Arms Control, Conflict and Peace Settlements: The Caucasus

CIAO DATE: 02/02

Arms Control, Conflict and Peace Settlements: The Caucasus

S. Neil MacFarlane ₁

1999 - 2000

The Geneva Centre for Security Policy (GCSP)

Introduction

Regional arms control and confidence-building measures - notably those adopted in the framework of the Conference on Security and Co-operation in Europe (CSCE) and later the Organisation for Security and Co-operation in Europe (OSCE), as well as the Conventional Forces in Europe (CFE) Treaty (1990) - played an important role in the stabilisation of Central and Eastern Europe during and after the Cold War. At the sub-regional level, the role of arms control in the regulation of civil conflict in the former Yugoslavia (e.g. the 1996 agreement on sub-regional arms control associated with the Dayton Accord) and the comparatively successful record of these arrangements have led analysts and policy-makers to consider whether such approaches might not be useful in preventing and resolving conflicts in other parts of Europe and farther afield.

The southern Caucasus ₂ is one of the more active zones of conflict in the OSCE space, and one in which existing conflicts have been especially resistant to durable resolution. Policy-makers have repeatedly expressed interest in the application of southeastern European models of conflict resolution to the conflicts in the Caucasian sub-region. ₃ This begs the question whether regional arms control arrangements on the Dayton model might have potential as part of a strategy of conflict resolution in the Caucasus.

This paper assesses whether sub-regional arms control and confidence-building arrangements can serve as useful elements of strategies of conflict prevention and resolution within and between the three southern Caucasian countries - Armenia, Azerbaijan, and Georgia. The paper begins with a brief discussion of the applicability of arms control in post-Cold War conflict, with particular reference to the southern Caucasus. It proceeds to a discussion of the roots and course of conflict in the southern Caucasus. It then discusses the regional and broader international significance of the conflicts and summarises the processes of mediation and the role that arms control initiatives have played therein. It continues with a consideration of the applicability of models of arms control in other regions for conflict resolution and prevention in the Southern Caucasus, with particular reference to arms control initiatives in Bosnia-Herzegovina. It concludes with a discussion of the potential utility of arms control in efforts to produce regional stability in the southern Caucasus.

The Role of Arms Control

It is perhaps useful to begin by clarifying the ways in which the term arms control will be used here. In its most general sense, arms control refers to quantitative or qualitative limitations on various categories of weaponry. These limitations may have many forms: the prohibition of (e.g. the chemical and biological weapons conventions) or constraint upon (e.g. ABM) specific forms of weapons development, limits on the deployment of various categories of weapons system (CFE) weapons export control (e.g. the Wassenaar process); control on the proliferation of weapons or weapons-related technologies (e.g. the MTCR), and so on. Arms control may or may not require disarmament measures (viz. the destruction of weapons associated with the CFE and START processes), but equally they may be directed at preventing acquisition (NPT) or channelling weapons development (e.g. the SALT process). Arms control measures are frequently associated with confidence and security building measures (CSBMs), notably those involving inspection and verification. But, equally, measures to insure transparency may have little if anything to do with arms control.

The traditional complexion of arms control initiatives was strongly affected by bipolarity and focused largely on the superpower relationship. Notably, it was not focused on conflict termination, but was instead an important element of managing relations between the USA and USSR in such a way as to prevent war, and to stabilise their strategic relationship and reduce the costs of the arms race, while retaining their mutual dominance over the global strategic system as a whole (e.g. the non-proliferation regime). In contrast, regional or state-specific arms control initiatives were few and far between, although not altogether absent (viz. the progress in Latin America on the establishment of a nuclear weapons free zone).

The end of the Cold War was accompanied by a rise in the significance attributed to intra-state conflict in international relations (Holsti, 1999), to ethnicity as a dimension of conflict (Survival, 1993), ₄ and to groups and individuals, rather than states, as referents of the discourse on security. (Buzan, 1991) The analysis of security has become less state-centric and more "human." This period has also been characterized by further reduction in the incidence of inter-state war, decline in the military expenditure and the size of forces of Western and former communist states in Europe. ₅ Finally, the period has witnessed an uneven growth in the activity of universal and regional multilateral organisations in the management of security and in efforts to resolve conflicts. This has been accompanied by dramatic expansion in the number of tasks undertaken by external actors in building peace, and by an equally dramatic expansion in the numbers and types of outside organisations taking on roles in conflict management and resolution and the humanitarian, transitional and developmental functions related to responses to conflict.

What role does arms control have in this new mix? In the first place, numerous important residues from traditional arms control linger in the international inter-state negotiating agenda (continuing reduction in strategic arms through the START process, implementation of conventions on chemical and biological weapons, review and implementation of CFE, NPT review, etc.). The arms control agenda has expanded to embrace new issues (e.g. the negotiation, ratification, and implementation of the land mines treaty, or the various efforts to control the proliferation of delivery technologies in the MTCR, and of conventional arms in the Wassenaar process). However, much of the focus has shifted to the utility of arms control as part of the effort to build peace in internal and (to a lesser extent) sub-regional conflict.

Internally, effective and agreed control over weapons can play a role in conflict prevention by limiting the arms racing and conflict-inducing consequences of emergent security dilemmas between groups within states. Exclusion zones and weapons free zones along lines of contact can assist in the stabilisation of cease-fires by reducing the probability of inadvertent resumption of hostilities. In the context of the political settlement of conflict, arms control can contribute to restoring confidence between groups that have been in conflict. Weapons reduction reduces the capacity for one group can mount a credible threat to the security of other groups within a society emerging from conflict. Intrusive monitoring and inspection regimes likewise can build confidence by reducing the threat of surprise attack. Disarmament, demobilisation, and reintegration of combatants are frequently see as essentially aspects of reconstructing peaceful social relations.(Berdal, 1996)

More broadly, it is plausible that the process of negotiating arms limitation arrangements may be useful in restoring communications between parties in conflict (and particularly between their militaries) while the continuous interaction characteristic of such negotiations may foster learning and empathy.

All of this said, Keith Krause is right in stressing the dangers of too direct a transplantation of traditional understanding of the nature and potential of arms control from the Cold War, bipolar, inter-state context into that of post-Cold War sub-state and sub-regional conflict. In "structural" terms, Cold War arms control was based on the presumption that each side had a right to exist, a belief that the resort to war was excessively costly and consequently a mutual desire to avoid it, a mutual understanding that a balance existed, a recognition that this balance was legitimate, and a shared belief that the point of arms control was to stabilise this balance. ₆ None of these presumptions obviously holds in the realm of inter-state, inter-group conflict. Moreover, success in arms control rested on the existence of coherent polities, with clear and reliable divisions of responsibility and lines of authority. This, too, is rare in contemporary conflict. Political structures are weak and contested, lines of authority are often ill-defined and ephemeral, and the polities or proto-polities engaged in conflict lack adequate control over other social forces operating in territory ostensibly under their control. These forces often pursue conflict-related agendas that coincide loosely, if at all, with those of the political "authorities." In many instances, groups in conflict have economic agendas that are served by prolonging war.(Keen, 1998; Berdal and Malone, 2000) Finally, the cultural contexts of contemporary conflict often differ substantially from those presumed in the traditional arms control paradigm, notably with regard to the desirability of agreement per se, the acceptability of transparency, norms of trust, honour, and shame, and culture's attitudes towards personal possession of weapons. (Krause, 1999, 4-10) As shall be seen below, many of these general propositions have strong resonance in the conflicts of the Southern Caucasus. There is, consequently, reason to view the potential for sub-regional arms control measures to contribute to conflict resolution with some scepticism in this zone of conflict.

The Origins and Course of Conflict in the Southern Caucasus

The previous section suggests that the utility of sub-regional arms control in conflict resolution depends strongly on the roots, course, and societal impact of the conflict(s) under consideration. With this in mind, this section examines the sources and dynamics of conflict in the southern Caucasus.

Origins

Since the collapse of the Soviet Union in 1991, the three republics of the Southern Caucasus (Armenia, Azerbaijan, and Georgia) have experienced three more or less serious internal conflicts - South Ossetia and Abkhazia in Georgia, and Nagorno-Karabakh in Azerbaijan.

All three conflicts have significant ethnic dimensions and were in important respects products of the uncertainties associated with the decline and then collapse of the USSR and the associated rise of titular nationalism in the union republics. This phenomenon was particularly strong in the Caucasus, as self-consciously ethnic nationalist movements took power in all three republics (the Karabakh Committee in Armenia, the Popular Front in Azerbaijan, and the Round Table-Free Georgia coalition in Georgia). As in many other instances of ethnic conflict, the problem was that the territories within which they sought to express their mythic vision of nation were not ethnically homogeneous (see Table 1).

Table 1. Ethnic Composition of the Southern Caucasian Republics (1989)

Group

Country	Armenian	Azerbaijani	Georgian	Russian	Other
Armenia	94	3		2	2
Azerbaijan	6	83		6	6
Georgia	8	6	70	6	10

Source: Dawisha and Parrott, 1994, 339-341

As is evident from the Table, only Armenia approached homogeneity, although it had a substantial Azerbaijani population divided between Yerevan and districts of eastern Armenia in proximity to the frontier with Azerbaijan.

Perhaps as importantly in accounting for conflict, a number of the minorities in the sub-region possessed autonomous structures of governance (see Table 2). The ethnic groups in whose behalf the jurisdictions had been created constituted the majority of the territory in question in three out of four instances. In the fourth (Abkhazia), the group in charge formed only 18% of the region's population.

Table 2. Autonomous Government Structures in the Southern Caucasian Republics

Armenia	Azerbaijan	Georgia
	Nagorno-Karabakh Autonomous Republic*	Ajar Autonomous Republic
		Abkhaz Autonomous Republic*
		South Osset Autonomous Oblast'*

* indicates a zone in conflict.

The elites clustered around these autonomous structures viewed the existence of the latter as the basis of their political power. The existence of these structures had always been contentious for the titular populations of the two republics in question. Indeed, the existence of the autonomies which risked the dismemberment of these new "nation-states," quickly became a cause célèbre amongst the burgeoning nationalist movements of Azerbaijan and Georgia.

The Soviet centre had been the principal means of protection from majoritarian nationalism for the region's autonomous jurisdictions. So, although in a general sense the collapse of the union generated considerable uncertainty - if not insecurity - for minority populations in the union republics, such perceptions were particularly strong amongst élites in areas which had enjoyed autonomy. Their default option in the context of rapid (and threatening) political change was exit. This explains, perhaps, why civil war in the region has centred on pre-existing autonomous entities, and why, in Georgia, ethnic civil war has engaged not the country's larger minorities (the Armenians, Azeris and Russians) but two of the smallest minority groups in the country (the Ossets at 3% of the country's 1989 population, and the Abkhaz at around 1%).

The combination of majoritarian chauvinism and minority secessionism - in the context of deepening general uncertainty at both mass and elite levels - destroyed the norms that underpinned the ethnic peace in Georgia and Azerbaijan. In both cases, the state was captured by one contending group and thus lost its capacity to act as a more or less impartial adjudicator of inter-group conflict. ₇ (Lake and Rothchild, 1996). Competing élites seized upon national contradictions as a legitimising device to secure their power.(Mansfield and Snyder, 1995) ₈ This "emergent anarchy" engendered a set of security dilemmas that deepened tension further.(Posen, 1993)

Evolution of the Conflicts

The first major outbreak of conflict in the sub-region came in 1988 in Nagorno-Karabakh. ₉ Change in the former Soviet Union brought increasing agitation for the removal on the Karabakh enclave from Azerbaijan and its unification with the Republic Armenia. Violence broke out in February 1988, rapidly followed by a series of pogroms against Armenians living in Azerbaijan's cities, and notably Sumgait. The rapidly declining security situation produced a massive exchange of population between Armenia and Azerbaijan, as minorities left for their kin states rather than facing the uncertainty of remaining within hostile host communities. This process was essentially complete (with the exception of the Karabakh enclave and the adjoining district of Shaumian in Azerbaijan) by the end of 1990. Meanwhile, the conflict escalated in a rather inchoate fashion until the collapse of the USSR in December 1991.

In 1992, with substantial assistance from the Republic of Armenia and from Russian forces stationed there, the Karabakh Armenians successfully consolidated their control over the Karabakh enclave itself while removing the minority Azeri and Kurdish populations resident in the region. This was followed by a successful breakout through the Lachin District to the Armenian border in mid-1992, establishing land communication with the kin state. From this point until mid-1994, the Karabakh Armenians and their allies succeeded in taking full control over districts between the enclave on the one hand and Armenia and Iran on the other, and establishing buffers to the east of Karabakh itself. Active hostilities ended in May 1994 in a Russian mediated cease-fire. Russia combined its mediating role with an effort to reinforce Armenian (and Karabakh) military capability through the massive transfer of heavy weaponry to Armenia in 1994-6 (the effects are evident in Table 4). Negotiations on a political settlement have proceeded since 1992 under the auspices of the OSCE (the Minsk Group), without success (see below).

The second conflict to erupt was that in South Ossetia in 1989-90. ₁₀ Tensions between Georgian and Osset populations within South Ossetia, and, more importantly, between rising Georgian nationalism and the identity concerns and perceived political interests of the leadership of South Ossetia produced small scale violence in South Ossetia at the beginning of 1990. The Supreme Soviet of the South Osset Autonomous Oblast' responded to these pressures in the autumn of 1990 by adopting a declaration transforming itself into the South Ossetian Soviet Democratic Republic. This initiated a sequence of constitutional initiatives by the Supreme Soviet that were subsequently annulled by the Georgian Parliament. This sequence culminated in December with a decision by the Georgian Parliament to abrogate the region's autonomous status.

Violence between the two sides broke out in December, engaging not only competing Georgian and Osset militias, but also a regiment of Russian MVD forces quartered in the regional capital, Tskhinvali. After the collapse of the USSR, this regiment withdrew, and the conflict intensified. During the period of active conflict, some 40,000 Georgians were displaced from the SOAO to the rest of Georgia while a substantial number of Ossets fled to North Ossetia, not only from the SOAO itself but from areas elsewhere in Georgia where they experienced "ethnic cleansing." The arrival of Osset refugees from Georgia had a knock-on effect in North Ossetia, as it increased competition for land and resources between Ingush and Ossets in the Prigorodnyi Raion (a region controlled by North Ossetia but claimed by Ingushetia). This contributed to the serious violence between Ossets and Ingush, and the cleansing of Ingush from their homes in areas under Osset jurisdiction.

The conflict continued until June 1992 when Russian President Boris Yel'tsin mediated the Sochi Accord, establishing a cease-fire and a framework for negotiation on a political settlement. The Accord provided for a mixed (Russian, Georgian, and Osset) peacekeeping force to be deployed between the two sides. It did not include any effective arms control or disarmament measures.

In December 1992, the OSCE established a mission of long term duration in Georgia and from that point forward, mission members have monitored the cease-fire and observed the activities of the peacekeepers. Although the situation has remained quite stable since then and significant progress has been made in re-establishing economic and other ties between the Osset community and its Georgian neighbours, little forward movement has occurred in the effort to define a mutually acceptable political settlement.

During the active phase of hostilities in South Ossetia, a civil war broke out amongst the Georgians themselves between supporters and opponents of elected president Zviad Gamsakhurdia. In December 1991, the president fled the country, first to Armenia and then to Chechnya. The Military Council that succeeded him invited Eduard Shevardnadze to return to Georgia, which he did in March 1992. Gamsakhurdia also returned in the spring of 1992, when his supporters mounted a guerrilla insurgency in his home region in western Georgia, Mingrelia. Georgian government and paramilitary efforts to suppress the rebellion were frustrated by the sympathy of the local population for the insurgents, by the ineffectiveness of Georgian forces and by the capacity of the zviadists to use sanctuaries in the Abkhaz Autonomous Republic as cover for their forces and as a hiding place for kidnapped officials of the central government.

Responding to this problem, Minister of Defence Tengiz Kitovani led a force into southern Abkhazia to deny the area to the rebels. Once there, he found the road open to the republic's capital, Sukhumi, and this at a time when the Abkhaz Supreme Soviet had just abrogated its Soviet era constitution, restoring the 1925 constitution, a document which significantly altered the relations between Abkhazia and the central Georgian government. This decision was annulled by the Parliament in Tbilisi, and Georgian forces attacked the Abkhaz Parliament, initiating the third conflict considered here. The Abkhaz portion of the Supreme Soviet fled to the north of the region and war began.

Over the course of the next year, benefiting from substantial assistance from Russian military units and north Caucasian volunteers from the Russia Federation, the Abkhaz slowly took control of much of northern Abkhazia and closed in on Sukhumi. In July 1993, active Russian mediation produced a cease-fire accord that contained important arms control provisions. Notably, it called for the withdrawal of Georgian forces and military equipment from the republic and for the encampment of Abkhaz troops and equipment under Russian supervision. The disengagement of the two sides and the establishment of Russian control over weapons in the zone of conflict was to enhance the confidence of the parties in Russian- and then UN-mediated efforts to reach a political settlement. At Georgian request, the United Nations (UN) agreed to the deployment of a small force to observe the implementation of the agreement.

Although Georgian forces more or less complied with the provisions of the accord, the Abkhaz did not, while the Russians failed to enforce the provisions of the agreement pertaining to Abkhaz military equipment. Consequently, the military balance shifted in favour of the Abkhaz. They took advantage of this a month later to evict Georgian authorities and the bulk of the Georgian (Mingrel) population from Abkhazia. At this stage, the Georgian government - facing renewed rebellion in Mingrelia - was near collapse and Shevardnadze travelled to Moscow, joined the CIS, agreed in principle to leasing arrangements for several Russian bases in Georgia, and requested Russian military assistance to stabilise the line along the Abkhaz border and to put down the rebellion in Mingrelia. The Russians complied, deploying between the two forces and dispatching marines to deal with the zviadists. This was followed by an agreed cease-fire and (in April 1994) by an interim agreement on cessation of hostilities. The agreement envisaged the establishment of a security zone and weapons exclusion zone on both sides of the border between Abkhazia and the rest of Georgia, and the return of displaced persons to their homes. The security zone was to be policed by a CIS peacekeeping force. Provisions for the weapons exclusion envisaged the concentration of heavy weaponry withdrawn from the line of contact to facilities where they would be monitored by the CIS peacekeeping force. In July 1994, the UN Security Council (UNSC) authorised the CIS peacekeeping mission, while expanding the UN observer force to 140 (UNSC/94/937).

Resolution 937 also called for the continuation of mediation by the Special Representative of the Secretary General in pursuit of a political settlement. These negotiations have continued for nearly six years; no real progress towards settlement has been made.

Significance of the Conflicts

The three conflicts under consideration here have had serious humanitarian, developmental, and political consequences. Together, they have resulted in some 50,000 deaths and a civilian population displacement in the area of 1.5 million people (an exchange of about 500,000 refugees between Armenia and Azerbaijan in 1988-90, displacement of some 600,000 Azeris from areas of Azerbaijan in and around Karabakh occupied by Armenian forces, some 100,000 refugees and displaced persons from South Ossetia and the rest of Georgia linked to the South Ossetia conflict, and some 350,000 persons displaced from Abkhazia, mainly to Western Georgia, but also to Russia, Armenia, and Greece. The conflicts have also hampered these state's political and economic transitions, and have impeded the development of regional co-operative arrangements that would further the economic development of all three. The persistence of the Abkhaz and Karabakh conflicts in particular considerably enhance the risks associated with the development of energy export infrastructure through western Azerbaijan and Georgia to the Black Sea and/or Turkey.

All three conflicts have had important external dimensions, complicating the fabric of interstate relations in the southern Caucasus. The conflict in Nagorno-Karabakh occasioned military intervention be the Republic of Armenia and reportedly by Russian troops stationed in Armenia. Azerbaijan mounted and has sustained an economic blockade of Armenia beginning since 1989, with disastrous humanitarian and economic consequences for the latter country. Turkey joined this blockade in 1993 after serious military reversals inflicted on Azerbaijani forces by the Karabakh Armenians. The conflict in Karabakh has prevented any normalization of the troubled Armenian-Turkish relationship.

In the case of South Ossetia, the insurgents were assisted by their co-ethnics in North Ossetia with the tacit approval of the Russian government. As already noted, the refugee flow from Georgia to North Ossetia has exacerbated processes of conflict in the northern Caucasus, and notably the simmering dispute between North Ossetia and Ingushetia over the Prigorodnyi Raion.

Russia also played an important role in the Abkhaz conflict through the provision of weapons and personnel to the Abkhaz side. Its performance as monitor and guarantor of the July 1993 cease-fire agreement resulted in the military advantage shifting to the Abkhaz side, laying the basis for the latter's victory in September 1993.

In a more general sense, all three conflicts serve the Russian objective of sustaining or enhancing the Federation's influence over the region and over its development. On the other hand, they also serve to sustain levels of regional instability that complicate Russia's effort to maintain order in the northern Caucasus.

Mediation in Southern Caucasian Conflicts

The effort to achieve political settlements to the conflicts of the southern Caucasus is divided between the UN and the OSCE. In the case of Karabakh, the OSCE assumed responsibility for the conflict in 1992, through the Minsk Group. ₁₁ The OSCE - in co-operation with the Russian Federation - also took the lead on the dispute in South Ossetia. In the case of Abkhazia, although originally both the OSCE and the UN engaged, the two organisations agreed in 1993 that the UN should take the lead in mediation under the auspices of the Special Representative of the Secretary-General (SRSG), of which there have been three - Eduard Brunner, Liviu Bota, and the current incumbent, Dieter Boden. The UNHCR led negotiations for a return of the displaced. UN-based mediation proceeds with Russian "facilitation." Several years ago, the UN and the Russians were joined in the effort to achieve a settlement by the so-called Friends of Georgia.

As noted, none of these negotiations has produced identifiable progress towards political settlement. The reasons for this are reasonably clear. Settlement of the Karabakh conflict is blocked by fundamental disagreement on status for the republic, with the Karabakh authorities and their Armenian supporters holding out for independence - or more modestly a horizontal or confederal arrangement with Azerbaijan - while Azerbaijan insists on a vertical relations. It has offered "maximal possible autonomy" but remains reluctant to spell out just what that means. Interim, step-by-step negotiation on such issues as return of the displaced to their homes in areas surrounding Karabakh or the return of some or all of the "occupied territories" to Azerbaijani jurisdiction is hampered by a number of factors. One is the Karabakh belief that these territories are their principal bargaining point in the overall negotiation. Another is the absolute Karabakh unwillingness to surrender control over the Lachin Corridor linking the enclave to Armenia in the absence of credible international guarantees of freedom of movement, access to Armenia, and security of the enclave itself. A third is the apparent incapacity of the international community to produce such guarantees.

In the case of South Ossetia, continuing disagreement between the parties regarding the level of autonomy that will be enjoyed by the region is again a significant obstacle. In that of Abkhazia, and similarly to Karabakh, the de facto authorities of the secessionist region are unwilling to accept a vertical relationship with Tbilisi or a substantial reduction of their capacity to defend themselves. The issue of return of the displaced is complicated by the fact that were the displaced ethnic Georgians to return they would probably constitute a majority in Abkhazia. In any subsequent election, the current authorities of Abkhazia would likely be removed. In all three conflicts, substantial economic interests (principally associated with smuggling) have developed in opposition to normalisation.

Underlying these specific impediments to settlement, a number of more general factors are relevant to the explanation of the intractability of the conflicts. One is, of course, the Russian factor. Many in the Russian policy-making apparatus apparently see the frozen conflicts as a source of leverage over Georgia and Azerbaijan on such issues as basing rights, co-operation on air defence and border control, and the profile of energy production and export from the Caspian Basin. In a more general sense, the conflicts may be seen as a means of preventing or impeding the westward turn of these two states.

Second, although the leaders themselves have on occasion demonstrated a willingness to compromise in pursuit of settlement, such initiatives are hard to sustain given the domestic political climate that they face. When, for example, Levon Ter Petrosyan (Armenia's former President) began to speak publicly of the need to recognise that Nagorno-Karabakh would remain part of Azerbaijan, he was removed from office. Evidence of movement towards peace in 1999 was accompanied by statements from the Azerbaijani opposition that any significant compromise would result in the removal of Geidar Aliev, while on the Armenian side, and probably for unrelated reasons, much of President Kocharian's government was eliminated in a mass assassination in the Parliament. Subsequent instability within the successor government makes it very difficult for the Armenian leadership to negotiate on the Karabakh question. In short, the capacity to make progress in negotiation requires that negotiators be able to bring domestic public and elite opinion along. This is an extremely difficult and delicate task.

That it is so difficult and delicate reflects a deeper problem. Short of secession (which would be unacceptable not only to the governments involved, but also, therefore, to the international community), a durable political settlement requires that the parties to conflict put their trust in shared institutions that they jointly control. (Lake and Rothchild, 1996) As noted earlier, the conflicts began in considerable measure because of uncertainty as to whether newly independent states were able to provide protection for minority populations and a deep and deepening distrust between the distinct populations of the region. The conflicts themselves, the economic havoc they have wrought, and the personal grievances they have engendered have degraded the climate of trust still further. There is good reason to doubt whether the societies of the region can generate levels of trust and perceived mutual interest sufficient to generate durable institutional solutions on their own. And there appears to be little evidence that international and regional organisations can compensate for this.

It is, therefore, not surprising that the record of efforts by international organisations to build on the region's reasonably durable cease-fire arrangements is not encouraging. It is noteworthy that the cease-fires themselves were a product of activity by a regional power (Russia) rather than by international organisations. International and regional organisations lack the will and the capacity to impose settlements on the parties in conflict. This contrasts rather starkly with the deep engagement of such organisations in such areas as the Balkans, and reflects the relative lack of interest of major states (other than Russia) in the affairs of the Southern Caucasus. Moreover, it is not clear whether Russia has reconciled itself or is likely to reconcile itself to such a proactive role by external organisations in the region. To judge from the record of the 1990s, Russia itself is unlikely to act as a surrogate for the international community in the quest for peace, because it is not at all clear that her perceived interests would be served by regional stabilization in the Southern Caucasus, because the Russia policy process has been so tangled that it is difficult for the Russian Federation to generate consistent and sustained policy on these issues, and because (despite the second round in Chechnya) it is not obvious that Russia has transcended the budgetary and personnel crises affecting her military during much of the 1990s.

Arms Control in the Southern Caucasus

Tables 1-5 in Appendix 1 provide data on the military balance in the Caucasian region. The record of arms control initiatives as an aspect of conflict management and resolution in this sub-region is rather limited. The only substantial case is that of Abkhazia in the context of stabilisation of cease-fires (the July 1993 encampment procedure and the 1994 weapons exclusion zone). The other serious contemplation of arms control as a component of political settlement occurred within the OSCE as part of planning for a multinational peace-keeping force of the Karabakh conflict. One element of the mandate was supervision and monitoring the withdrawal and concentration of heavy weapons, and "any zones of limitation," as well as the "disarmament and disbandment of uncontrolled armed groups," monitoring the air exclusion zone and the control of "air assets for military purposes." (OSCE, 1995) Of these two initiatives, the first was essentially coerced and the second consent-based. This presumably accounts for the relative success of the first; an external actor was willing and had the capacity and leverage to impose these limited arms control measures on both parties.

Little has been said about arms control arrangements as part of political settlement in the three conflicts, not least since there is profound disagreement as to the extent to which insurgent authorities will retain military capability. More broadly, there has been no substantial consideration of arms control as part of regional stabilisation, although the determination and adjustment of CFE flank limits has some bearing on the regional military balance and on confidence building amongst the states of the region. One might expect arms control to play a significant role in the various regional "zone of peace" and "stability pact proposals," ₁₂ but these are on the whole declaratory and quite general. Little effort has been expended in fleshing them out in sustained processes of negotiation. This is unsurprising, since Armenia and Azerbaijan - whose participation is essential to the success of such enterprises - remain locked in conflict over Karabakh, while there is no agreement amongst regional and external actors as to which contiguous states should be included in the process. One has to wonder also about the merit of regional security and CSBM arrangements that are limited to state actors, when half of the parties to the actual conflicts in the region are non-state actors.

Conclusion: The Potential for Arms Control in Settlement Processes in the Southern Caucasus

The emergence of effective sub-regional arms control and confidence-building arrangements in other zones of conflict raises the question whether such experiences might serve as models for similar security-enhancing measures in the southern Caucasus. In Bosnia-Herzegovina, for example, Annex 1B of the Dayton Agreements mandated the OSCE to develop sub-regional and regional arms control and CSBM measures for Bosnia-Herzegovina. In January 1996, the Agreement on CSBMs in Bosnia and Herzegovina was concluded under the auspices of the OSCE. The agreement "provided for a comprehensive set of measures to enhance mutual confidence and reduce the risk of conflict - such as exchange of military information, notification as well as observation and constraints on certain military activities, restrictions on military deployments and exercises in certain geographic areas, withdrawal of forces and heavy weapons to cantonments or designated emplacements, etc."(OSCE, 1999) No problems have been reported with the implementation of the Agreement, which has included over 130 inspections.

This was followed six months later by the Florence Agreement on Sub-regional Arms Control between the three parties in Bosnia-Herzegovina and also Croatia and the Federal Republic of Yugoslavia, establishing ceilings on five categories of conventional armaments. By the end of the statutory reduction period, some 6,600 weapons were put out of service by the parties.

Both agreements are judged to have been extremely successful. In some respects (e.g. high levels of inter-group animosity, very low levels of trust both between groups engaged in conflict, and of the state as a potentially impartial adjudicator of conflict), there is some reason to believe that similar arrangements might have a similarly positive effect in the southern Caucasus. To take one example, monitored demilitarisation of the occupied territories around Nagorno-Karabakh would facilitate the return of displaced persons to their homes in conditions that would reassure the Karabakh Armenians. Efforts to disarm paramilitary partisans operating from bases in Georgian-held territory and sheltering amidst spontaneous returnees in Abkhazia's Gali District would reduce the sense of insecurity of the Abkhaz authorities as they contemplate movement towards an agreement on status. At a regional level, control over imports into the region and intrusive monitoring regimes could reduce the probability of an arms race between Armenia and Azerbaijan. Concentration and encampment of military personnel and heavy weapons are a necessary element of any settlement of the civil conflicts in both Georgia and Azerbaijan. Joint patrols and inspection regimes along the border between Georgia and Chechnya ₁₃ and between Azerbaijan and Daghestan would reduce Russian concerns that the territories of these neighbouring states were being used as staging areas for ongoing insurgencies in the Northern Caucasus.

However, there is a key contextual difference between the Balkan and the southern Caucasian situations. In the latter case, the agreements were part of a conflict resolution framework imposed by the United States and enforced by the deployment of a substantial NATO force (IFOR-SFOR) into the country with a deterrence and stabilisation mandate and the firepower to fulfil that mandate. There is little if any prospect that a similar level of external intervention and long term commitment would be forthcoming in the context of settlement of the conflicts in the southern Caucasus. Absent such intervention, it is extremely difficult for the parties themselves to develop the negotiating momentum to achieve stabilising arms control arrangements. They remain locked in a zero-sum game, deeply sensitive to relative gain. Moreover, it is noteworthy that the subregional arms control and confidence-building regime in Bosnia followed the parties' acceptance of a formal political settlement. There is little evidence of progress towards political agreements on the issues separating adversaries at both the intra-state and inter-state levels.

The regional context is also considerably different. The former Yugoslavia is surrounded by states that share a strong perceived interest in conflict resolution and regional détente. The southern Caucasus is surrounded by regional great powers (Turkey, Iran and Russia) that remain deeply suspicious of each other's intentions in the region and locked in a competition for influence over its affairs.

And finally, recalling the structural and cultural impediments mentioned in the earlier discussion of the post-Cold War role of arms control, several points deserve emphasis. The region taken broadly is not characterised by a relative balance that could operate as a basis for arms control. The imbalance favouring Russia is perceived by Georgia and Azerbaijan to be a profound security problem. Russia on the other hand would be unlikely to give up this asymmetry, given its difficulties in the Northern Caucasus, as well as its continuing strategic concerns regarding Turkey. In substate conflicts, the very legitimacy of the idea of balance is questionable for the governments involved. State and sub-state actors in the region do not necessarily share a sense that the costs of the use of force are prohibitive. Indeed, Osset, Abkhaz, and Karabakh uses of force were successful in attaining the military/strategic objectives of these parties.

The parties do not fully control the military effectives supporting them; nor do they possess a relative monopoly of force. Chains of command are frequently loose, the region is afflicted by numerous military/economic actors who benefit from instability, and small weapons are widely spread throughout society. The latter is related to cultural issues: possession of weapons is considered culturally acceptable if not desirable in traditional southern Caucasian cultures. Moreover, the recent experience of widespread disorder, the manifest incapacity and unwillingness of police forces to maintain law and order, and the systemic corruption of state judicial and police agencies render it unlikely people would be willing to surrender their personal means of protection.

All these factors suggest that we are a long way from effective arms control and CSBMs in the region. Although such efforts are desirable, the impediments to an effective use of such tools in the pursuit of regional stability should not be underestimated.

Appendix 1. The Military Balance in the Caucasus ₁₄

Table 3. The Military Balance in 1991-97

Country

Year

Def

Exp

($m)

Def Bdgt

($m)

Total Armed Forces

Reserve

Army

Navy

Air Force

MBT

AIFV (APC/ACV)

TOTAL ARTY

SAM

(HEL)

Armenia

1990

1992

1994

1996

1998

122

n.k.

147*

50000

32700

57400

55700

n.k.

300000

50000

32700

56600

52000

n.k.

3700

258

77.

120

102

n.k. (641)

n.k (189)

225 (108)

202 (83)

168 (50)

375

160

225

n.k.

0 (7)

3 (13)

6 (30)

ε7 (21)

6 (16)

Azerbaijan

1990

1992

1994

1996

1998

125

132

130

n.k.

165*

132

115

190

ε47000

56000

70700

72150

n.k.

560000

575700

5000

49000

57000

55600

n.k.

3000

2200

n.k.

2000

11200

10350**

391

134

279

300

270

n.k. (1285)

n. k. (113)

320 (502)

431 (149)

287 (74)

463

126

350

302

301

n.k.

some

60+

124 (24)

15 (9)

48 (23)

46 (18)

37 (15)

Georgia

1990

1992

1994

1996

1998

106

108

n.k.

20000

n.k.

33200

n.k.

250000

3000

10000

ε10000

12600

n.k.

2000

n.k.

200

1000

3000**

850

n.k. (1054)

n.k. (28)

both 70

65 (12)

both 111

363

80-100

107+

n.k.

110

some

245 (48)

15 (15)

2 (1)

9 (8)

Russia

(North Caucasus

MD)

1990

1992

1994

1996

1998

n.a.

n.k.

65000***

56000***

n.k.

65000

54500

n.k.

1500

n.k.

1army

1Army

550

380

ε536

180 (250)

0 (0/1270)

0 (130/2227)

530

630

ε524

n.k.

150

125

880 (0)

ε709(60)****

ε775 (80)****

* Converted according to the official rate of rouble in 1991.

** including Air Defence.

*** plus 1 tactical Air Army and 3 regular AD.

**** including 200 training L-39.

Source: The IISS Military Balance (1990-91, 1992-93, 1994-95, 1996-97, 1998-99).

Table 4. Manpower and Treaty Limited Equipment: Current holdings and CFE limits on the Forces of the Countries of the Caucasus

(derived from data declared as of 1 January 1998)

Country	Manpower Holdings	Manpower Limit	Tank Holdings	Tanks Limit	ACV Holdings	ACV Limit	Artillery Holdings	Artillery Limit	Attack Hel Holdings	Attack Hel Limit	Combat AC Holding	Combat AC Limit
Armenia	60000	60000	102	220	218	220	225	285	7	50	6	100
Azerbaijan	69941	70000	270	220	557	220	301	285	15	50	48	100
Georgia	30000	40000	79	220	111	220	107	285	3	50	7	100
Russiaª				1300		1380		1680		na		na

ª Flank zone.

Source: The IISS Military Balance (1990-91, 1992-93, 1994-95, 1996-97, 1998-99).

The SIRPI Yearbook 1993 p. 597.

Table 5. Russian armed forces and UN observers in the Caucasus

Host Country	Year	Pers	Equipment	Peacekeeping	Obs
Armenia	1992 1994 1996 1998	23000 5000 4300 4100	MBT 250, 350 AIFV, 350 arty, 80 SAM, 7 hel, MBT 80, APC190, arty 100 MBT 75, ACV140, arty 85, 1 ac sqn MBT 74, APC 17, ACV148, arty 84, 1 ac sqn	— — - -	— — - -
Azerbaijan	1992	62000	MBT 400, AIFV 800, arty 470, 14 hel, ac 120, SAM 135	—	—
Georgia	1992 1994 1996 1998	20000 20000 8500 9200	MBT 850, AIFV 680, arty 370, 48 hel, ac 240, SAM 175 MBT 230, ACV 300, arty 220, AC and hel 35 MBT 110, ACV 510, arty 238, AC and hel 35 MBT 140, ACV 500, arty 173, AC and hel 35	1 AB bn ε 2500 ε1700 (1AB regt, 2MR bn) ε1500 (3 MR, 1 inf bn) ACV 118	— 21 83 127

Source: The IISS Military Balance (1990-91, 1992-93, 1994-95, 1996-97, 1998-99).

Table 6. Opposition Forces

Country1

Year

Oppostion Force

Total

Personnel

Equipment

Azerbaijan

1992

1994

1996

1998

Armenian Armed Groups

in Nagorno-Karabakh2

ε 30-50000

ε20000

ε20-25000

n.k.

some MBT, APC, arty

MBT 50+; AIFV/APC 100+; arty 100

MBT151+; ACV 298; arty 298+

Georgia

1992

1994

1996

1998

Abkhazia

South Ossetia

Abkhazia

South Ossetia

Abkhazia

South Ossetia

Abkhazia

South Ossetia

no details

4000

no details

ε5000

ε2000

ε5000

ε2000

n.k.

some MBT, ACV, arty

n.k.

MBT 50+; AIFV/APC 80+; arty 80+

MBT 30+; AIFV/APC 30

MBT 50+; AIFV/APC 80+; arty 80+

MBT 5-10; AIFV/APC 30, arty 25

Source: The IISS Military Balance (1990-91, 1992-93, 1994-95, 1996-97, 1998-99).
____________________________
1 Armenia has no armed opposition forces
2 including volunteers from Armenia
______________________________

Table 7. Navy Statistics in the Caucasus

Country

Year

Personnel

Equipment

Azerbaijan

1991

1993

1995

1997

n.k.

ε 3000

ε 2200

2200

Some 5 mine warfare

16 naval units

2 principal surface combatants, 21 patrol and coastal combatants, 14 mine

countermeasures, 1 amphibious, 2 support and miscellaneous

2 principal surface combatants, 18 patrol and coastal combatants,

15 mine countermeasures, 4 amphibious, 2 support and miscellaneous

Georgia

1991

1993

1995

1997

n.k.

2000

n.k.

Some units of the Black Sea Fleet

15 patrol and coastal combatants (2 frigates, 1 torpedo craft, 12 patrol craft)

12 patrol and coastal combatants, 6 amphibious

Source: The IISS Military Balance (1990-91, 1992-93, 1994-95, 1996-97, 1998-99).

Appendix 2: Acronyms

ABM          Anti-Ballistic Missile
CFE          Conventional Forces in Europe
CIS           Commonwealth of Independent States
CSBM      Confidence and Security Building Measures
CSCE      Conference on Security and Co-operation in Europe
KFOR       Kosovo Force (NATO)
MTCR       Missile Technology Control Regime
NATO        North Atlantic Treaty Organization
NPT           Non-Proliferation Treaty
OSCE       Organisation for Security and Co-operation in Europe
SALT         Strategic Arms Limitation Treaty
SRSG       Special Representative of the Secretary-General
START      Strategic Arms Reduction Process
TLE           Treaty Limited Equipment (CFE)
UN             United Nations
UNHCR    United Nations High Commissioner for Refugees
UNOMIG   United Nations Observer Mission in Georgia
USSR        Union of Soviet Socialist Republics

Appendix 3: Bibliography

Berdal, Mats, 1996. Disarmament and Demobilisation after Civil Wars, Adelphi Paper No 303 (London: IISS).

Berdal, Mats and David Malone, 2000. Economic Agendas in Civil Wars (Boulder: Lynne Rienner).

Buzan, Barry, 1991. People, States and Dear (London: Harvester Wheatsheaf).

Dawisha, Karen and Bruce Parrott, 1994. Russia and the New States of Eurasia (Cambridge: Cambridge University Press).

Holsti, Kal, 1999. "The Coming Chaos? Armed Conflict in the World's Periphery," in T.V. Paul and Robert A. Hall, eds., International Order and the Future of World Politics (Cambridge: Cambridge University Press, 1999), 283-310.

Keen, David, 1998. The Economic Functions of Violence in Civil Wars, Adelphi Paper No. 320 (London: IISS).

Krause, Keith, 1999. "Structural and Cultural Challenges to Arms Control in Intra-State and Post-Conflict Environments" (paper prepared for the 1999 Nobel Symposium), 2 October, 1999, mimeo.

Lake, David and Donald Rothchild, 1996. "The Origins and Management of Ethnic Conflict," International Security XXI, No.2 (Fall).

MacFarlane, S. Neil, Larry Minear, and Stephen Shenfield, 1996. Armed Conflict in Georgia: A Case Study in Humanitarian Action and Peacekeeping Occasional Paper No.21 (Providence, RI: The Watson Institute).

MacFarlane, 1997. "Democratization, Nationalism and Regional Security in the Southern Caucasus," Government and Opposition XXXII, No. 3 (Summer).

MacFarlane, S. Neil and Larry Minear, 1997.Humanitarian Action and Politics: The Case of Nagorno-Karabakh, Occasional Paper No. 25 (Providence, RI: The Watson Institute).

Mansfield, Edward and Jack Snyder, 1995. "Democratization and the Danger of War," International Security XX, No.1 (Summer).

OSCE, 1995. OSCE High Level Planning Group, "Concepts for the OSCE Multinational Peacekeeping Mission for the Nagorno-Karabakh Conflict" (Vienna, 15 November, 1995).

OSCE, 1999. OSCE Handbook, Online Edition. Http://www.osce.org/publications/handbook/9htm.

Posen, Barry, 1993. "The Security Dilemma and Ethnic Conflict," Survival XXXV, No. 1 (1993), 27-47.

Survival, 1993. "Ethnic Conflict and International Security," Survival, XXXV, No. 1 (Spring 1993).

_____________________________________________

Note 1: S. Neil MacFarlane is Lester B. Pearson Professor of International Relations and Director of the Centre for International Studies at the University of Oxford. He is grateful for the assistance of Georgre Khelashvili in the preparation of this paper. Back.

Note 2: This paper deals with the southern Caucasian republics - Georgia, Armenia and Azerbaijan - except to the extent that consideration of events and trends in the northern Caucasian subjects of the Russian Federation impinge significantly on the security of these three states. Back.

Note 3: Officials and politicians in both Nagorno-Karabakh and Georgia reacted to the Dayton Accords by clling for Dayton-type solutions to the Karabakh and Abkhaz conflicts. In 1999, Georgian officials called similarly for a KFOR-type solution to the Abkhaz conflict, an odd suggestion, given that in the case of Kosovo, NATO intervened in behalf of a minority struggling for self-determination. Back.

Note 4: This is not to say that such phenomena were absent earlier. Many Cold War conflicts were as much internal as external and had ethnic dimensions (e.g. Angola, Guatemala). The Cold War period witnessed numerous internal conflicts that had no obvious or causally significant connection to the Cold War (e.g. the civil war in Nigeria or the simmering dispute in Ceylon/Sri Lanka). Back.

Note 5: This has translated into a decline in global military expenditure (Krause, 1999), but this level of aggregation masks considerable regional variation. Until the Asian economic crisis of 1998, for example, there was little evidence of regional decline in Asia. Likewise, although there has been dramatic decrease in military spending by South Africa, there is little evidence of regional pattern. Finally, focusing on government budgets in estimating trends in global military spending ignores the rising significance of non-state actors in the military realm. Back.

Note 6: The use of the word "structural" is interesting in this context, since each of these aspects is as much cognitive as material. Back.

Note 7: For a discussion of this role of the state in multi-ethnic societies, see Lake and Rothchild, 1996. Back.

Note 8: This should not be read as an uncritical embrace of instrumental explanations of nationalism during periods of democratic and liberal transition. For a more complete account, see MacFarlane, 1997. Back.

Note 9: The origins and course of the Karabakh conflict are treated in greater detail in MacFarlane and Minear, 1997,7-36 and 83-100. Back.

Note 10: This conflict is described more fully in MacFarlane, et.al., 1996, 8-9, 49-50. Back.

Note 11: The Minsk Group was established by eleven interested OSCE members to lay the basis for a conference in Minsk, Belarus where a formal settlement of the Karabakh dispute was to be signed. Since the basis was never laid, the conference has yet to be held. Back.

Note 12: The first of these was an initiative by Dzhokhar Dudaev in 1991. In 1996, Eduard Shevardnadze launched the "Peaceful Caucasus Initiative," involving Russia, Georgia, Azerbaijan and Armenia. In November of 1999, Azerbaijan's President Geidar Aliev proposed a Caucasus Security Pact at the Istanbul OSCE Summit, expanding the circle to include Turkey, the US and the European Union. This proposal was adopted by Turkish President Suleyman Demirel by year's end. Armenia also embraced the idea in principle, but added Iran in a three (Armenia, Azerbaijan, Georgia as major parties) plus three (Turkey, Russia, Iran as guarantors of regional stability) plus two (the US and the EU providing external support) variant. Back.

Note 13: Such arrangements are developing. The Georgian Government has accepted in principle participation of Russian personnel in border patrols while the OSCE is mounting a monitoring mission along the border. Back.

Note 14: These data do not take into account changes in the disposition of Russian forces in the northern Caucasus as a result of the renewal of conflict in Chechnya. Back.