Military confidence-building in crises: lessons from Georgia and Ukraine

ABSTRACT What role can conventional arms control (CAC) and confidence- and security-building measures (CSBMs) play in crises and conflicts? By examining the use of CAC and CSBMs during the Russian-Georgian war in 2008 and during Russia’s annexation of Crimea and the war in eastern Ukraine in 2014, the present article argue that CAC and CSBMs are unable to prevent intentional conflict. Their more realistic value in crises is to function as early warning mechanisms that raise the bar for and increase the costs of conflict as well as serving as instruments to monitor conflicts. To increase their effectiveness, the links between early warning and early action need strengthening and several ambiguities need to be removed, particularly form the Vienna Document, in order to improve indicators, increase warning times and raise the political costs of non-compliance. Nonetheless, when one or two sides seek a conflict, CAC and CSBMs do not provide a remedy for conflict prevention.


Introduction
Since Russia's illegal annexation of Crimea and military aggression in Ukraine, a political-military climate of mistrust has taken hold in Europe as Russia stands against the West in a new geopolitical confrontation. In response to Russia's violation of Ukraine's territorial integrity, the United States and the EU introduced economic sanctions against Russia, and NATO decided to establish an enhanced forward presence along its eastern flank to deter Russia. At the other end of the spectrum, some political leaders and experts have increasingly emphasized the need for rebuilding confidence and trust in Russia-West relations for the ultimate purpose of reincorporating Russia in a revamped European security architecture (OSCE 2016a;The Economist, 2019, 9).
In the latter category, one stream of thinking advances the idea of re-launching a dialogue on conventional arms control (CAC) and confidence-and securitybuilding measures (CSBMs) in the Organization for Security and Cooperation in Europe (OSCE) as a way forward. Since Russia unilaterally suspended its application of the Treaty on Conventional Armed Forces in Europe (CFE Treaty) in 2007, the system of CAC in Europe has been in a process of steady decline. During Donald Trump's presidency, the US has consistently withdrawn from several arms control regimes, including the Open Skies Treaty. Moreover, in recent years a range of compliance problems has also beset another key pillar of CAC and CSBMs in Europe -the Vienna Document on Confidence-and Security-Building Measures in Europe (Vienna Document).
The idea to revive CAC and CSBMs was first launched by then-German Minister of Foreign Affairs Frank-Walter Steinmeier (2016) in the summer of 2016. His proposal was subsequently channelled into a broader compromise -the Hamburg declaration -in which the OSCE's participating states "committed to launching a structured dialogue on security and arms control" (OSCE 2016b). Since then, several experts and scholars have provided support for the proposal arguing that efforts in the field of CAC and CSBMs can provide much needed transparency and conflict preventive tools in times of instability and unpredictability. From different angles, analysts argue the merits of conventional arms control as an approach to reduce tensions, avoid an arms race, rebuild trust between Russia and the West, and ultimately prevent the outbreak of armed conflicts (Zellner (2015)); Richter (2016), (Richter 2017); Kulesa (2017); Perthes and Meier (2017); Zellner et al. (2018); Buzhinskiy and Shakirov (2019); Shakirov (2019).
While these analysts compellingly identify various problems to which CAC and CSBMs can offer potential solutions, the discussions nonetheless tend to be argumentative rather than based on historical record of how the instruments were tried and tested during actual crises situations. In the past decade, two armed conflicts have taken place in Europe -in Georgia in 2008 and the war in Ukraine, since 2014. By examining how CAC and CSBMs were deployed in order to stabilize these two conflicts, this article aims to draw empirically grounded lessons regarding what can realistically be expected of CAC and CSBMs for restraining and managing crises and conflicts in contemporary Europe, including insights into how the flexibility and resilience of the instruments could be strengthened. Consequently, the article does not focus on evaluating the overall strategic significance of arms control and confidence-building for European security, but limits the investigative scope to how the instruments can play a role in addressing conflict escalation. Special attention is paid to the Vienna Document, which is the most frequently used tool in these conflicts.
The article proceeds by first briefly sketching the basic tenets of the major arrangements in the interconnected fields of CAC and CSBMs, including the role of the OSCE. The second part details how CAC and CSBMs were used in attempting to defuse the situation between the conflicting parties on the ground in Georgia in 2008 and Ukraine in 2014. Thereafter, the third section identifies the major lessons from the use of CAC and CSBMs for raising the bar for military conflict in Europe. The final section sets out the conclusions as well as some principal recommendations.
Temporally, the study focuses on the course of events taking place in 2008 and in 2014 respectively. A detailed account of the background to and the aftermath of the conflicts thus fall outside the scope of this article. The article draws on a combination of written and oral sources. Among the written sources are official documents and statements. The secondary literature on the armed conflicts also discusses the various efforts to use arms control instruments. To complement the written documentation, the author conducted interviews with selected undisclosed diplomats and military experts involved in conflict prevention work in various professional capacities before, during and after the outbreak of the wars in Georgia and Ukraine.

OSCE and the interlocking web of CAC and CSBMs
The OSCE plays a key role as a venue for CAC and CSBMs among what is now 57 participating states. The OSCE framework for arms control from 1996 stresses that arms control, including disarmament and CSBMs, constitutes an integral part of the organization's concept of security. In the sphere of CAC and CSBMs, the three major instruments -the CFE Treaty, the Vienna Document and the Open Skies Treaty -are described as a web of interlinking and mutually reinforcing commitments (OSCE undated).
The CFE Treaty -known as the military cornerstone of European security -was negotiated during the final years of the Cold War, and was signed on November 1990 by the then 22 members of the North Atlantic Treaty Organization (NATO) and the members of the Warsaw Pact. Since both the Warsaw Pact and the Soviet Union broke up shortly afterwards, 30 countries eventually ratified the legally binding treaty in 1992. The CFE Treaty regulates the possession and location of five categories of heavy military equipment, or so-called treaty limited equipment (TLE) -battle tanks, armored combat vehicles, heavy artillery pieces, combat aircraft and attack helicopters -from the Atlantic to the Urals. The basic principle behind the CFE Treaty is to limit the quantities and locations of weapons in order to reduce the risk of surprise attacks and the concentration of forces. To ensure compliance, the treaty stipulates a comprehensive set of intrusive verification measures that emphasizes onsite inspections (Treaty on Conventional Armed Forces in Europe 1990, 1). Taken together, the central pillars of limitations, information exchanges and verification made the CFE Treaty a critical part in shifting the European security environment away from insecurity and distrust to cooperation and confidence-building. Thus, the CFE Treaty had a profound imact on military security in Europe by reducing tensions, dismantling heavy military equipment and introducing a level of transparency that would have been unimaginable at the height of the Cold war.
The CFE Treaty was however designed for the Cold War and in order to adjust it to the post-bloc reality in Europe, which had rendered the limitation regime of the original bloc-based treaty increasingly obsolete, a new Adapted CFE Treaty was negotiated and signed in Istanbul in 1999. While Russia ratified the new treaty, western states rejected ratification on the ground that Moscow had failed to adhere to the so-called "Istanbul commitments" -politically binding pledges to withdraw its forces and equipment from the territories of Georgia and Moldova (Kühn 2013, 191). Dissatisfied with western refusal to ratify the Adapted CFE Treaty, and frustrated by the restrictions of the outdated original treaty, Russia decided to unilaterally suspend its application of the CFE in 2007. Since then, Russia has refused to provide any data or allow any inspections, leaving the treaty essentially dead.
The Vienna Document on CSBMs was originally created in 1990, although it has been slightly revised in several rounds since then (in 1992, 1994, 1999 and 2011). It is the essential OSCE document on CSBMs and constitutes an integral part of the organization's all-encompassing approach to security as first outlined in the Helsinki Final Act of 1975 and subsequently confirmed in the Paris Charter of 1990. Politically but not legally binding, the Vienna Document's most significant contribution to European security lies in enhancing military transparency among the 57 OSCE participating states. Its major provisions are on exchanges of military information and details of defense policy and expenditure, and enabling inspection and observation of certain military activities, as well as rules on notification of exercises and new deployments (of over 9,000 troops) and rules on observation of certain military activities (exceeding 13,000 troops). As a CSBM, the Vienna Document is intended to promote transparency concerning how participating states organize their military forces (Vienna Document 2011 on Confidence-and Security-Building Measures 2011, p. 1).
The Open Skies Treaty was created when a group of 27 states in March 1992 decided to open their airspace to unarmed observation flights, and to collect data on military forces and activities on the territories of other signatories to the treaty. After a lengthy ratification process, the Open Skies Treaty entered into force on 1 January 2002. Shortly thereafter eight additional states joined, among them Sweden, Finland and the Baltic states. The geographical area of application of the Open Skies treaty extends beyond that of the CFE Treaty by incorporating North America and Siberia. Its declared purpose is to: "improve openness and transparency, to facilitate the monitoring of compliance with existing or future arms control agreements and to strengthen the capacity for conflict prevention and crisis management in the framework of the Conference on Security and Co-operation in Europe and in other relevant international institutions" (Treaty on Open Skies 1992, 1). Thus, if the Vienna Document is intended to increase confidence to facilitate disarmament and the CFE Treaty is intended to prevent massive surprise attacks, the Open Skies Treaty has a less clear-cut purpose that includes conflict prevention and crisis management. In May 2020, the US announced its withdrawal from the treaty, citing repeated Russian violations of its terms as the reason. Thus, further undermining the post-Cold War system of CAC and CSBMs.
The OSCE also has a broader political-military toolbox at its disposal for confidenceand security-building purposes. Since the early 1990 s, the OSCE has been involved extensively in the prevention, management and resolution of conflicts through its field operations. The OSCE field missions in various countries operate only with the consent of the host country, and a number of them concentrate specifically on early warning and conflict prevention. In the case of an emerging or ongoing crisis, field missions observe and report on developments on the ground as well as assist with facilitating dialogue and mediation and promote confidence building.

Georgia
In the early 1990 s, nationalist animosities between an ultra-nationalist Georgian government in Tbilisi and separatist movements in Abkhazia and South Ossetia escalated into two successive secessionist wars. Russia intervened in the conflicts at an early stage and after the conclusion of ceasefire agreements Moscow gained a role as peacekeeper and negotiator in the conflicts. Over the next 15 years, no progress on conflict settlement occurred, not least since Russia used its leading role to support the de facto secession of Abkhazia and South Ossetia from Georgia. In August 2008, a five-day war broke out between Russia and Georgia over South Ossetia. This short war took the world by surprise, and dramatically changed the geopolitical landscape not only in the Caucasus region, but also in Europe. In 2012, Vladimir Putin admitted that a plan for military action against Georgia had been prepared as early as 2006 and approved by him in 2007 (Felgenhauer 2012).
Escalation had begun the year prior to the war with the two serious military incidents in March and August of 2007. While the first incident, related to a Russian strike against Georgian villages in Abkhazia, received scant international attention, the second, an incident with an undetonated missile near a Georgian radar station just outside South Ossetia, brought a stronger reaction. Two intergovernmental fact-finding missions conducted onsite inspections and determined that aircraft flying from Russian airspace were responsible (Report from the International Group of Experts 2007; Second Independent Intergovernmental Expert Group 2007). As a response to the missile incident, several participating states pushed for increasing the number of OSCE military observers in South Ossetia. Since observers can only be deployed by consensus, the attempt to expand the field operation failed to materialize due to Russia opposing such decision (Interview diplomat, 19 February 2019).
In the spring and early summer of 2008, the international community attempted several CAC and CSBMs, with some degree of success, to stabilize the situation in and around Abkhazia, which was at that time the prime area of tension in Georgia. The CFE Treaty, the Open Skies Treaty and the Vienna Document were all used at various points in the prelude to the Russian-Georgian conflict in August 2008. First, both Georgia and Russia were signatories to the Open Skies Treaty. On April 1-4, 2008, Italy led an Open Skies observation flight over Georgia. The flight, which was scheduled in accordance with the quota distribution procedures of the treaty in 2007, monitored Georgian military activities near Abkhazia and reported unusual military preparations on the Georgian side (Schmidt 2016, p. 270;Grigalashvili 2018, 8).
Second, as Russian military activity intensified in the North Caucasus, Georgia bolstered its military presence near Abkhazia and sent out unmanned aerial vehicles (UAVs) to try to assess the situation. On April 20, a Russian MiG-29 aircraft downed a Georgian reconnaissance UAV in the airspace over Abkhazia. According to an international investigation of the incident, the activities of both parties were in violation of the Moscow Agreement on a Ceasefire and Separation of Forces of 1994. Moreover, the Russian aircraft responsible for the downing was not part of the peacekeeping force, which represented an additional violation (UNOMIG undated). In relation to the incident, the conflicting states, Georgia and Russia, in late May activated the Mechanism for Consultation and Cooperation as Regards Unusual Military Activities, paragraph 16 of the Vienna Document's Chapter III on Risk Reduction. 1 This began when the Georgian delegation to the OSCE sent a note to its Russian counterpart requesting an explanation for the involvement of a Russian aircraft in the destruction of the Georgian UAV on April 20. Russia, in turn, retorted with a request of its own regarding the legitimacy of Georgia sending UAVs over the Georgian-Abkhaz conflict zone, seeing it as an infraction of the 1994 Moscow Agreement and of United Nations Security Council resolution 1808. Since neither state was satisfied with the respective responses, bilateral talks on the subject were held in the presence of the OSCE Chairman in Office (CiO) and other delegations on June 4. Having failed to resolve the issue, the discussions moved on to a joint FSC-PC meeting on June 11 (Pacer 2016, 92-4). On behalf of the OSCE, a joint statement of the Forum for Security and Cooperation (FSC) and Permanent Council (PC) chairpersons declared that the meeting "showed the continuing relevance of this Mechanism and the OSCE as a forum for dialogue on the security concerns of participating States" (OSCE 2008b). However, in a context of escalating hostility and deep mistrust, others have seen the activation of the consultative mechanism in the Vienna Document as a pawn that the two adversaries were able to use against each other, and resulted in no specific proposals for de-escalating the conflict (Pacer 2016;OSCE 2008a).
Third, unlike Russia, which suspended its application of the CFE Treaty in December 2007, Georgia provided information on its Treaty Limited Equipment (TLE) and allowed CFE inspections right up until the war. As part of its military modernization program under President Mikheil Saakashvili, Georgia reported a substantial increase in troops and TLE falling under the CFE between January 2005 and January 2008. While Russia criticized the military build-up and labeled Georgia "the most militarized country on the planet", Georgian armament nonetheless fell below the ceilings stipulated for the country in the treaty (OSCE 2008e). In the middle of June 2008, Georgia received a multinational CFE inspection near Gori close to South Ossetia. The inspection found "that Georgia no longer had 70% of its troops at the peacetime location" (Schmidt 2009, p. 22;Interview military expert, 12 December 2018). However, at the request of the Georgian escort team, the follow-up inspection of these formations was cancelled. According to Schmidt (2009, 22): "The results of this inspection were not recognized sufficiently on an international level before the war".
On the ground, the OSCE had military monitoring officers in the South Ossetian capital of Tskhinvali observing the security situation in that region (Lynch 2010a, 144). The OSCE Mission to Georgia carried out early warning and conflict prevention activities. Throughout the summer, the mission sent Activity and Spot Reports back to Vienna, indicating rising tensions that called for an international response. In early July, the OSCE sent a delegation of 22 ambassadors, accompanied by military advisors from Germany, Latvia and Sweden, to South Ossetia, including the town of Java, Tskhinvali and even the Roki tunnel. Discussions with representatives of the South Ossetian de facto leadership, Georgian authorities and Russian peacekeepers led the delegation to assess that the situation was so tense that the logical next step in the conflict was war (Interviews military expert, 12 December 2018; diplomat, 19 February 2019). Less than a month later, the missiles were launching.
Before that, in mid-July, Russia began a major military exercise Kavkaz 2008 in what was then the North Caucasus military district, next to the border with Georgia. Officially, the exercise activated 8,000 troops, 700 combat vehicles and more than 30 aircraft and attack helicopters, just below the threshold for inviting international observers. Several of the units that participated in the exercise were later used in the five-day war (Larsson 2008, 27). 2 While the exercise ended on August 2, some troops and equipment remained by the Georgian border rather than returning to their home bases. At the same time, in parallel with mounting escalation in South Ossetia, the US-led annual military exercise "Immediate Response" took place at the Vaziani base outside Tbilisi, with participating troops from Georgia, the US, Armenia, Azerbaijan and Ukraine.
In early August, OSCE monitors documented that tensions had been aggravated further on the ground. On August 4, a Spot Report (OSCE 2008c) informed the OSCE participating states of exchanges of small arms fire and mortar shelling in what the report described as the most serious outbreak of firing since 2004. The Spot Report (OSCE 2008d) on August 7 warned of significant Georgian troops and equipment moving in the direction of the zone of conflict. An international fact-finding mission on the conflict later documented the subsequent course of developments. On the night of August 7-8 Georgia's armed forces launched a large-scale military attack on Tskhinvali, the provincial capital of South Ossetia, "to restore constitutional order" and halt what the Georgian authorities had come to believe was an impending, or even actual, Russian invasion. However, the international fact-finding mission could not substantiate the claim of a Russian military incursion into South Ossetia prior to August 8 (Independent International Fact-Finding Mission on the Conflict in Georgia Report, 2009). Beyond doubt is that Russian troops quickly arrived and after three days of fighting had expelled Georgian troops from South Ossetia. Then, they proceeded beyond the conflict zones, deep into Georgian territory.
What seems clear was that the escalation of military activities on both sides indicated a prisoner's dilemma dynamic. The complete absence of trust on both sides meant that no side to the conflict took any steps to manage escalation by creating credible commitments. Instead, the dominating strategy was to fool the other. Consequently, the Georgian decision to militarily enter Tskhinvali proved a misread attempt to deter the enemy by striking first. It seems plausible to assume that the Georgian strategy was underpinned by an equally misread interpretation of the nature of its "special relationship" with the US.
Overall, the catalogue of activities and reporting mapped above demonstrate that the OSCE was able to vent early warning signs in the run-up to August 2008; however, "sufficient early action did not follow" (Lynch 2010b, 150).

Ukraine
In the immediate aftermath of the collapse of Ukrainian President Viktor Yanukovych government on 22 February 2014, tensions rose on Crimea over the future belonging of the Ukrainian peninsula. Concomitantly, Russian President Putin ordered a major snap readiness exercise of Russian troops in the country's Western and parts of its Central Military Districts. The drill involved 150,000 servicemen, 90 aircraft, 120 helicopters, 880 tanks, 1,200 pieces of equipment and 80 ships (Ministry of Defense of the Russian Federation 2014). The exercise focused on large-scale deployments far from Ukraine's borders, and seemed designed to provide a distraction from the operation to establish military control over Crimea, which began on February 27 when Russian special operations forces seized administrative buildings and installed a pro-Russian local government. The following day a full-scale invasion took place as Russian units spread out across the peninsula taking control of strategic objectives in a remarkably quick and wellorganized operation (Lavrov 2015, pp. 163-70). On March 16, Moscow backed a contested referendum in Crimea with an overwhelming majority of the voters allegedly supporting reunification with Russia. Soon after the illegal annexation of Crimea, Russia's military aggression against Ukraine turned to the eastern Donbas region; and a war that is yet to end broke out in early April 2014.
Thus, events in Ukraine in 2014 moved rapidly from a political crisis to a major armed conflict in Europe within a few months. In that period, several signatory nations used the Vienna Document, the Open Skies Treaty and the CFE Treaty to monitor the situation in and around Ukraine. On February 26, Russia notified snap exercises in Pskov and Rostov-on-Don. On March 1, while the seizure of Crimea was ongoing, a Latvian-led multilateral team arrived in Pskov within the framework of the Vienna Document to inspect the whereabouts of the 76th air assault division only to find an empty field. The military escort's explanation for the division's absence was that "they were all playing sport" (Interview military expert, 5 February 2019). Six months later, President Putin awarded the 76th division the Order of Suvorov for the "successful completion of military missions" (Presidential Decree 2014). Two days later, Ukrainian forces captured two infantry fighting vehicles belonging to the just decorated division in Luhansk (Fokus 2014).
As events in Crimea quickly unfolded, discussions in the OSCE grappled with how to respond to the Russian occupation of the peninsula and what tools to use in order to hold Russia accountable for its actions. Eventually, participating states' efforts coalesced around applying the Vienna Document. In accordance with Paragraph 18 of Chapter III, Ukraine reported security concerns about unusual Russian military activities and asked Russia to allow voluntary hosting of visits to dispel these concerns. Russia refused. In a creative interpretation of the same paragraph, Ukraine then requested the OSCE to gather an observation team to investigate the military situation in Ukraine instead. Thus, since Russia would not allow inspections on its territory within the framework of Chapter III, the idea was to try to observe Russian military activities from Ukrainian territory (Interview diplomat, 10 December 2018).
The first verification visit began on March 5 when a group of up to 40 international inspectors landed in Odesa. From Odesa, the inspection team made four attempts to enter Crimea at three different checkpoints. On each occasion, masked uniformed personnel refused them access at gunpoint (Interview military expert, 29 November 2018). Even if the attempt to visit Crimea proved unsuccessful, the verification visit, by default, enabled the OSCE to provide evidence that Crimea was no longer under the control of the Ukrainian government (Interview military expert 23 November 2018).
In addition, the mechanism for consultation -Paragraph 16 of Chapter III -was invoked 18 times in 2014 by Ukraine and several other OSCE participating states in order to address concerns regarding unusual military activities conducted by Russian troops along the Ukrainian border. According to NATO estimates, Russia had amassed 20,-000-40,000 soldiers at the border without advance notification. Eventually, three joint meetings of the FSC and the PC took place in April 2014 (OSCE 2015;Schmitt 2018, 280). However, Russia refrained from participating in any of the meetings. This led the delegations of Ukraine and several western states to condemn Russia for being uncooperative and not following the spirit of the Vienna Document (United States Mission to the United States Mission to the OSCE 2014). Moscow in turn accused its critics of misusing the Vienna Document to "stir up a scandal" (Ministry of Foreign Affairs of the Russian Federation 2014).
After the verification visits described above, enabled by the unconventional interpretation of Chapter III, smaller nation-led multilateral inspection teams were deployed to Ukraine in accordance with the Vienna Document's Chapter IX on Compliance and Verification and Chapter X on Regional Measures. To enable the observation of developments over time, NATO and partners covered Ukraine with inspection visits from March to June 2014 (Interviews with military experts, 23 November 2018 and 5 February 2019). 3 In one of these inspections, separatists took a German-led team of seven inspectors' hostage in the eastern Ukrainian town of Slovyansk from April 25 to May 3 (Sonne 2014). In 2014 alone, 17 verification visits took place in Ukraine and a further six visits in Russia (OSCE 2015). Even though these inspections did not provide clear-cut evidence of Russian troop activities in Ukraine, they may have had the effect of raising the threshold for direct Russian intervention. At the very least, the hostage-taking incident aimed at scaring away the inspectors would indicate that these measures were a source of irritation to both Russia and the separatists it was supporting.
The Open Skies Treaty provided the framework for two different types of observation flights in response to the Ukraine crisis. The first type of flights, granted under Annex L of the Treaty on Open Skies (1992, p. 96), permits states "on a bilateral and voluntary basis to conduct observation flights over the territory of each other". Referencing this Annex, Sweden carried out an observational flight over Ukraine on March 13 for the first time ever, while the United States followed up with an observation flight over Ukraine on March 14. Due to the security situation, none of the flights covered Crimea. A week later Ukraine conducted an observation flight over Russia within the same procedural framework. In addition, the regular Open Skies quota flights distributed on an annual basis were used to try to monitor military movements in both Russia and Ukraine (Rowberry 2014). After a missile shot down Malaysian Airlines flight MH-17 over eastern Ukraine on 17 July 2014, however, Open Skies overflights were suspended for safety reasons (Bieri and Nünlist 2018, 419). In Crimea, the Russian annexation complicated the situation still further. Although Moscow invited Open Skies flights over Crimea after March 2014, western states did not want to implicitly recognize Russia's land grab and therefore refrained from such missions (Bieri and Nünlist 2018, 419). 4 While the CFE instrument was unavailable for collecting information on Russian military forces, its mechanisms for information exchange and verification provided NATO and its partners with insights on aspects such as the location of Ukrainian equipment, troop morale and training (Interview military expert, 5 February 2019).
On 21 March 2014, the OSCE's PC decided, with the consent of all OSCE participating states, to send a Special Monitoring Mission (SMM) to Ukraine to monitor the crisis (Neukrich 2016, pp. 229-39). The OSCE deployed the first monitors in less than 24 hours. The quick deployment betrayed a delicate negotiation process. Indeed, for a while it appeared unlikely that the OSCE participating states would agree on the mission at all. Facing the risk of a stalled process, some EU member states in a parallel track pushed for an EU mission. On March 20, the day before the OSCE participating states agreed to the SMM, the European Council turned the heat on Russia by putting the option of an EU observer mission on the table: "In the absence of an OSCE mission in the coming days, the European Union will launch an EU observer mission" (European Council 2014). Moscow, fearing an EU mission, responded the next day by accepting the OSCE SMM, the size, geographical scope and management of which it had managed to constrain during the negotiations (Interview diplomat, 10 December 2018). For example, Crimea -the intended geographic area of operation -was excluded altogether due to Russia's refusal. In addition, the mission had been intended to deploy military observers, but Russian objections meant that it ended up with civilian monitors (Interview diplomat, 10 December 2018).

Lessons to learn
What lessons can be drawn from the events unfolding in Georgia in 2008 and Ukraine in 2014? On the basis of the preceding account of the armed conflicts, four principal aspects merit a closer discussion: the role of military exercises; the use of the inspection instrument; the implications of Russia's suspension of the CFE Treaty; and an analysis of the realistic role of arms control in the various points of a conflict cycle.

Military exercises and conflict escalation
On July 15 2008, Russia began a major military exercise, Kavkaz 2008, next to the border with Georgia with the official participation of 8,000 troops. The scenario of the exercise included counterterrorism operations and preparation for special peacekeeping operations, such as supporting Russian peacekeepers in Abkhazia and South Ossetia. Prominent Russian and western analysts saw Kavkaz 2008 as a rehearsal for a military operation in the region (Felgenhauer 2008;Shanker 2008). Even at the time of the exercise, the former commander of the Group of Russian Forces in the Transcaucasus, retired Lieutenant General Yuri Netkachev, said that the official figures were underestimated in order to avoid the possibility of international observers (Nezavisimaya gazeta 2008). In Tbilisi, the suspicion was that the exercise was intended to conceal Russian mobilization along the Georgian border. 5 While the exercise had ended before the war broke out, some, Russian troops and equipment remained close to the Georgian border, enabling Russian forces to quickly advance into Georgia through the Roki tunnel once the order was given (Independent International Fact-Finding Mission 2009, 217).
The Vienna Document stipulates that military activities exceeding certain thresholds require a 42-day notice. However, exercises carried out without prior notification to the troops involved are exempted from this rule. If such snap exercises exceed the thresholds, notification need to be given when they begin (Vienna Document, paragraph 41.1). Such exercises are exempted from the need to invite observers unless they exceed 72 hours of duration (Vienna Document, paragraph 58). For some time, Russia has been criticized for conducting large-scale snap exercises and reports them as falling below the Vienna Document's threshold of 13,000 troops (van Ham 2018, 17). Moreover, the inspection mechanism does not apply to de facto large-scale exercises formally reported as smallscale and under separate commands. Russia has taken advantage of this by presenting large-scale exercises west of the Urals as a series of smaller exercises that did not meet the threshold for evaluation. In the run-up to the annexation of Crimea, Russia formally compartmentalized the large readiness inspection in February 2014 as a sequence of snap exercises (Gressel 2018, 6). Sutyagin and Bronk (2017, 13-4) found two organizational aspects noteworthy. First, to avoid the notification requirements of the Vienna Document, Moscow tailored troop deployments to keep them under the threshold of 9,000 ground troops in any single zone. Second, the Russian armed forces organized troops as battalion tactical groups that appeared to operate independently of each other, and not to be under central command, thereby allowing Russia to bypass Paragraph 40, Chapter V of the Vienna Document. Thus, Russia complied with the letter, if not the spirit, of the document (Interview military expert, 29 November 2018).
In sum, in 2008 and again in 2014, Russian military exercises near the borders with Georgia and Ukraine preceded the actual military operation. These exercises were no coincidence, and served as both diversions from, and support for, the subsequent operations. Kavkaz 2008 not only had an escalatory effect on the Georgian government but also pre-positioned troops and equipment for the military invasion. In Ukraine, the sequence of snap readiness exercises in the spring of 2014 provided cover for the simultaneous mobilization for the Crimea operation. These exercises kept Georgia and Ukraine guessing when the Russian military would "shift from exercises to preparation for war" (Gressel 2018, 6).

Limits of the inspection instrument
While the Vienna Document's Chapter III mechanism on unusual military activities was activated in the build-up to the Russian-Georgian war, the document contains other mechanisms that were not used in the summer of 2008. Chapter IX stipulates quotas for military inspections. A small number of inspections are available annually of every country. For Russia, the annual quota is three inspections and two evaluation visits. The low level of the quota typically leads to a race for inspections at the beginning of the year (Richter 2010). By the time Russia's military activities in and around Georgia intensified in 2008, the annual quota for Russia had already been exhausted. Similarly, by April 2008 participating states had already used all three inspections for Georgia (Schmidt 2009, 22). Unlike the Vienna Document, which had run out of inspections for Georgia, one inspection under the CFE Treaty and three Open Skies overflights were still available but not used before the outbreak of war.
In 2014, states had already exhausted the annual quota of inspections for Russia before the end of March. Two military inspections, conducted by Latvia in the Pskov region and Switzerland in the Moscow region, took place around the time of the annexation of Crimea, while the final one took place in Belgorod and Kursk, near the Russian-Ukrainian border, on March 19-20. While some OSCE participating states were critical of the unusual application of the Vienna Document, notably the repeated inspection visits from March to June 2014, others saw this as a way of demonstrating that Ukraine had lost control of Crimea as well as a political statement against Russia's aggression, and thus as raising the political costs of further Russian action (Interviews with military experts, 23 November 2018 and 5 February 2019). Irrespective of position, there appears to be a consensual understanding that the recurring verification visits did not reveal any Russian activities in Ukraine (Interviews with military experts, November 23 and 29 November 2018). The lack of substantive information produced by these verification visits in combination with the SMM being up and running eventually led to a reduction in the frequency of inspections after June 2014 (Interview military expert, 29 November 2018). Overall, CAC and CSBMs should not be conflated with intelligence gathering. Rather than sharpening the intelligence gathering capabilities, the main advantage lies in the transparency-fostering function. Transparent limitations, information sharing and verification can provide a legitimate foundation for public discussion and, ideally, for demanding accountability.
The possibility of detecting military build-ups with the use of Open Skies observation flights, which was used as a kind of inspection instrument, is also questionable. First, and most importantly, the intention with Open Skies was never to spot tanks in the forest, but it can be used to support crisis management. Second, their procedural underpinnings also set limits. In preparation for each observation flight, the observing party is required to send a message notifying the observed state no less than 72 hours before the estimated time of arrival. The observed party is obliged to acknowledge the notification within 24 hours. On arrival, the observed party then conducts a pre-flight inspection of the aircraft (Treaty on Open Skies 1992, 11-2). Taken together, this procedure provides the observed state with significant time to warn military units that may be conducting unusual activities (Interview military expert, 23 November 2018). These types of verification activities have also become more challenging because force sizes have shrunk dramatically since the early 1990 s.

What if Russia had not suspended the CFE Treaty and the west had adapted the adapted CFE?
The purpose of the legally binding CFE Treaty is to prevent large-scale military mobilization and massed armored attack. In comparison with the restrictive inspection quotas of the Vienna Document, which also does not contain an intrusive verification component, the CFE treaty entails a much stricter verification mechanism with different types of onsite inspections. When Russia suspended its implementation of the CFE Treaty in December 2007, all limits on Russian military deployments along its flanks were removed. Since then, Russia has provided no notification of troop movements and no information on changes in the location of TLE. Nor has it accepted any intrusive verification measures (McCausland 2008;Felgenhauer 2015). This facilitated preparations for the war in Georgia as well as the annexation of Crimea.
Released from the CFE Treaty limitations, the Russian military intensified its build up troops and armaments in the North Caucasus and also deployed heavy military equipment to Abkhazia in the spring of 2008. However, even before it suspended implementation of the CFE Treaty, Russia had allegedly stockpiled TLE in Abkhazia and maintained a garrison at the Gudauta base in breach of treaty stipulations (Socor 2008). Nonetheless, at least in theory, CFE inspections could have provided advance notice of Russia's military build-up in Georgia and warned of possible early preparations for military actions that prefigured the annexation of Crimea (Felgenhauer 2015). If nothing else, Russia's suspension of the CFE Treaty represented a form of political warning in itself.
As it was, Russia's suspension of the CFE Treaty left the international community with no possibility of activating intrusive verification measures, such as calling for a challenge inspection, to determine the location of Russia's TLE. If Russia had not suspended the treaty, this intrusive inspection mechanism would have increased the chances of other signatory states providing early warning of Russia's military deployments. Left with only the Vienna Document, the force deployments prior to the August war did not even require invitations to international observers.
In 2014, Russia moved troops near the border with Ukraine enabling its armed forces to rapidly enter Ukraine. If Russia were still participating in the exchanges of information stipulated in the treaty, it would have been subjected to 25 inspections, in addition to the special challenge inspection, and 100% accountability over the location of every piece of TLE. Taken together, this would have made its operation, and the transfer of military equipment to separatists in Ukraine, much harder to hide (Interview military expert, 5 February 2019).
Then, there is the question of what a ratified Adapted CFE could have done to alleviate situation. The Adapted CFE aimed to address the post-Cold War geostrategic changes and replace the alliance ceilings with national and territorial limitations for each state party. In particular, it intended to prevent destabilizing sub-regional concentration of forces. By ratifying the Adapted CFE in 2004, Russia accepted limitations in the flank areas, including the north Caucasus. Thus, if ratified, the Adapted CFE would have made it more difficult for Russia to move forces in border areas. Moreover, the verification provisions of a ratified Adapted CFE Treaty would have provided for an increase of inspection quota by one third.
That said, none of the instruments, whether legally binding or politically binding, are designed to withstand intentional warfare. Verification is not an end in itself but serves the purpose to ensure compliance with the objectives and provisions of the treaty. However, when parties chose to implement the agreements as they see fit, the stabilizing function of CAC and CSBMs is severely undermined.

Arms control and the conflict cycle
The experiences from Georgia and Ukraine identify three distinct points in the conflict cycle: the initial conflict situation, a hot war and an unresolved (frozen) conflict. In this cycle, the tasks evolve from conflict prevention via conflict management to conflict resolution. The expectations on CAC and CSBM regimes differ depending on the phase in which the instruments are used.
First, by activating various tools, such as fact-finding missions, monitoring missions and consultations and inspections within the framework of the Vienna Document, the OSCE provided early warning on the conflict escalation in Georgia in 2008 and in Donbas in 2014. In Georgia, OSCE observers were instrumental in raising international awareness of the conflict both within the OSCE and in other multilateral institutions (Interview military expert, 19 December 2018). These warnings also spurred several diplomatic initiatives, including the ambassadorial visit a month before the war. Nonetheless, the OSCE failed to adapt policies in response to the emerging escalatory logic documented by observers on the ground. In fact, as the reports issued by OSCE observers in early August were documenting a serious escalation, crisis diplomacy petered out, suggesting that the reports failed to register in Vienna (Stöber 2011, pp. 203, 216). In short, in a consensual organization like the OSCE, early action ultimately boils down to the measures that participating states manage to agree on (Interview military expert, 19 December 2018).
In 2014, the OSCE demonstrated mixed abilities to react. Since there was no organized international presence on the ground in Crimea, the OSCE and other international organizations were, as one close observer notes, "blind and unable to engage in early crisis response". Furthermore, the sheer speed of the military operation meant that the international community "was soon confronted with a fait accompli" (Tanner 2016, 242). When hostilities moved to eastern Ukraine, several participating states attempted conflict prevention measures, not least by activating a series of inspection visits under the Vienna Document, and crisis management initiatives, such as the SMM. However, the stabilizing measures deployed did not prevent the outbreak of outright war in eastern Ukraine in April 2014. This was not necessarily due to flaws in the CAC and CSBMs. As one expert notes, the tools worked in the sense that there was "more than enough information to act -if we wanted to. It is all about political will" (Interview military expert, 5 February 2019). Thus, overall, the major problem in 2008 and 2014 was that onsite early warning failed to generate a sufficiently strong political response.
Second, during a hot war, the experiences from Georgia and Ukraine underline the conventional wisdom that CAC regimes are largely obsolete by the time a conflict has escalated to a regular war. By that point, the conflict parties are unwilling to disclose information on heavy weapons and therefore unlikely to submit themselves to transparency measures. If intrusive inspections are unlikely to be allowed by the conflict parties in the war phase, the policy of short visits stipulated in for example the Vienna Document is also insufficient at this stage. In Ukraine, the OSCE acted on this realization by deploying a novel initiative in the form of the SMM to ensure a permanent presence on the ground that at least in theory could monitor escalation and have a constraining effect on the conflict. In other words, at a certain point in the escalation ladder inspections are unlikely to add value and should therefore give way to alternative crisis management measures. On the one hand, this calls for flexibility and adaptability among outside actors. On the other hand, the need for flexibility also creates room for wavering.
In both conflicts, Russia deployed what is usually referred to as "hybrid warfare" tactics, such as the use of cyberattacks and the spread of mass disinformation and propaganda along with military warfare (Nilsson 2018). In Georgia, Russia justified its involvement partly in terms of a humanitarian campaign aimed at stopping an ongoing Georgian genocide in South Ossetia, and partly as part of its peacekeeping mission in the region. None of these justifications withstood scrutiny ("Independent International Fact-Finding Mission 2009, pp. 188-90, 422). Similarly, the annexation of Crimea was justified by the need to defend the Russian-speaking population on the peninsula from a purportedly "fascist regime" in Kyiv (Presidential Administration of the Russian Federation 2014). Warfare in Ukraine initially relied on soldiers without insignia, which granted the Russian government a certain degree of deniability regarding the involvement of Russian forces in Crimea and Donbas. From the outset, the Russian leadership branded the conflict in Ukraine as a civil war and, despite convincing proof to the contrary, still insists to this day that Russian soldiers involved in the conflict are volunteers and not following orders from Moscow. Thus, the use of hybrid warfare, use of proxy forces and blurred distinction between foreign troops and local insurgents had not been foreseen when existing arms control instruments were conceived. It is unrealistic to envision that they could ever handle such complex realities, even under an entirely new type of regime.
Finally, in OSCE circles, some experts have introduced the concept of status-neutral security and arms control regimes for conflict regions with unclear international status, such as Abkhazia, South Ossetia and now Crimea, and possibly Donbas if that conflict would eventually freeze. By applying arms control instruments, the idea is to address the lack of military transparency in territories plagued by protracted conflicts. Since agreements on CAC and CSBMs are non-functioning in these conflict regions, there is a high degree of uncertainty about the activities and capacities of the military forces stationed in these territories. Developing special CAC and CSBMs that are applicable to conflict regions, irrespective of their political status, therefore, represents one potential way out of this dilemma, according to some advocates (Kapanadze et al. 2016, 8-9). However, those skeptical of this depoliticized status-neutral approach to so-called frozen conflicts argue that CAC and CSBMs cannot be developed in a manner that directly or indirectly legitimizes Russia's violations of neighboring countries' sovereign right to territorial integrity and their right to make their own security policy choices freely (Persson, ed. 2018). Classical arms control instruments may serve to build political-military trust and transparency in frozen conflicts, but only as part of a post-conflict rehabilitation process. Otherwise, such measures are likely to freeze a conflict further without contributing to its resolution. Ultimately, the sticking point for contested territories is related to how much exchange the international community can have with Abkhazia and South Ossetia without de facto recognizing them as independent political entities, and in the case of Crimea without accepting Russia's annexation.

Conclusion
What realistic role can CAC and CSBMs play in a military crisis? It should be emphasized that the effectiveness of the instruments depends on the nature of the conflict. In a local conflict that might erupt due to mistakes on the ground, CAC and CSBMs have a potentially crucial role to play in alleviating tensions by providing transparency and predictability. In a conflict that is escalating due to careless political leadership, CAC and CSBMs could play a calming role by providing the tools for clarifying the military situation. However, in the case of leaders willfully seeking an armed conflict, there is not much help to be found from CAC and CSBMs in terms of conflict prevention. When one or several parties deliberately seek war, they tend to undertake measures to dilute compliance, to diminish transparency and to conceal intent through disinformation. In short, international agreements function well when the parties seek to avoid unintentional conflict, but are non-functioning when one side seeks conflict.
Concerning the realistic role of the OSCE, it possesses political-military tools to provide early warning signals on which its participating states and others can act. There are, however, several thresholds to pass. The first concerns how to enable early warning initiatives to begin with, especially when a participating state is the aggressor in the conflict. The second concerns how to move from early warning to early action. Again, the main obstacle is that in a consensus-based organization, one particular state can successfully block any attempt at early action. For the future effectiveness of OSCE early crisis management, it would seem imperative to establish a more robust link between early warning and early action. One potential way forward would be to devise methods to facilitate the OSCE's ability to act without the need for consensus. A minimum level of ambition should be to allocate special budgetary funds for fact-finding missions, the deployment of special envoys, shuttle diplomacy and so on, but also to increase the possibility of quick and effective leadership. That said, even if such measures may provide additional possibilities, they cannot substitute for the actual political will to act.
The experiences from Georgia and Ukraine demonstrate that the Vienna Document, as the main tool used, has its limits with regard to providing information and transparency in an environment of mistrust, conflict escalation and hostility. These experiences indicate the need to remove certain ambiguities from the document. This would help to strengthen the capacity of the OSCE in general and the Vienna Document in particular to provide the kind of transparent informational basis for dialogue and decision-making required.
First, in its present configuration, the Vienna Document is ill-equipped to handle the smaller armies and different force structures of the 21st century. Since the end of the Cold War, military deployments have changed from large tank armies to smaller, more readily deployable units. The thresholds in Chapter V of the Vienna Document for prior announcement and observation of military exercises, however, still reflect the larger composition of military exercises that was common in the past. For example, lower thresholds would at least have made it more difficult for Russia to avoid inspection of its military exercise Kavkaz 2008. Moreover, since 2013, Russia has systematically, including in conjunction with the Crimea operation, used snap exercises to avoid international scrutiny. One way to address this would be to clearly specify how participating states must provide notification about snap exercises. Removing the complementary provision that exempts from inspection exercises not presented as under single command would also be a way of strengthening the inspection regime and the ability to gather early warning information.
Second, the current low inspection quotas mean that participating states tend to have exhausted their quotas early in the year, thereby nullifying this tool in case of military escalations later in the year. Hence, the Vienna Document's relatively high threshold for information and observation in combination with the low quota for onsite inspections prevented any effective use of verification activities at the time of crisis escalation in 2008 and 2014. Lowering the threshold for information and observation and expanding quotas would therefore represent concrete measures that could have had a bearing on the conflicts in Georgia and Ukraine.
Third, to sharpen the Vienna Document's responsiveness to crisis escalation, another plausible idea would be to allow for an additional, or bonus, inspection mechanism. One idea might be to expand Paragraph 16 of Chapter III to include an inspection mechanism as a last resort in the consultative mechanism. This consultation mechanism allows a state to request an explanation for military activities that they perceive to be a security concern. The other party must provide a response within 48 hours. If the statement fails to bring clarity to the issue, the requesting state may then call for a meeting with the responding state within a further 48 hours. If still unresolved, a third step is to take the dispute to a joint FSC/PC meeting with all OSCE participating states within 48 hours. This mechanism was activated in both 2008 and 2014 without making much headway in defusing the tensions. An additional fourth step could therefore be to activate an obligatory inspection to address the security concerns. Given the difficulty of unanimously agreeing on such a measure, an exploration of the possibilities of activating this mechanism without consensual approval would appear to be necessary.
To sum up, under current security conditions, the prospects for a wholesale reform of the existing system of CAC and CSBMs appear slim. Lessons from the experiences in 2008 and 2014, however, suggest that the OSCE participating states could improve the effectiveness of the Vienna Document by undertaking a few concrete modernization measures in line with the themes suggested above, all of which have in some form already been placed on the table in Vienna. That said, while closing the gaps and shortcomings in the Vienna Document, or any other agreement, is desirable, it does not provide a remedy for conflict prevention. The best to hope for is to strengthen indicators, increase warning times and raise the political costs of cheating. The rest, such as converting early warning into action, will require political will.