Trade as a confidence-building measure in protracted conflicts: the cases of Georgia and Moldova compared

Academic and policy literature on the post-Soviet region’s protracted conflicts frequently assumes that trade is an effective confidence-building measure (CBM) and as such also has the potential to facilitate progress toward conflict settlement. We probe these assertions through a detailed empirical comparison of the effects that regulated and unregulated forms of trade have had in the cases of Abkhazia, South Ossetia, and Transnistria. Relying on a wealth of primary and secondary sources, including 29 semi-structured interviews, as well as direct and participant observation, we find that, absent political will by authorities on the ground and their respective international partners, trade remains by-and-large unregulated and has limited and generally not sustainable confidence-building and stabilizing effects, as evidenced by the cases of Abkhazia and South Ossetia. Where local and geopolitical conditions are more conducive, trade becomes more regulated and significantly contributes to confidence building and stabilization and has some positive spill-over effects into non-economic areas of confidence building, as is evident from Transnistria. This finding has potentially important implications for other cases in which governments and relevant regional and international actors are faced with difficult choices over whether and how to engage with unrecognized entities, such as in Ukraine. ARTICLE HISTORY Received 23 September 2019 Accepted 5 December 2019


Introduction
Do confidence-building measures (CBMs) facilitate conflict settlement in the post-Soviet space? This is a pertinent question for three reasons: the connection between CBMs and conflict settlement is frequently asserted and confidence building remains at the core of the initiatives of many regional and international organizations in the area, such as the OSCE, EU, UN(DP) and their participating and member states. Yet, the lack of any substantial progress in resolving longstanding protracted conflicts in this area appear to indicate the limited impact of CBMs on conflict settlement. At best, CBMs have stabilized, but arguably also entrenched, an existing status quo. This does not render CBMs useless or counter-productive but requires a re-assessment of what they can realistically contribute to the management of protracted conflicts.
CBMs have a long history in East-West relations during the Cold War. Economic engagement, one of the three baskets of the 1975 Helsinki Final Act, is one avenue that can be explored in the search for measures through which confidence can be built among parties. It can be viewed as a first step on a path toward a settlement as opposed to a temporary "fix" of an acute problem. Trade features prominently in the so-called second basket of the Helsinki Final Act, entitled "Co-operation in the Field of Economics, of Science and Technology and of the Environment". In particular, the participating states recognize "that trade represents an essential sector of their co-operation" and acknowledge "the importance of the contribution which an improvement of business contacts, and the accompanying growth of confidence in business relationships, could make to the development of commercial and economic relations" (Conference on Security and Co-operation Europe 1975). Trade is, thus, one area in which self-interested elites can see opportunities for constructive engagement and which offers the international community an opportunity to contribute not only to confidence-building and improving the livelihoods of the people affected by conflicts but also to building a foundation upon which comprehensive conflict settlement in the future might be possible.
Examining past and present efforts of trade facilitation in three of the protracted conflicts in the post-Soviet space -Transnistria, Abkhazia, and South Ossetiaour main aim is to investigate when and how trade contributes to meaningful engagement between the conflict parties that can pave the way toward conflict settlement. We follow Keshk, Pollins, andReuveny (2004, 1176) who have proposed that the scholarly debate on trade should be framed in the context of "when" and "how" trade promotes peace rather than "whether" it does so.
In the next section, we first define what CBMs entail and briefly summarize existing research on the nexus of trade, confidence building, and conflict settlement. We then outline our approach to operationalizing the question when and how CBMs facilitate conflict settlement in protracted conflicts, before turning to an analysis of trade-focused confidence-building initiatives in the cases of Georgia-Abkhazia, Georgia-South Ossetia, and Moldova-Transnistria. In the final section, we offer a summary and interpretation of our findings based on a comparative assessment of what accounts for the successes and/or failures of these initiatives.
Conceptual background: confidence building, conflict settlement, and trade Cooperation in Europe 2012, 9) as "sales people are usually reluctant to fight their customers" (World Trade Organization 2011, 1).
This sentiment is echoed in democratic peace theory, which posits that liberal democracies are less likely to engage in military confrontations with one another (Doyle 1983). In a similar manner, proponents of the interdependenceconflict theory have emphasized the pacifying effects of trade interdependence: the broader, deeper, and more intense inter-state and inter-community trade relations are, the lower is the likelihood of these states and communities fighting each other (Oneal 2003). More recently, Lee and Rider (2016) have re-tested the 'trade-reduces-conflict' hypothesis by focusing on a particular type of conflict, namely conflict over high-salience issues such as territory, and found, in line with the established liberal peace consensus, that trade "interdependence decreases the incentives for trading partners to engage in militarised conflict" (Lee and Rider 2016, 2, 4f.). The few studies that have questioned that trade interdependence reduces the likelihood of conflict (Barbieri 1996; Barbieri and Levy 1999;Keshk, Pollins, and Reuveny 2004) have been counter-challenged effectively (Gartzke and Li 2003;Xiang, Xiaohong, and Keteku 2007;Hegre, Oneal, and Russett 2010).
We thus have robust evidence of trade interdependence being negatively correlated with the onset of interstate conflicts. Equally, a substantial body of literature has established the utility of confidence building as a means to progress toward tension reduction and conflict settlement. Yet, many of the assertions in the policy literature that trade can promote confidence and thus pave a way toward conflict settlement in intrastate conflicts (Mirimanova 2006(Mirimanova , 2017 International Crisis Group 2018) lack empirical backing for such claims. While individual cases offer some evidence of a plausible connection between the promotion of trade links, enhanced confidence between the conflict parties, and reduced tensions (Krepon 2000;Kyris 2014;Pandian 2005;Singh 2009;Rolandsen 2019;Mirimanova et al. 2015), no systematic assessment of the effectiveness of trade as a confidence-building measure in relation to the protracted conflicts in the post-Soviet space exists, and we cannot simply assume that trade, especially in the context of the heightened complexity of intrastate conflicts involving unrecognized entities, has similarly pacifying effects.
Our contribution seeks to address this knowledge gap by focusing specifically on trade as a confidence-building measure in two post-Soviet countries and the protracted separatist conflicts that they have experienced for the past quarter of a century. While we are generally interested in the question whether (the promotion of) trade contributes to increasing confidence between the sides and whether this enhances the prospects of conflict settlement, we need to be realistic in our ambitions. With none of the three conflicts in Abkhazia, South Ossetia, and Transnistria actually resolved, framing our contribution too broadly would make our analysis redundant or at best highly speculative. While we are, thus, not able to provide a definitive answer to these more general questions, our comparative approach within and across these three cases offers important insights into the conditions under which trade, and what kind of traderegulated or unregulatedis likely to build confidence and whether such confidence in economic relations is likely to spill over into more political, and status-related, issues.
This distinction between regulated and unregulated trade is a critical one for the cases we examine, and beyond them. The so-called de-facto or unrecognized states that are one of the actors in these conflicts have frequently been associated with large-and small-scale smuggling as one of the principle economic lifelines of regimes that have otherwise few options of revenue generation, which has been documented for both the post-Soviet cases (e.g., Chamberlain-Creanga and Allin 2010;International Crisis Group 2006;Istomin and Bolgova 2016;Kolstø 2006) and elsewhere (e.g., Hafeez 2014; Mylonas and Ahram 2015;Rolandsen 2019). At the same time, efforts have been under way over the years to regulate these trade flows with a view to curb organized criminal activity and to engage with business elites as a potential peacebuilding constituency without threatening the economic survival of the entities in question (e.g., Axyonova et al. 2018;Całus 2016b;Mirimanova 2006;Popescu and Litra 2012). While one could argue that even unregulated trade can have confidence-building effects and thus contribute at least to tension reduction and de-escalation (Remler 2016;Wolff, Remler, and Davies 2017), from the perspective of longer-term conflict settlement, regulated trade would appear as more promising and hence CBMs that seek to promote regulation in trade relationships should be preferable (Douglas and Wolff 2018). However, the more regulated trade relations become, the more the specter of potential recognition rises which poses difficulties for the parent states of unrecognized entities and their international partners (de Waal 2018;Ker-Lindsay and Berg 2018;Bouris and Kyris 2017), which in turn may stabilize a conflict situation on the ground, yet entrench a status quo in which conflict settlement becomes ever less important as the stable co-existence between an unrecognized entity and its parent state benefits multiple constituencies (Beyer 2010;Beyer and Wolff 2016;Wolff 2012).
In addressing the question when and how regulated and unregulated trade contributes to meaningful engagement between the conflict parties that can pave the way toward conflict settlement, we use three case studies as our primary empirical basis for the analysis that follows. Choosing the conflicts in Moldova (Transnistria) and Georgia (Abkhazia and South Ossetia) offers us a rich basis of empirical data given that we can look back at more than 25 years of attempts to build confidence between the conflict parties, including through trade. Yet, as noted above none of the three cases have seen substantial progress toward conflict settlement. While this requires us to scale down our ambition to understand the contribution of economic confidence building to conflict settlement, it also represents an opportunity to focus on the distinction between regulated and unregulated forms of trade and their respective effects. Across our three cases, we find sufficient variation in the relationship between these forms of trade and the levels of confidence being built and sustained, as well as their stabilizing, violence-reducing effects. This adds important nuance to an established, yet inconclusive debate on how to engage with unrecognized states, for what purpose and with what acceptable trade-offs. As such, our analysis and findings have importance beyond the three cases we study here. Ukraine currently faces important choices about whether and how to engage with the non-government-controlled areas of Donbas, and further afield governments, as well as regional and global actors, face similar dilemmas of engagement from Western Sahara to Somaliland, from Palestine to Kosovo and to Taiwan. While we do not claim that our findings here are transferable to any of these cases or that these cases do not have insights of their own to offer on these questions, our within-case and cross-case comparisons generate an important new hypothesis on the impact of regulated and unregulated trade on confidence building. Evidence from our analysis suggests that this hypothesis, and its underpinning causal mechanism, is highly plausible, but further research in other cases will be required before broader generalization is possible.
In terms of data, we rely on a wide range of sources and methods of data collection. We have followed events in all three cases closely for more than a decade, visited both Moldova and Georgia, and some of the conflict zones, on dozens of occasions, and worked with some of the institutions and organizations active in the field of confidence-building. We have thus built up significant and in-depth personal knowledge of each of the three cases. Additionally, we collected information from academic sources, reports by think tanks and NGOs, media reports, and statistical data from governmental and international organizations. We also conducted 29 interviews between 2016 and 2019 with government officials, representatives of international organizations, nongovernmental organizations, and individuals associated with the authorities in Abkhazia, South Ossetia, and Transnistria (see Appendix 1). We were thus able to triangulate our data between different sources and build a robust analytical narrative for each of our cases as a basis for the within-case and cross-case comparison upon which our study rests.

The case of Georgia
From the late 1980s onwards, separatist tensions built up in both Abkhazia and South Ossetia which turned into violent conflicts in the wake of the dissolution of the Soviet Union. Ceasefires achieved in both conflicts through Russian mediation (1992in South Ossetia, 1994 in Abkhazia) meant that Georgia effectively lost control over both regions (except for the Kodori gorge in Abkhazia). Until the Russia-Georgia war in August 2008, periods of relative calm were pierced by periodic violent escalation. At the end of August 2008, Russia formally recognized Abkhazia and South Ossetia as independent states.
Following the Rose Revolution in Georgia, which swept Mikheil Saakashvili to power, relations between Tbilisi and the two separatist regions had deteriorated from 2004 onwards, as had relations between Georgia and Russia. Consequently, "prospects for substantive progress in negotiations about status, security, or return of persons displaced by conflicts in the 1990s or the 2008 war" were considered "minimal" toward the end of Saakashvili's second term in office (Toal and Grono 2011, 656). Since the Georgian Dream coalition came to power in 2012, Tbilisi has maintained a more conciliatory policy toward its breakaway regions (as well as Russia). The previous government's relatively aggressive rhetoric against Russia was significantly tempered down, the use of the word "occupation" by senior government officials was minimized, and Georgia renewed trade talks with its northern neighbor. This policy change also involved an easing of the security measures on the so-called Administrative Boundary Lines (ABLs), including re-opening of all initial six legal crossing points at Georgian-Abkhaz ABL in December 2012 as part of an agreement reached at the Geneva International Discussions (Interview 25).

Georgian-Abkhaz relations
Throughout the 1990s, contraband trade across the main transit route of the Enguri/Ingur river grew significantly, and according to some analysts (e.g., Kukhianidze, Kupatadze, and Gotsiridze 2004;Mirimanova 2006) reached especially high levels in the period of 1998-2003 (which followed the deterioration of the security situation in the region and the escalation of violence in May 1998). According to some estimates, the turnover of goods between the sides had reached $10-15m annually by 2006 (Baratelia 2011, 14). Moreover, although initially mainly ethnic Georgians from Gali participated in this transborder trade (Kolossov and O'Loughlin 2011), by 2005-6 some ethnic Abkhaz also traveled to Georgia to trade (Baratelia 2011, 14).
In the aftermath of the crisis in summer 2006, when violence again significantly escalated, ABL crossing points were closed and the turnover of goods between the two sides reportedly shrunk more than ten-fold (Baratelia 2011, 14). This had a negative impact on cross-border inter-community trade relations, especially in light of the slow but steady progress that had been made in the region over the preceding years.
In 2007, Sergei Bagapsh, then de facto President of Abkhazia, signed a decree that any movement of commercial goods to and from the Georgian side was illegal (Charaia 2015). The ban was partially lifted in 2015 when Bagapsh's successor, Raul Khajimba, authorized hazelnut exports (International Crisis Group 2018). Georgia, too, considers almost all economic activities, including trade, on the territory of both Abkhazia and South Ossetia illegal under the terms of the Law of Georgia on the Occupied Territories, but also views the movement of goods on the territory of Abkhazia as "domestic turnover" within Georgia (Charaia 2015).
These official positions notwithstanding, unregulated, informal trade (i.e., smuggling) has continued. In 2013 this was estimated at around $15m per year (Baratelia 2011;Mirimanova 2015). As one official from the European Union Monitoring Mission in Georgia (EUMM) observed in 2016, "if you were to stand by the crossing points at Enguri/Ingur river, you will see that there is always some small-scale trade going on" (Interview 03).
Thus, trade between the sides was affected by conflict in two ways: each major military crisis led to significant trade disruption (the original secessionist war in 1992/3, the 1998 crisis, and the 2008 war with Russia), and after each such crisis the authorities on both sides partly encouraged, partly tolerated the rebuilding of trade links. Even after the 2008 war, the level of regulated and unregulated (or formal and informal) trade recovered to pre-war levels. Given the recurrent violent crises since the 1992/3 secessionist war, whatever trade links existed between the sides did not prevent the outbreak of major violence. In light of the only limited progress on normalization between the sides after the 2008 war, let alone any sustainable settlement, the question arises what accounts for this repeated absence of sustainable confidence building through trade?
A 2006 report by International Alert highlighted the potential positive impact that local, Gali-based small-business entrepreneurs might have on bridging the gap between Abkhaz and Georgian communities and potentially on changing the conflict dynamics (Mirimanova 2006, 526-27). In their interviews, they mentioned the need for peace and trade to re-build relationships between the two sides: "We need to restore relationships between the Abkhaz and Georgians in Abkhazia. Nothing can do it better than joint business" (cited in Mirimanova 2006, 529). People on the ground, thus, seemed to echo the proponents of the "trade-causes-peace argument". As one of the respondents of the study put it at the time, "[p]eople who have got a business are neither angry nor aggressive. They want to expand their business, and another war would mean losing everything again" (cited in Mirimanova 2006, 529). Similar sentiments were highlighted also in surveys conducted in both Abkhazia and South Ossetia in 2010 (Kolossov and O'Loughlin 2011). More recent reports, too, found that Georgian IDPs from Abkhazia and Abkhaz living in Abkhazia and in Georgia-controlled territory "under certain security guarantees, . . . will be happy to engage in bilateral business development, especially in trade, small-scale agriculture and small-scale services" (Macharashvili 2018, 3, 26).
This local enthusiasm, however, has been curbed by developments in the international arena and Georgia's internal politics. Internationally, funding began to dry up as donors focused more on the crises in Afghanistan and Iraq (Klein and Pentikainen 2012) and bought into their own narrative of a stable status quo (Cohen 2012). This coincided with the "Rose Revolution" which brought political change to Georgia in more than just generational terms. The new government under Mikheil Saakashvili viewed Russia as the main obstacle to conflict settlement and pursued a policy of internationalizing Georgia's conflicts in order to secure more substantial Western support, framing the conflict as Georgian-Russian rather than Georgian-Abkhaz. However, this approach also led to a deterioration of the relations with the Abkhaz who perceived it as a strategy to further isolate them (Cohen 2012) and left little room for Georgian-Abkhaz dialogue. Combined with a deterioration in relations between Russia and the West and Western disengagement from the region, this meant that authorities in Tbilisi and Sokhumi/Sukhum had little interest in exploring any confidence-building effects of trade and viewed civil society initiatives as un-if not counter-productive and, therefore, less welcome (Cohen 2012; also Interview 07).
As the political and security situation around Abkhazia further deteriorated in the spring of 2008, the Georgia and its European partners made a fresh attempt at utilizing trade as a CBM. In March 2008, Saakashvili proposed to turn the Gali and Ochamchire districts in Abkhazia into "free economic zones" as part of his government's new peace plan for Abkhazia. Economic relations were also an important part of the second stage of the three-phase peace plan proposed by then German Foreign Minister Frank-Walter Steinmeier during his visits to Tbilisi, Sokhumi/Sukhum, and Moscow in July 2008. Both initiatives, however, collapsed in the wake of war between Georgia and Russia in August 2008 (Fuller 2008).
It took a decade before interest in the regulation of informal economic activities across the ABLs was renewed and before any serious consideration of the potential of trade for building and sustaining confidence was back on the agenda with the publication of "A Step to a Better Future"a Georgian peace initiative made public in April 2018. As a senior Georgian government official explained, there are 3 major dimensions to this peace initiative: to facilitate trade across the conflict divides, making education opportunities more accessible, and making different services more accessible to the residents of the occupied territories (Interview 25).
Under the trade facilitation dimension of this initiative, goods produced in Abkhazia and South Ossetia can be transported to "the Georgian-controlled territories, using status-neutral labeling". As the document states, "[i]n particular, it may be sufficient to include only the name of the producer, the name of the city/settlement (e.g. Sukhumi, Gagra, etc.) and/or address (street, number) on the product. . . . The name of the product and other information on the label may also be placed in the Abkhazian language" (Office of the State Minister of Georgia for Reconciliation and Civic Equality 2018, 14f.). Furthermore, the 2014 "Produced in Georgia" program includes the "Enterprise for a Better Future" subprogram which offers business development grants, including for cross-ABL projects (Interviews 24, 26). In addition, a Special Independent Fund was to be established in order to "mobilise financial resources and contributions from different donors and actors, international partners and private organisations, to serve the purpose of facilitating trade across the dividing lines" (Office of the State Minister of Georgia for Reconciliation and Civic Equality 2018, 14f.). These initiatives could pave the way to potential Abkhaz involvement in Georgia's Association Agreement (AA) with the EU and the related Deep and Comprehensive Free Trade Area (DCFTA), yet no tangible progress has been made in this direction to date (Interviews 23, 25, 27).

Georgian-Ossetian relations
As in Abkhazia, the seemingly "frozen" nature of the conflict led to both a lack of regulation of economic relations and a flourishing of unregulated, informal economic activity in the form of smuggling goods, drugs, weapons, and people for much of the 1990s and early 2000s (International Crisis Group 2004a, 9;Kartli 2017). Illegal trade was thriving because neither Georgia nor South Ossetia were willing to cooperate on regulating it under any legal framework. South Ossetia would not let Georgian customs officials on its territory and Georgia would not set up customs control checkpoints on their side of the ABL in order not to grant any legitimacy to the Ossetian claim of independence (International Crisis Group 2004a, 10). Moreover, authorities on both sides as well as Russian troops stationed in South Ossetia profited significantly from smuggling operations (Kukhianidze, Kupatadze, and Gotsiridze 2004;International Crisis Group 2004a), which further limited any incentives for formal trade regulation.
Located on a highway, which links the Georgian-controlled Gori to Tskhinvali in South Ossetia, Ergneti market was one such place where unregulated (or informal) trade was taking place. It gained particular prominence after 1999 when smuggling alcohol from Georgia to Russia became less profitable and instead people started smuggling cigarettes, petrol, stolen cars, agricultural and industrial products into Georgia via Ergneti, while Georgian agricultural produce would end up in Russia through the Ergneti route (Mirimanova 2006, 535). Since the goods were not taxed, prices for them were among some of the lowest in the region. Ergneti market quickly turned into "a price-setting mechanism for the entire region", providing a rather "reliable source of income for thousands of Ossetians and Georgians" (Mirimanova 2006, 535;Anonymous 2014). According to some estimates, annual turnover at Ergneti market was between $101-133m (Mirimanova 2006, 535;Kukhianidze, Kupatadze, and Gotsiridze 2004, 19) in the early 2000s.
Ergneti market was largely perceived as "an important development mechanism in South Ossetia, with significant peacebuilding implications" and even authorities in Tbilisi and Tskhinvali acknowledged at the time "that the organisation of Ergneti market had been their main achievement" (Mirimanova 2006, 535). As trust plays an important part in business, flourishing trade increased the quality of trust (Interview 20).
However, after the 2003 "Rose Revolution", the Georgian government took the issue of illegal smuggling more seriously and by early June 2004 Ergneti market had been closed down. While Georgian customs reportedly claimed a fourfold increase in revenue (International Crisis Group 2004a, 10f.; factcheck.ge 2019), relations between ethnic Georgians and Ossetians in the conflict zone further deteriorated and eventually resulted in the brief outbreak of violence in summer 2004. Livelihoods of the local population on both sides of the ABL were also negatively affected as their only source of income evaporated. More importantly, however, "the market was a means for average Georgians and Ossetians to meet, build contacts, and identify common interests after the war years [of the early 1990s]" (International Crisis Group 2004a, 10f.).
At the time as well as since then, many have criticized the decision by Saakashvili's government to close this market down. As one of our sources makes clear, "up to 3,000 people, both ethnic Georgians and Ossetians were involved in this market one way or another . . . this is a huge number for such a small region as South Ossetia. It served as a place of inter-ethnic mixing through trade" (Interview 07).
The importance of the Ergneti market was also recognized by However, in another implicit recognition of the benefits of the trade relations that the market had helped to sustain, there has been an attempt to create an Ergneti-style market near the Georgian-Abkhaz ABL. Located in the Georgian village of Rukhi near the Enguri/Ingur bridge, the initiative was praised by many of our Georgian government interlocutors at the time of its opening in 2016. However, a year later, there was an acknowledgment that "the expectations we had [about Rukhi trade centre] haven't materialised . . . It had big potential . . . But unfortunately, our attempt to turn it into a multi-purpose . . . market didn't work" (Interview 20).
By 2018, however, the idea was back on the agenda when "A Step to a Better Future" envisaged the creation of "Special Economic Zones" in Rukhi and Ergneti with several "service-oriented public institutions" such as "offices of the Revenue Service, commercial banks, currency exchanges, and notary services to be opened in the two villages. In addition, agricultural and auto markets are to be built in these areas, along with a multi-functional training center, a business incubator, and a free industrial zone" (OC Media 2018a).
It is possible that a Rukhi-type trade center in Ergneti has a better chance of success than it did in Abkhazia. One of the reasons is that the population of South Ossetia largely depends on imports from Georgia, especially during the harsh winter period when the Roki Tunnel connecting South Ossetia to the Russian Federation is often closed. The products in Gori (on the Georgiacontrolled territory) are reportedly two to three times cheaper than the same products in Tskhinvali, and "up to five times" cheaper than products imported from Russia (International Crisis Group 2018, 6).
These nascent renewed trade relations, however, are severely hampered by Russia's borderisation policy in South Ossetia (Interview 15). As one EUMM official (Interview 03) pointed out, despite the fact that "people are used to living together" new wired fences significantly hinder the freedom of movement of local populations. Currently, only some 4,000 people with permanent residence in Akhalgoni, the only majority ethnic-Georgian town in South Ossetia, are allowed to cross into Georgia-controlled territories (Interviews 03, 16).
Despite these restrictions, trade has been on the rise between South Ossetia and Georgia since 2015. On average, up to 1,500 tonnes of cargo arrive in Akhalgori, "the only crossing for motor vehicles . . . connecting the South Ossetian town of Akhalgori to the Mtskheta-Mtianeti region of Georgia", every month (International Crisis Group 2018, 5f.). According to South Ossetian regulations, everyone who passes through the checkpoint has to operate within the imposed limit of 50 kg per person on goods they can take back to South Ossetia. In 2017, de facto authorities tried to restrict these weight limits even further; however, this was met with a series of boycotts by truck drivers "that emptied South Ossetian markets" and forced de facto authorities to reconsider further restrictions (International Crisis Group 2018, 6).
In February 2017, Georgian and Russian authorities also announced negotiations about the possible formalization of a transport link that would go through the Roki Tunnel, which links South Ossetia to North Ossetia (Russian Federation) and was initially envisaged in the 2011 customs agreement -"a precondition for Georgia giving the green-light to Russia's World Trade Organisation membership that year" (OC Media 2018b). Officially known as the "Agreement between the Government of Georgia and the Government of the Russian Federation on the Basic Principles for a Mechanism of Customs Administration and Monitoring of Trade in Goods", the accord was brokered by Swiss mediators and envisions opening of three trade corridors that will go through the Kazbegi-Upper Lars crossing, the Roki Tunnel, and along the de-facto Black Sea border between Russia and Abkhazia (Civil Georgia 2011). According to this agreement, a neutral third party -Société Générale de Surveillance (SGS), a multinational company with headquarters in Genevawould "be tasked with monitoring the goods entering in and coming out of the trade corridors" (Bilanishvili 2018). For years, the negotiation process on the implementation of this agreement was stalled, but after months of intense talks, "in December 2017, Georgia unexpectedly signed its contract with SGS", followed by Moscow in May 2018 (International Crisis Group 2018, 16). By February 2019, it seemed that the formal procedures necessary to start "the practical implementation of the agreement were completed" (JAM News 2019), but this has not yet resulted in the opening of the crossings on the territories of Abkhazia and South Ossetia. In fact, at the State Ministry of Georgia for Reconciliation and Civic Equality officials were quick to point out that the implementation of this agreement was a separate matter and had nothing to do with the government's 2018 "A Step to a Better Future" peace initiative (Interview 27).
At the moment it seems that the opening of the trade corridor through South Ossetia might only be possible in emergency situations, as was the case in November 2016. In a mountainous region prone to landslides and avalanches, this would still be a significant step forward. While there remain a lot of questions regarding status-related issues concerning passport control and customs at the checkpoint, the fact that this discussion has even got to this point is worth of some optimism. With rapidly falling Russian financial assistance, this transit trade corridor might solve at least some of South Ossetia's financial woes. From Tbilisi's perspective, "[i]t could help unlock the region, which has remained [largely] isolated since the 2008 war" and remains "highly dependent on Russia's security support" (International Crisis Group 2018, 22).

The case of Moldova
With a territory of only 12.4% and a population of 15.2% of the then Moldovan Socialist Soviet Republic, today's Transnistria and Bender/Bendery together accounted for about 40% of the union republic's gross national product in 1990. The region provided transport links to Ukraine's seaport of Odessa and the Donetsk region in eastern Ukraine, was industrially highly developed, and hosted the largest hydroelectric power station generating 85-90% of Moldova's electricity. Transnistria's de-facto secession in 1992-3 thus affected the rest of Moldova not only politically but also economically in a much more significant way than similar developments in Georgia. Already one of the smallest economies in the former Soviet Union, Moldova's economy was now split "into two even smaller economies" and the country experienced "one of the steepest economic declines" among former Soviet republics (World Bank 1998, 3f.). Subsequently, Transnistria embarked on a path of decoupling its economic institutions from the rest of the country establishing its own central bank, issued its own currency, and set up separate customs services.
However, "the level of integration prior to the conflict was so high that trade links have been partially maintained despite all the difficulties, and the exchange of goods remain[ed] rather intensive" throughout much of the 1990s (World Bank 1998, 6; see also Prohnitchi and Lupusor 2013). As one of our Transnistrian interlocutors pointed out, "we had a war, but we still had [intercommunity] trade. War stopped and trade continued" (Interview 12). In fact, trade not only continued, but from the mid-1990s onwards a degree of economic (re-) integration has always been a reality. At the time, this coincided with a shift in Russian foreign policy toward more diplomacy (International Crisis Group 2004b, 5f.). As a result, relations between both Moldova and Russia and Moldova and Transnistria improved, and several agreements were signed between Chisinau and Tiraspol, including an agreement on regulating the monetary and credit systems. As part of this agreement, Moldova opened a branch of its National Bank office in Tiraspol. This was followed by a customs agreement which abolished customs posts between the sides and established joint customs points at the border with Ukraine. Although Chisinau also shared customs stamps and seals with Tiraspol (International Crisis Group 2004b, 15;World Bank 1998, iif.), the customs agreement did not achieve a common customs space, but it simply legalized large-scale smuggling via Transnistria. Similar to Georgia, this created and entrenched powerful vested interests in Transnistria and Moldova, as well as in Ukraine and Russia, keen on maintaining the status quo. In fact, as one EU official put it, the current situation in Transnistria is "one of comfort rather than a conflict. There is nothing to fight about. All there is are vested economic interests" (Interview 04).
The lucrative economic situation of the late 1990s began to change from 2001 onwards as the newly elected president, Vladimir Voronin (in office 2001-2009) tried to use political and economic pressure to coerce Tiraspol into a settlement, additionally hoping for parallel Russian pressure on Transnistria in the context of the so-called Kozak plan (Beyer and Wolff 2016). As this initiative faltered in 2003 in the wake of Western pressure on Voronin and his own fears of losing power given the popular discontent with the settlement proposed in the Kozak plan (Hill 2012), Voronin's "European turn" facilitated further measures that increased economic pressure on Transnistria. The first step in this direction was the introduction of new customs stamps and seals in line with WTO standards (International Crisis Group 2004b, 7). This was followed by the signing of a number of Protocols between Moldova and Ukraine in 2003, which demanded that goods transported through the Ukrainian-Moldovan border, including its Transnistrian segment, have official customs documents issued by Chisinau. However, in accordance with this new rule, Transnistrian enterprises could register with the Moldovan State Registration Chamber and receive certificates of origin from the Moldovan Chamber of Commerce and Industry. With approximately 1,000 Transnistrian companies registering this way over the following decade (Interview 08), economic relations between Moldova and Transnistria further consolidated, which, in turn, contributed to the stabilization of the status quo, political tensions notwithstanding.
In the wake of the Orange Revolution in Ukraine in 2004, and a subsequent Ukrainian conflict settlement initiative, the so-called Yushchenko Plan, several agreements were signed that had significant implications for Transnistria. On 7 October 2005, a Memorandum of Understanding between the European Commission and the Moldovan and Ukrainian Governments established the legal basis for the EU Border Assistance Mission to Moldova and Ukraine (EUBAM), which was launched later that year. This was followed by another agreement between Chisinau and Kyiv on 30 December 2005 "on normalizing the trade regime" on the border between the two countries (Socor 2006).
Despite political tensions, trade relations continued and acquired a new dimension as Chisinau intensified its efforts to establish stronger relations with the EU, which are also seen as "the necessary precondition for the country's sustainable reintegration" (Groza 2015, 5, also Interviews 08, 09, 16). Particularly important in this context were the launch of the EU's Eastern Partnership (EaP) program, which saw Moldova beginning to negotiate its own AA and the related DCFTA from 2010 onwards under a government of the Alliance for European Integration which had succeeded Voronin's communist government in 2009 in events akin to yet another color revolution in the post-Soviet space.
Crucially, and in clear difference to Georgia, Chisinau invited Tiraspol to participate in this Moldova-EU negotiation process as observers after a change in government in Tiraspol from long-term leader Igor Smirnov (in power since 1991) to Yevgeny Shevchuk (in office 2011-16) in December 2011 (Gumene 2019;Interviews 04, 13). For Chisinau it was important to apply the AA throughout the entire country, including Transnistria (Groza 2015, 5 see also Interviews 01, 02, 04). The negotiations were completed in 2013, and in June 2014, Moldova and the EU signed the AA, the application of which, together with the DCFTA, provisionally started from September 2014. By that time, the EU Council had decided to extend the Autonomous Trade Preferences (ATPs) offered to Moldova (including Transnistria) since 2008 until the end of 2015 when they were to be replaced by the DCFTA from 1 January 2016 (Groza 2015;Całus 2016b). Under this scheme, Moldova and Transnistria "could export certain quotas of their goods to the EU market without paying customs fees" (Całus 2016b). Already in 2012, for instance, Transnistria's exports to the EU accounted for just under 30% of all exports (Prohnitchi and Lupusor 2013). Just three years later, "the economic interest of the Transnistrian region is linked, with over 70%, to the Western economic space" and that without the DCFTA Transnistrian budget "could lose about 50 million USD, which would represent about 5% of the GDP" (Groza 2015, 6; see also Całus 2016b). However, despite economic rationality seemingly pointing toward Transnistria participating in the DCFTA, its authorities continuously insisted on the region's economic and trade links with Russia and the Eurasian Economic Union (EEU). In fact, as one official argued at the time, "the extension of the implementation of the Association Agreement into the Transnistrian region would be part of a geopolitical war against Transnistria" (quoted in Groza 2015, 6).
Thus, it seemed that Transnistrian authorities were not willing to make any concessions on this issue and were to miss out on having access to EU markets. In the early period after the EU and Moldova started negotiations on the AA in 2010, the Transnistrian side generally did not seem to appreciate the potential magnitude of the (negative) impact of being left out of a DCFTA with the EU, nor did political and economic dependence on Russia help to change this generally anti-DCFTA attitude until 2015 (Lupusor 2015;Całus 2016a;Interviews 04, 13). The critical role of the knowledge gap was recognized by the EU, and the British Embassy in Chisinau, in particular, organized training events to raise awareness of the DCFTA in the Transnistrian business community. Over time, capacity building was intensified especially at the level of working jointly with the Moldovan and Transnistrian chambers of commerce on trade relations and the impact of the DCFTA. This created both greater awareness on the Transnistrian side of the challenges and opportunities presented by DCFTA implementation and helped to put into place a broader coalition of actors on the left bank willing to reach a compromise with the EU. 1 Thus, following intensive negotiations between Brussels and Tiraspol, a deal was reached in December 2015 under which the EU accepted the inclusion of Transnistria into the application of the DCFTA for Moldova as of 1 January 2016 (Gumene 2019;Interviews 04, 13, 27).
While Tiraspol is strengthening its ties with the EU, even if rather unwillingly, there is hope in Chisinau and in Brussels and Vienna that this might lead to stronger ties with Moldova as well. However, the change in Transnistria's economic policy orientation toward the EU does not necessarily indicate its willingness to reintegrate with Moldova. On the east bank of the Nistru/ Dniester, there is widespread acknowledgment that "the EU is an important trading partner" they cannot afford to lose 19). Indeed, Transnistria's exports to EU countries have overtaken its exports to the EEU. This has created some tensions, as one of our interlocutors pointed out, "on the one hand, we want to have free trade with the EU which implies harmonisation of our legislation in accordance with EU standards, but at the same time we are intrinsically linked with Russia and the EEU . . . so we have to balance all this" (Interview 12). Part of this balancing act is the extension of the DCFTA to Transnistria because it further consolidates economic relations between Moldova and Transnistria, stabilizes the latter economically as well as socially and politically, and demonstrates that (economic) engagement with Moldova and the EU is not equivalent to severing ties with Russia (cf. de Waal 2016). This is critical for the stability of the Transnistrian economy which in 2017 depended for more than 60% of its exports on access to EU markets and for just under 75% of its imports on CIS countries, with 34% of imports being oil and gas, which in turn make exports to the EU competitive (see Gumene 2019; Interview 27).
These issues around the DCFTA, and the ups and downs in trade relations between Moldova and Transnistria in the 1990s and 2000s more generally, indicate, that the quality of economic relations depends, like in the case of Georgia, as much on domestic politics as on the geopolitics surrounding the conflict (Ademmer, Delcour, and Wolczuk 2016;Beyer 2010;Beyer and Wolff 2016;Douglas and Wolff 2018). Equally importantly, Transnistria's economy almost entirely depends on trade, remittances, and Russian subsidies. Tiraspol uses income from exports (rather than taxes) to generate money for its budget, which makes it much more vulnerable to external pressures (Interviews 01,10,13,22,24). A steep 20% decline of the economy in 2015 (Lupușor et al. 2016) on top of the steady economic decline over the two previous years meant that by 2016 "Transnistria faced the threat of economic catastrophe" (Całus 2016b; see also Lupușor et al. 2016;Leskaj and Jonasson 2016) and businesses there began to lobby not only their own authorities but also the Moldovan Ministry of Economics to take mitigating action (Interviews 11,13,26). Due to its landlocked geographic position, Transnistria has also been particularly affected by the ongoing conflict in Ukraine. Devaluation of the national currencies of its main trading partners (Moldova, Russia, Ukraine) also played a role. A weaker Russian economy, largely due to Western sanctions since 2014, falling oil prices, and the cost of the wars in Syria and Ukraine also meant that financial support from Moscow was gradually decreasing (Interviews 18,26). Under these circumstances, the authorities in Tiraspol were under significant pressure, especially from the Transnistrian business community, to accept the deal offered by the EU on the application of the DCFTA to Transnistria (Interviews 01,13,21,22,24). At the same time, as one EU official pointed out, "Transnistria does not have to assimilate with Moldova, but it has to harmonise [its procedures]" (Interview 21). This allows the Transnistrian authorities to maintain their stance that more regulated economic relations with Moldova and the EU do not undermine its independence aspirations (Interviews 12, 19) while EU and OSCE officials can claim that the further approximation of the Transnistrian economy to DCFTA standards will, at a minimum, make future reintegration easier (Interviews 01, 24, 28, 29).
Russia's acquiescence to this new arrangement was most likely also driven by such status-quo considerations (Interviews 01, 10, 13, 22). On the one hand, Moscow did not really lose any of its influence in Transnistria itself or any leverage that results from it. On the contrary, the economic stabilization in Transnistria facilitated by the deal with the EU contributes to maintaining the status quo and reduces any pressure for a political settlement. Economically, too, there are clear benefits, as many of the Transnistrian businesses whose continued and increasing access to the EU is assured by the application of the DCFTA are owned or controlled by Russian conglomerates (Chamberlain-Creanga and Allin 2010; also Interview 10).
For the EU and Moldova, the benefits were primarily political and in line with the expectation that trade can facilitate confidence building and as such lay the foundations for a future conflict settlement. While the latter part of this equation still needs to be realized, the agreement on the so-called Berlin Plus package (or Package of Eight) in 2016 and its near-complete implementation by mid-2019 indicates that there was a degree of spill-over from the EU and Chisinau's constructive approach to the DCFTA negotiations into other areas of the conflict management, if not settlement, process (Interviews 01,20,21,26,28).

The primacy of politics?
For a quarter of a century, the conflicts related to Abkhazia, South Ossetia, and Transnistria have remained unresolved. Their management has been conducted primarily through efforts to build confidence between the sides, and the promotion of economic relations, and trade in particular, has been a prominent feature among such efforts. Whether it is through programs and projects sponsored by the EU, through dialogue formats involving international organizations like the UN or OSCE, or initiatives promoted by the Georgian and Moldovan governments, there is an evident expectation that trade between the sides will contribute to increased confidence and eventually to conflict settlement. The application of the DCFTA to Transnistria appears the most promising development in this regard. As such, the process of negotiating and implementing this specific measure which had at least some confidencebuilding effects warrants more detailed analysis, especially with regard to any lessons that could be applied to Georgia. The two cases are, of course, very different. Unlike in Transnistria, it is more difficult to use trade as leverage in Abkhazia and South Ossetia, as an incentive or otherwise, because of their proximity to Russia and the structure of their economy (Interviews 01,02,05,06,18). At the same time, however, Russian financial assistance to the two regions has decreased and traditional sources of income (such as Russian tourism in Abkhazia) have dried up (Gaprindashvili et al. 2019;Charaia 2015). The economic uncertainty thus created, and the further deterioration in living conditions that result from it, have heightened a sense of pragmatism in both break-away regions, and more so in Abkhazia where businesses are keen on diversifying economic and trade relations beyond Russia and would welcome more European goods into their markets too, "which tend to be considerably cheaper than their Russian equivalents" (International Crisis Group 2018, 10f.). However, as one economic analyst from Sokhumi/ Sukhum put it, "[t]rade with the West is possible, but with too many headaches" (International Crisis Group 2018, 10). For de facto authorities in Abkhazia, the main "headache" is more about political challenges rather than trade-related and is intrinsically linked to their sensitivity about the status of Abkhazia. Nevertheless, despite publicly rejecting Georgia's recent peace initiative, in private conversations, Abkhaz officials, businessmen, and analysts seem to take a more pragmatic stance (Interviews 17,25,27).
Another key difference between Moldova and Georgia is the fact that Moldovan-Transnistrian relations have a much longer and more successful track record of economic cooperation, which has, among other things, resulted in greater awareness of the benefits of trade, including from a purely financial perspective. By contrast, as one EU official pointed out, when it comes to Abkhazia and South Ossetia, "we need to start from the basics here and think about restoration of some basic relations. . . .. Motivation of every business is to make money so as long as there is safe environment to do this, the opportunity of re-building economic links, and potentially restoring good neighborly relations between communities, is already there actually" (Interview 02). Yet, one of the acute problems that the conflicting sides often face is "isolation from knowledge . . . which sometimes leads to paradoxical policies" (Interview 02; see also Interviews 01, 23).
The longer and more successful track record of confidence building in the Moldovan case and the isolation from knowledge that is apparent in the cases of Abkhazia and South Ossetia, in part, also explain that Georgian-Abkhaz/South Ossetian attitudes toward one another diverge significantly from the Moldovan-Transnistrian attitudes. This is, of course, also due to the different history and intensity of these conflicts, but their divergent trajectories and attitudes over time have become mutually constitutive. In the absence of political will on both sides of the respective conflicts in Georgia, it is difficult to see how lack of engagement, including through economic confidence building, can lead to anything but further hardening attitudes and vice versa. Likewise, in the Moldovan case, it will require political will on both sides to maintain, and potentially expand, current levels engagement and confidence building and prevent retrenchment into antagonistic attitudes. Thus, while political will is required to prevent an unraveling of progress achieved through confidence building, another question is whether regulated or unregulated trade can have a spill-over effect beyond economic confidence building and pave the way to actual conflict settlement.
Despite significant setbacks over the past 25 years, the Georgian government's 2018 "A Step to a Better Future" peace initiative demonstrates that the expectation of confidence building through trade remains alive and well. Yet, this does not imply that an extension of the DCFTA to Abkhazia (similar to what was done in Moldova with Transnistria) is feasible, let alone viable. As one interlocutor in Brussels pointed, "there seems to be a lot of excitement about the DCFTA and a desire to copy it in other cases but there are a lot of strings attached to it which makes it quite a fragile tool. Transnistria is a unique case and therefore the DCFTA deal there is not repeatable any time soon anywhere else" (Interview 01, also Interviews 02, 14). While "the EU's quiet diplomacy with Abkhazia on the DCFTA extension" (International Crisis Group 2018, 27) may be a first step in the direction to ascertain the transferability of the Transnistrian experience to Georgia or search for alternatives, the experience of Moldova and Georgia also indicates that for trade to have confidence-building effects, there needs to be political buy-in from relevant authorities in the separatist regions, the national government, and their respective international partners. More importantly, it is this comprehensive political buy-in from all sides specifically that facilitates the kind of regulated trade relations that are predominant in the Moldovan case and which have proven more resilient over time than the by-and-large unregulated economic transactions that prevailed in Georgia.

Conclusions
We started this article by asking whether confidence-building measures facilitate conflict settlement in the post-Soviet space. In light of the evidence that we gathered from our case studies, the answer to this general question has to be no. Yet, if we break this question down and ask more specifically when and how trade contributes to meaningful and purposeful engagement between the conflict parties that can pave the way toward conflict settlement, this "no" has to be qualified in some important ways.
The case of Transnistria, in general, and specific periods in relation to the conflicts in Abkhazia and South Ossetia, illustrate the clear benefit of trade across conflict lines, and the overall stabilizing, violence-reducing effects that such economic relations can bring about. These benefits are broadly recognized as is evident from the resumption of trade after every major crisis in Georgia and the continued commitment to facilitating, or at least tolerating, trade relations by the Georgian and Moldovan governments and their international partners. The authorities in the separatist regions, too, acknowledge such benefits, albeit to different degrees and often in deed more so than in word. Equally importantly, the people and businesses directly affected by the divisive legacies of the conflicts in, and since, the early 1990s, value economic linkages for their economic and stabilizing effects.
However, while one could take the view that any form of trade is better than no trade, it is important to note several significant caveats that arise when taking a long-term view on our three cases. Unofficial, unregulated economic activity limits the potential for confidence building and for leveraging and strategically locking in any potential confidence gains. It can provide examples of the possibility of inter-ethnic and cross-community collaboration, but it also creates and entrenches constituencies with little interest in moving beyond the status quo, precisely because they benefit from the unregulated nature of their activities. This decreased the likelihood of a resurgence of violence in the case of Transnistria, but simultaneously blocked meaningful progress toward conflict settlement, while also solidifying bad practices of governance.
Three further insights follow from this. First, where unregulated trade does not create broader constituencies in favor of peace (even if it is the cold peace of a status quo), the resumption of violence is more likely. Despite significant cross-ABL trade in Georgia in the 1990s and early 2000s, violent crises could not be averted. Although trade recovered after every such crisis, its unregulated nature clearly has had limited violence-averting effects.
Second, the case of Ergneti market indicates the stabilizing effect of "any" trade relations: while there were no major violent incidents involving South Ossetia after the initial violence in the early 1990s, the market's forced closure in June 2004 was followed by a major resurgence of violence the following summer. The subsequent recognition of the importance of places of trade as places of contact, and thus potentially confidence building, by leaders on both sides has led to new initiatives, such as for the establishment of the so-called trade centers near the ABLs in Georgia. In other words, disbanding rather than tolerating unregulated trade is counter-productive unless alternative and sustainable arrangements are put in place.
Third, gradual or parallel regulation of at least some trade relations has greater potential for stabilization and violence prevention. This is demonstrated by the case of Transnistria in two ways. On the one hand, the greater degree of regulation and legalization of trade over the years has spilled over into other areas of relations and has led to a wide range of agreements being concluded, most recently in relation to the recognition of status-neutral number plates, the re-opening of a major transport link between the sides, and telecommunications (Interviews 22,28,29). On the other hand, the relationships built between business and political communities on both sides of the divide and with the EU, enabled progressive engagement on further trade facilitation that first saw ATPs extended to Transnistrian companies and then the DCFTA applied to the region as well.
All three cases thus also underline that the feasibility and viability of trade as a confidence-building measure is highly contingent upon the context in which it happens, and success in one case is not necessarily transferable to another one. This is particularly obvious in relation to the EU-Moldova DCFTA and its application to Transnistria. There was a clear Transnistrian interest (to retain the ability to export goods to the EU and to avoid a deeper economic crisis), a similarly clear Moldovan interest (to avoid a formal internal economic boundary), and a clear EU interest (to prevent further political and economic destabilization in the region). Moreover, as these interests partially overlapped, i.e., were shared among some or all actors, finding a solution was possible. Equally importantly, there was an inclusive process of finding this solution, and one which was not blocked by Russia. In addition, the pre-existing level of regulation significantly eased some of the technical difficulties of the application of DCFTA to Transnistria, reduced the risks involved in this, and could build on an existing network of trade professionals in the private and public sectors on both sides and in the EU to negotiate, implement, and operate the new arrangements. Replication of this deal to Georgia might prove more difficult (albeit not impossible), given the very different circumstances in Abkhazia and South Ossetia: direct access to and from Russia, seemingly significantly lower importance of trade with Georgia and the EU, a lower density of pre-existing linkages across the divides, and arguably a different level of geopolitical stakes on both sides. Put differently, the balance of power and capabilities between coalitions for and/or against change in the case of Transnistria are different from the situation in Abkhazia and South Ossetia.
That said, our cases do point to the need, and to some extent appetite, for the regulation of economic relations. This is most obvious in the case of the application of the DCFTA to Transnistria, but also apparent in recent efforts in Georgia. In this way, considering the underpinning infrastructure for trade, such as transportationboth in the sense of physical infrastructure of roads, railways, and border crossings, and in terms of associated regulation, such as the recognition of number plates, for instanceis equally important. Agreements in these areas facilitate trade and thus create positive, mutually reinforcing spill-over effects, whereas protracted failures to reach, or implement, agreements have the opposite effect.
Confidence building, through trade or other mechanisms, is ultimately a question of political will which is often lacking. Yet, it can be fostered by focussing on the capacities and interests of the relevant actors on both sides of a divide and on the structures in which they operate. Future confidencebuilding initiatives will have to recognize the primacy of politics and thus reflect this need to build capacities, re-orient interests, and create structures that enable confidence building as a foundation for feasible and viable conflict settlements. Trade can contribute to this, but even in its more regulated forms, it is insufficient in itself to live up in reality to the often-lofty expectations associated with it in some of the academic and policy literature. Note 1. One of the authors acted as an advisor on this project.