facto authorities relied on this narrative for their internal legitimacy (Caspersen, 2015), and vital diaspora support was mobilised through a narrative of a pan-Armenian struggle that brought together Armenians in Karabakh, Armenia and across the world (MFA, Nagorno Karabakh, 2022a). Deep divisions with the patron would weaken that narrative and could reinforce the perception that Karabakh was a lost cause. As noted earlier, this perception had already started undermining diaspora support (Interview B, 2022; Interview D, 2022). As a result, the preferred strategy was to try to influence Yerevan behind closed doors (Interview B, 2022). Keeping disagreements private made it possible to preserve an image of ethno-national unity. Moreover, although Pashinyan was deeply unpopular in NK (Interview A, 2022), the de facto leaders were unable to mobilise popular opinion in the patron state or effectively align themselves with the opposition. Immediately after the 2020 war, the head of the NK security council supported attempts by the Armenian opposition to force Pashinyan’s resignation (Socor, 2021a). However, this strategy was abandoned once the patron leader regained control of the situation. Popular support for the protects was limited and declining, and the NK authorities did not publicly back them (Caucasian Knot, 2022).

Renegotiated relationship: the return of the patron

As a result of the constraints analysed earlier, both cases saw the return of the patron. However, the strategies used by the client to restore autonomy had an impact and the two patron–client relationships were renegotiated.

In the case of Nagorno Karabakh, the immediate post-war period pointed to the disengagement of Armenia from Karabakh, as Yerevan’s involvement became limited to financial support (Grigoryan, 2021). This more passive patron role was prompted by the terms of the ceasefire agreement and the chaotic situation in Armenia, but also reflected the client’s disappointment with the patron as well as the emergence of a potential new patron. However, tensions between the patron and the client subsided once Pashinyan consolidated his position. The de facto state reengaged with its patron, but on somewhat different terms.

The link with the patron was weaker, as Armenia no longer provided a security guarantee, but it was not without significance. First, the patron still represented the client in peace talks. Second, economic support from the patron continued and actually increased due to the greater need following the war (Interview A, 2022; Interview B, 2022; Interview C, 2022). Other ties between government bodies in Karabakh and Armenia were either restored or new ones were created (Grigoryan, 2021). Third, the continuation of strong relations with Armenia was part of a narrative that described Karabakh’s claim to independence and the protection of the entity as an all-Armenian struggle. Speeches and statements from the Karabakh authorities repeatedly referred to the importance of ‘pan-Armenian unity’ and the ‘Armenia-Artsakh-Diaspora trinity’ (MFA, Nagorno Karabakh, 2022a). Even if the patron no longer supported its independence, the client had little choice but to stay close: it was in a very vulnerable position, its new patron was unreliable and support from the old patron was intimately tied up with continued diaspora support. However, public statements pointed to a renegotiated relationship with the (old) patron, which formed part of a structure of external supporters: it was important but not necessarily privileged. For example, the NK foreign minister stated that the entity maintained its ‘traditional priorities’: aiming for international recognition, seeking de facto relations with other states, primarily with ‘fraternal Russia’, and reinforcing inter-Armenian relations, i.e. with Yerevan and the diaspora (Socor, 2021b). Moreover, the patron’s influence in the de facto state was not as strong as it used to be (Interview A, 2022). This can be explained by the lack of military dependence on the patron, but also by the unpopularity of the patron leader and the mistrust of his intentions (Interview A, 2022). Despite the client’s weakness, the patron–client relationship had become less hierarchical.

Notwithstanding the absence of a significant security threat, the client’s agency was also significantly constrained in the case of Republika Srpska. Continued economic dependence made the RS leaders unable to increase their autonomy from the patron. The entity still faced international sanctions and lacked access to international reconstruction aid. Consequently, the patron remained the main source of economic opportunities, especially since the Dayton Agreement allowed the client and patron to sign a treaty of ‘Special Parallel Relations’ (Politika, 1997). Moreover, the ethno-national narrative remained important to the legitimacy of the RS leadership and severing ties would mean losing a significant part of the entity’s raison d’être. Finally, as demonstrated earlier, the RS leaders were constrained by the patron’s influence over their domestic political competition. However, this influence was beginning to wane. In 1997, Milošević brokered an agreement to settle an RS leadership dispute, but this outcome had been dictated to him by the international High Representative, Carl Bildt, and went against the patron leader’s preferred outcome (Glas srpski, 1997). Following this crisis, the newly elected RS President, Biljana Plavšić, began to comply with the directives of the international community. As Bildt (1998) wrote in his memoirs, the ‘Milošević channel’ had run dry, and the progressive decline of the patron’s influence in RS resulted in a rebalanced relationship.

Two factors appear to have been crucial: (1) the weakening of Milošević’s position in Serbia and of Serbia’s position in the region and (2) the client’s greater access to international support. When Milošević’s regime was embedded in accusations of electoral fraud and became an international outcast due to the war in Kosovo, the RS leadership distanced itself from the patron leader. Although vocally supporting the Serbian claim to Kosovo, the RS leadership disavowed Milošević as leader of Yugoslavia and therefore patron of Republika Srpska (Glas srpski, 2000a). The new RS government had become too reliant on international aid to openly support Milošević, and with this alternative external support came new constraints. The fall of Milošević in 2000 created new opportunities for renegotiating the patron–client relationship. As soon as the new Serbian President, Vojislav Koštunica, was elected, the RS National Assembly held a session with the headline message ‘Bez barijera na Drinu’ (Glas srspki, 2000b), which translates as ‘No barrier on the Drina river’ (the border between RS and Serbia). During a visit to Belgrade, RS Prime Minister, Milorad Dodik, disassociated RS from Milošević’s legacy including the very hierarchical approach to patron–client relations, arguing for a new phase of bilateral relations (Glas srpski, 2000c). The signing of a second Agreement on Special Parallel Relations marked the beginning of a more horizontal partnership (Glas srpski, 2000c). Serbia was no longer a regional hegemon and the client’s economic dependence on its patron was reduced: the relationship had become less asymmetrical, and this reduced the constraints on the client’s agency.

Conclusion

The dependence of de facto states on external patrons is well-established in the existing literature, but we lack detailed knowledge of the dynamics of this patron–client relationship, including opportunities for client agency and catalysts for change. This article examined these issues through an in-depth comparative study of two patron–client relationships that had reached a crisis point: the relationship between Serbia and Republika Srpska after the signing of the Dayton Agreement and between Armenia and Nagorno Karabakh after the 2020 war. The resulting out-of-equilibrium behaviour presented an opportunity for analysing the underlying mechanisms of the relationship. The article demonstrated the usefulness of expanding the patron–client framework beyond its usual state-centric focus. The comparative analysis showed how the client’s leverage is limited by the context of non-recognition, but it also provided additional evidence for the dynamic nature of patron–client relations.

We developed a typology of patron–client crises – abandonment, significant reduction in support or enforced compromise on the issue of independence – and of client strategies for reclaiming autonomy or ensuring greater patron compliance with their preferences. We also identified factors likely to constrain such strategies. In our in-depth case studies, we found that the type of crisis and patron intentions were less important for client responses than the significant constraints faced. In both cases, the de facto leaderships were constrained by the absence of reliable alternative supporters, by remaining structural links and continued dependence on the patron’s support, and by the existence of powerful ethno-nationalist narratives. Lack of sovereignty matters: the patron–client relationship is characterised by multiple reinforcing hierarchies, and the impact of non-recognition is especially noticeable in the case of severe security threats. Unlike patron–client relationships involving sovereign states, where the existence of alternative patrons provides the clients with leverage (Cooley and Nexon, 2020), a patronage monopoly still effectively exists for de facto states, due to their international isolation.

Our article shows how this constrains de facto states, yet our findings also demonstrate that even the weakest clients have agency. They do not have access to an ‘à la carte hierarchy’ (Sharman, 2017), but they retain room for manoeuvre. The analysis contributes to the existing patron–client literature by providing a detailed, bottom-up examination of the strategies used by the weaker party. As suggested in the literature, alternative supporters were courted. This strategy is not without risks, as it exposes de facto states to geopolitical volatility. However, reduced dependence on the patron, due to complimentary sources of external support was an important catalyst for changing dynamics – economic support in the case of Republika Srpska, security in Nagorno Karabakh. Reduced patron capacity was another catalyst for change which undermined the patron’s ability to enforce compliance in the case of Republika Srpska. In the case of Nagorno Karabakh, patron leverage was also reduced because of mistrust, thereby demonstrating that not just relative capacity matters. Affectivity was undermined, resulting in increased cost of compliance. Moreover, other strategies were also pursued by the client: Attempts were made to alter the patron’s position by affecting the dynamics of domestic contestation; and there were even examples of direct obstruction. The client controls local resources and can use this for leverage. As a result, while the de facto states were unable to sever links with their patron, the crisis led to a renegotiated patron–client relationship.

No matter how hierarchical, patron–client relationships are also reciprocal. If the patron fails to live up to its side of the bargain, if it is ‘no longer what it used to be’, it must either renegotiate the contract or accept reduced leverage over its client.

Footnotes

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Research for this article was generously funded by a grant from the Norwegian Research Council: Dynamics of de facto state patron-client relations (301277).

ORCID iD

Nina Caspersen

Notes

Author biographies

Nina Caspersen is Professor of Politics at the University of York, the United Kingdom. She is a leading expert on de facto states, conflict dynamics and peace agreements. Her work has been funded by the British Academy, the European Social and Research Council, the Medical Research Council and the Norwegian Research Council. She is the author of three monographs and numerous journal articles.

Sophie Gueudet is a research fellow working on the project RE-ENGAGE (Horizon Europe 101132314) at the Scuola Superiore Sant’Anna in Pisa, Italy. She is also an associate researcher at LSE IDEAS as part of its work for the PeaceRep-Ukraine programme. A historian of war and conflict, she specialises in separatism and secession, contested states and unresolved territorial conflicts, and civil-military relations in intra-state wars in Southeastern and Eastern Europe and the Caucasus.

References

Adler-Nissen

Zarakol

(2021) Struggles for recognition: The liberal international order and the merger of its discontents. International Organization 75(2): 611–634.

Adler-Nissen

De Carvalho

Leira

(2021) Goods substitution at high latitude: undermining hegemony from below in the North Atlantic. In: Andersen

Cooley

Nexon

(eds) Undermining American Hegemony: Goods Substitution in World Politics. Cambridge, MA: Cambridge University Press, p. 120.

Aghavalyan

(2022) Twitter, 14 April. Available at: https://twitter.com/aghavalyan/status/1514680329180790793?s=20&t=jfgSFelqJ8o7vOPOBrqhKw (accessed 7 December 2022).

Avetisyan

(2009) Nagorno-Karabakh Must No Longer Be Barred from the Negotiating Table. REF/RL, 14 July. Available at: https://www.rferl.org/a/NagornoKarabakh_Must_No_Longer_Be_Barred_From_The_Negotiating_Table_/1776580.html

Azatutyun (2021) Armenian military demands government’s resignation, 25 February. Available at: https://www.azatutyun.am/a/31121104.html

Bercovitch

(1991) Superpowers and client states: Analysing relations and patterns of influence. In: Efrat

Bercovitch

(eds) Superpowers and Client States in the Middle East. London: Routledge, pp. 13–19.

Berg

Vits

(2018) Quest for survival and recognition: Insights into the foreign policy endeavours of the post-soviet de facto states. Ethnopolitics 17(2): 390–407.

Biermann

(2024) Conceptualising patron-client relations in secessionist conflict. A research agenda. Territory, Politics, Governance 1–19: 2318467.

Bildt

(1998) Peace Journey: The Struggle for Peace in Bosnia. London: Weidenfeld & Nicolson.

10.

Blakkisrud

Kolstø

(2012) Dynamics of de facto statehood: the South Caucasian de facto states between secession and sovereignty. Southeast European and Black Sea Studies 12(2): 281–298.

11.

Blakkisrud

Kemoklidze

Gelashvili

, et al (2021) Navigating de facto statehood: trade, trust, and agency in Abkhazia’s external economic relations. Eurasian Geography and Economics 62: 3347–3371.

12.

Broers

(2022) With Russia Distracted, Azerbaijan Escalates in Karabakh. Chatham House, 30 March. Available at: https://www.chathamhouse.org/2022/03/russia-distracted-azerbaijan-escalates-karabakh

13.

Bullough

(2008) Grateful Abkhazia thanks Russia on national day. Reuters, 30 September. Available at: https://www.reuters.com/article/idUSLU133758/

14.

Carney

(1989) International patron-client relationships: A conceptual framework. Studies in Comparative International Development 24: 42–55.

15.

Caspersen

(2008) Between Puppets and Independent Actors: Kin-State Involvement in the Conflicts in Bosnia, Croatia and Nagorno Karabakh. Ethnopolitics 7(4): 357–372.

16.

Caspersen

(2010) Contested Nationalism: Serb Elite Rivalry in Croatia and Bosnia in the 1990S. Oxford: Berghahn Books.

17.

Caspersen

(2012) Unrecognised States: The Struggle for Sovereignty in the International System. Cambridge: Polity Books.

18.

Caspersen

(2015) Degrees of legitimacy: Ensuring internal and external support in the absence of recognition. Geoforum 11(1): 184–192.

19.

Caspersen

(2023) Political parties in De Facto States: Links with external patrons. In: Ishiyama

Sindra

(eds) The Effects of Rebel Parties on Governance, Democracy and Stability after Civil Wars. London: Routledge, pp. 27–36.

20.

Caucasian Knot (2022) Armenian political analysts note absence of public support of opposition, 13 June. Available at: https://www.eng.kavkaz-uzel.eu/articles/60416/

21.

Chollet

(2007) The Road to the Dayton Accords: A Study of American Statecraft. New York: Springer.

22.

Comai

(2018) What Is the Effect of Non-recognition? the External Relations of De Facto States in the Post-soviet Space. PhD Dissertation, Dublin City University, Dublin. Available at: http://doras.dcu.ie/22159/

23.

Cooley

(2019) Ordering Eurasia: The rise and decline of liberal internationalism in the post-communist space. Security Studies 28(5): 588–613.

24.

Cooley

Nexon

(2020) (No) Exit from liberalism? New Perspectives 28(3): 280–291.

25.

Crisis Group (2020) Improving Prospects for Peace after the Nagorno-Karabakh War. Briefing #81, Europe and Central Asia.

26.

Crisis Group (2021) Post-war Prospects for Nagorno-Karabakh. Report #264, Europe and Central Asia.

27.

De Waal

(2018) Uncertain ground: Engaging with Europe’s de facto states and breakaway territories. Washington, DC: Carnegie Endowment for International Peace.

28.

Dembińska

Mérand

(2019) The role of international brokers in frozen conflicts: The case of Transnistria. Asia Europe Journal 17: 15–30.

29.

Donia

(2008) Izvodi iz govora u Skupštini Republike Srpske,1991 – 1996, Tuzla, University Press. Fondacija Istina, pravda, pomirenje.

30.

Donia

(2014) Radovan. Karadžić: Architect of the Bosnian Genocide. Cambridge, MA: Cambridge University Press.

31.

Glas srpski (1996a) Pocela evakuacija sarajevskih Srba, 21 February

32.

Glas srpski (1996b) Nesredna Ikonija u hotelu Fortuna, 20 February

33.

Glas srpski (1997) Parlamentarni izbori 15. novembra, predsednički 7. decembra, 25 September

34.

Glas srpski (2000a) Wolfgang Petrić: Milošević je negativan faktor, 9–10 September.

35.

Glas srpski (2000b) Susreti u Patrijaršiji i Vladi Srbije, 6 November

36.

Glas srpski (2000c) Srbija za Dodika, 7 November

37.

Glas srpski (2022) U Banja Luci, 26. maja, biće otvorena filijala Ruske ambasade, 26 May.

38.

Graham

Horne

(2012) Unrecognized States: A Theory of Self-determination and Foreign Influence. Los Angeles, CA: University of California Press.

39.

Grigoryan

(2021) One year after the signing of the ceasefire statement: Serious security challenges persist. EVN Report, 9 November. Available at: https://www.evnreport.com/spotlight-karabakh/one-year-after-the-signing-of-the-ceasefire-statement-serious-security-challenges-persist

40.

Grono

(2016) Isolation of Post-Soviet Conflict Regions Narrows the Road to Peace. Brussels: International Crisis Group.

41.

Gueudet

(2018) Guerre d’agression ou guerre par procuration? L’armement des forces bosno-serbes par le régime de Milošević. Stratégique 118(1): 229–240.

42.

Gueudet

(2022) Serbes contre Serbes: les relations politico-militaires en Republika Srpska durant la guerre de Bosnie-Herzégovine (1991-1996) Revue Historique des Armées 304(1): 26–40.

43.

Hauer

(2021) Bitter military losses lead to power struggle in Nagorno-Karabakh. RFE/RL, 10 April. Available at: https://www.rferl.org/a/karabakh-power-struggle-haratiunian-balasanian-russia-armenia-azerbaijan/31196526.html

44.

Hicken

(2011) Clientelism. Annual Review of Political Science 14: 289–310.

45.

Hobson

Sharman

(2005) The Enduring Place of Hierarchy in World Politics. European Journal of International Relations 11(1): 63–98.

46.

Holden

(2004) Armies without Nations: Public Violence and State Formation in Central America, 1821-1960. Oxford: Oxford University Press.

47.

Interview

(2022) Tigran Grigoryan, analyst. On Zoom from Yerevan, 18 July.

48.

Interview

(2022) Anonymous. On Zoom from Stepanakert, 17 August.

49.

Interview

(2022) Tevan Poghosyan, analyst. On Zoom from Yerevan, 17 August.

50.

Interview

(2022) Richard Santourian, President of the South section of the Fond Arménien de France, Aix, 20 October.

51.

Isayev

Kucera

Mejlumyan

(2022) EU emerges as major player in armenia-azerbaijan negotiations. Eurasianet, 25 May. Available at: https://eurasianet.org/eu-emerges-as-major-player-in-armenia-azerbaijan-negotiations

52.

Jacobsen

(2020) Greenland’s Arctic advantage: Articulations, acts and appearances and sovereignty games. Cooperation and Conflict 55(2): 170–192.

53.

Kanol

Köprülü

(2017) Quality of democracy in unrecognized states: Lessons from Northern Cyprus. Southeast European and Black Sea Studies 17(3): 389–402.

54.

Karlén

(2017) The legacy of foreign patrons. Journal of Peace Research 54(4): 499–512.

55.

Kaufman

(1974) The Patron-Client Concept and Macro-Politics: Prospects and Problems. Comparative Studies in Society and History 16(3): 284–308.

56.

Ker-Lindsay

(2013) Engagement without recognition: The limits of diplomatic interaction with contested states. International Affairs 91(2): 267–285.

57.

Koljević

(2014) Creating the Republika Srpska. Provo, UT: Unwritten History.

58.

Kolstø

(2020) Biting the hand that feeds them? Abkhazia–Russia client–patron relations. Post-soviet Affairs 36(2): 140–158.

59.

Kolstø

Blakkisrud

(2008) Living with Non-recognition: State- and Nation-building in South Caucasian Quasi-states. Europe-Asia Studies 60(3): 483–509.

60.

Kolstø

Pauković

(2014) The short and brutish life of republika Srpska Krajina: Failure of a De Facto State. Ethnopolitics 13(4): 309–327.

61.

Kosienkowski

(2019) The patron-client relationship between Russia and Transnistria. In: Hoch

Kopecek

(eds) De Facto Sates in Eurasia. Abingdon: Routledge, pp. 56–78.

62.

Kucera

(2022) Armenia signals willingness to cede control over Karabakh. Eurasianet, 1 April. Available at: https://eurasianet.org/armenia-signals-willingness-to-cede-control-over-karabakh

63.

Ladwig

(2017) The Forgotten Front: Patron-client Relationships in Counter Insurgency. Cambridge, MA: Cambridge University Press.

64.

Lake

(1996) Anarchy, Hierarchy, and the Variety of International Relations. International Organization 50(1): 1–33.

65.

Lake

(2017) Hierarchy and international relations: Theory and evidence. In: Thompson

(ed.) Oxford Research Encyclopedia of Politics. New York: Oxford University Press.

66.

Lekic

(1995) Night Moves, Vreme News Digest Agency No 220, 18 December.

67.

Maksić

(2017) Ethnic Mobilization, Violence, and the Politics of Affect. New York: Palgrave Macmillan.

68.

Malkasian

(1996) Gha-ra-bagh: Emergence of the National Democratic Movement in Armenia. Detroit, MI: Wayne University Press.

69.

Mejlumyan

(2022) Officials in Karabakh break with Armenia over negotiations. Eurasianet, 15 April. Available at: https://eurasianet.org/officials-in-karabakh-break-with-armenia-over-negotiations

70.

Mejlumyan

Isayev

Kucera

(2022) Ahead of key summit, Armenians of Karabakh rally against Azerbaijani control. Eurasianet, 31 October. Available at: https://eurasianet.org/ahead-of-key-summit-armenians-of-karabakh-rally-against-azerbaijani-control

71.

MFA Nagorno Karabakh (2022a) Annual Press Conference of Foreign Minister of the Republic of Artsakh, 19 January. Available at: http://www.nkr.am/en/news/2022-01-19/press_conference (accessed 6 April 2022).

72.

MFA Nagorno Karabakh (2022b) The recognition of independence of the Republic of Artsakh. Available at: http://www.nkr.am/international-recognition-of-karabakh (accessed 6 April 2022).

73.

MFA Nagorno Karabakh (2022c) International Cooperation. Available at: http://www.nkr.am/en/international-cooperation (accessed 6 April 2022).

74.

MFA Republike Srpske (1995a) Facsimile from Aleksa Buha, Minister of Foreign Affairs of Republika Srpska to Government of Republika Srpska in Pale, 02-99/95, 7 December. Narodne Arhive Republike Srpske, Box 11.

75.

MFA Republike Srpske (1995b) Telegram od 5 Dr. Radomir Lukić, pomoćnik ministra inostranih poslova Republike Srpske OPUS Pale, 05.12., 01-02/9, box 11.

76.

MFA Republike Srpske (1995c) Telegram ministra Aleksa Buha Agrobanci, Dokument 04-93-18/95, 23. avgust, kutija 11.

77.

MFA Republike Srpske (1995d) Facsimile addressed to Milan Milutinovic Minister of Foreign Affairs of the Federal Republic of Yugoslavia, 01-100/95, 16 December.

78.

MFA Republike Srpske (1995e) Declaration concerning the Bosnia reconstruction conference, no date but estimated December 1995, Box 11.

79.

MFA Republike Srpske (n.d.) Untitled document, n°01-96/95, Box 8.

80.

Mnatsakanyan

(2021) Heatwave, drought and war leave Nagorno-Karabakh short of water. Eurasianet, 8 September. Available at: https://eurasianet.org/heatwave-drought-and-war-leave-nagorno-karabakh-short-of-water

81.

NK parliament (2022a) Statement of the National Assembly of the Republic of Artsakh, 14 April. Available at: http://www.nankr.am/en/4682 (accessed 7 December 2022).

82.

NK parliament (2022b) Statement of the National Assembly of the Republic of Artsakh, 30 October. Available at: http://www.nankr.am/en/4955 (accessed 7 December 2022).

83.

Novinska Agencija Republike Srpske (1996) Daily report. 5-6-10 March.

84.

Ó Beacháin

Comai

Tsurtsumia-Zurabashvili

(2016) The secret lives of unrecognised states: Internal dynamics, external relations, and counter-recognition strategies. Small Wars and Insurgencies 27(3): 440–466.

85.

O’Connor

(2020) Exclusive: Official at center of Armenia, Azerbaijan conflict calls it fight to death. Newsweek, 2 October. Available at: https://www.newsweek.com/exclusive-existential-fight-center-armenia-azerbaijan-conflict-1536087

86.

Peace Implementation Council (1995) Conclusions of the London Conference, 8–9 December.

87.

Politika (1997) ‘SPS pozdravlja sporazum između SRJ i RS’, 1 March.

88.

Ramet

(2006) The Three Yugoslavias: State-building and Legitimation, 1918-2005. Bloomington, IN: Indiana University Press.

89.

RFE/RL (2021) Russian language to get official status in Nagorno-Karabakh, 25 March. Available at: https://www.rferl.org/a/russian-language-official-status-nagorno-karabakh/31169752.html

90.

Rodman

(1997) Patron-client dynamics: Mapping the American-Israeli relationship. Israel Affairs 4(2): 26–46.

91.

Salehyan

(2010) The delegation of war to rebel organizations. Journal of Conflict Resolution 54(3): 493–515.

92.

Shahverdyan

(2023) PM says Armenia ready to recognize Karabakh as part of Azerbaijan. Eurasianet, 23 May. Available at: https://eurasianet.org/pm-says-armenia-ready-to-recognize-karabakh-as-part-of-azerbaijan

93.

Sharman

(2017) Sovereignty at the extremes: Micro-states in world politics. Political Studies 65(3): 559–575.

94.

Shoemaker

Spanier

(1984) Patron-client State Relationships: Multilateral Crises in the Nuclear Age. Santa Barbara, CA: Greenwood.

95.

Smolnik

(2016) Secessionist Rule: Protracted Conflict and Configurations of Non-state Authority. Frankfurt: Campus.

96.

Socor

(2021a) Vitaly Balasanian, Karabakh’s Strongman-in-Waiting. Eurasia Daily Monitor, 8 January. Available at: https://jamestown.org/program/vitaly-balasanian-karabakhs-strongman-in-waiting/

97.

Socor

(2021b) Russia’s karabakh protectorate taking clearer shape (Part Two). Eurasia Daily Monitor, 22 March. Available at: https://jamestown.org/program/russias-karabakh-protectorate-taking-clearer-shape-part-two/

98.

Svarm

(1995) ‘Blackmailing Sarajevo’. Vreme NDA No 218. Available at: https://serbiandigest.libraries.rutgers.edu/218/blackmailing-sarajevo (accessed 3 December 1995).

99.

Veenendaal

(2017) Analyzing the foreign policy of microstates: The relevance of the international Patron-Client model. Foreig Polic Anal 13(3): 561–577.

100.

Vukušić

(2022) Serbian Paramilitaries and the Breakup of Yugoslavia: State Connections and Patterns of Violence. Abingdon: Taylor & Francis.

101.

Waterbury

(2021) Divided nationhood and multiple membership: A framework for assessing kin-state policies and their impact. Ethnopolitics 20(1): 39–52.

102.

Westad

(2005) The Global Cold War: Third World Interventions and the Making of Our Times. Cambridge: Cambridge University Press.

103.

Wong

(2020) Managing small allies amidst patron–adversary rapprochement: A tale of two Koreas. Asian Security 16(1): 107–126.