Taking ambiguity seriously: Explaining the indeterminacy of the European Union conventional arms export control regime

Rational institutionalism

Rational accounts, most notably, rational (choice) institutionalism (RI), have long devoted attention to why international agreements acquire particular designs. RI assumes that states and other actors behave instrumentally and strategically to pursue more or less fixed interests, and turn to multilateral cooperation when unilateral action cannot secure the desired benefit. Interests are seen in terms of a material, utilitarian, strategic calculus, and so are treaties, which are signed on the basis of material cost–benefit calculations (Abbott and Snidal, 2000; Koremenos et al., 2001; Lipson, 1991). Accordingly, the EU arms export control regime represents a strategic response to an arising interest to converge arms export control on material grounds. RI is utilized here to identify states’ or other key actors’ material incentives in regime formation, evolution and design. The arms trade is embedded with material interests: arms export serves suppliers’ security, power and status through enabling them to sustain the defence-industrial base, back allies and friends, weaken adversaries, establish ties with strategically important countries, and secure jobs and revenue (Hook and Rothstein, 2005). As returned to later, the material interests most clearly associated with the control of arms export in the context of the EU are security, export promotion and sovereignty.

Both during regime formation and change, regime design results from the sum of self-conscious, rational, utilitarian, purposive interactions among states (and, to a lesser extent, of interest groups and corporations) (Koremenos et al., 2001: 762). Ambiguity, in particular, results from compromises and/or a preference for ambiguity (Fearon, 1998; Koremenos et al., 2001). Regarding the former, compromises are common when heterogeneous groups negotiate a treaty. Where heterogeneity threatens agreement, ambiguity is often unavoidable and language is watered down to the lowest common denominator focal point in order to seal the deal. As a result, as the history of EU integration clearly illustrates, ambiguity and heterogeneity typically grow proportionately (Abbott and Snidal, 2000: 445). In particular, unevenly distributed collaboration costs will often impede precision because states facing high collaboration costs will be likely to favour imprecision and low obligation, or opt for special treatment (Koremenos et al., 2001). This dynamic is likely to influence the export control regime because the size and export-dependence of member states’ defence industries varies. Such variation has resulted in ambiguity in comparable arms export control regimes; for instance, Greene et al. (2002) show that the United Nations’ (UN’s) Programme of Action was watered down to secure ratification in the presence of major exporters’ resistance.

Moreover, even under homogeneity, all states may prefer to maintain flexibility, particularly if they face high collaboration and sovereignty costs. In this understanding, the resulting ambiguity is tailor-made and purposeful, or what US Secretary of State Henry Kissinger famously termed ‘constructive ambiguity’, which can be understood as ‘the deliberate use of ambiguous language in a sensitive issue in order to advance some political purpose’ (Berridge and James, 2003: 51). Following Kissinger, ambiguity can be used strategically to obscure obligation in the sense that clear, behavioural proscriptions are either purposefully avoided or sufficiently open to be circumvented. This is arguably a commonly used technique in arms export control regimes; for instance, Yanik (2006) describes arms export control regimes in general as flexibly worded in order to enable the pursuit of material interests.

In sum, this suggests that ambiguity correlates closely with the heterogeneity and resistance associated with material interests, unevenly distributed collaboration costs, and sovereignty costs. The article explores whether the ambiguity of the EU export control regime is traceable to any of these dynamics.

Constructivism

Constructivism also ascribes ambiguity to heterogeneity. To constructivism, ambiguity is a key mechanism of norm diffusion because ‘detail is not necessarily conducive to agreement’ (Wiener, 2004: 198). Ambiguity also enables the content of norms to be ‘filled in many ways and thereby to be appropriated for a variety of different purposes’ (Krook and True, 2012: 104). Hence, regimes may be adopted precisely because they mean different things to different people (Krook and True, 2012; Van Kersbergen and Verbeek, 2007). This implies either unavoidable or purposeful ambiguity, depending on the nature of the underlying heterogeneity. Hence, RI and constructivism pull in the same direction: ambiguity is a symptom of heterogeneity and enables regime adoption and change through a ‘one size fits all’ formula that maximizes the potential for consensus by obscuring obligation.

Nonetheless, the accounts disagree about the dynamics of heterogeneity. The view that arms export control regimes are adopted as a function of utilitarian, material interests is widely criticized by constructivist research on arms trade regulation due to the incidences of regime adoption left unexplained when focusing on material interests (e.g. Müller, 2013a). In contrast, constructivist accounts argue that ‘norm dynamics’ are vital to explaining arms export control regime adoption, evolution and design (Lantis, 2011: 545; Wunderlich, 2013). ⁶

Following recent constructivist scholarship, regimes are constructed and continuously evolve in response to the (1) the external environment and (2) internal regime contestation, notably, by (3) triggering agency (civil society, key individuals and states) response. These processes and pressures are the norm dynamics decisive to the formation and evolution of regimes and the norms constituting them (see Wunderlich, 2013). First, the external environment of regimes refers to the ever-changing domestic and external structures that enable and constrain regimes through determining their ideational foundation (Cortell and Davis, 2000: 65–66; Lantis, 2011; Wunderlich, 2013). Changes in domestic politics, sudden events like crises or wars, long-term events like gradual shifts in power, technological change, and the processes of other norms ‘out there’ will typically trigger regime adoption and change (Krook and True, 2012; Lantis, 2011; Wunderlich, 2013). Second, internal regime contestation refers to the fact that once in place, ambiguity is itself a source of regime evolution, and may facilitate a ‘battle over meaning’ over the definition of norms, what they proscribe and what constitutes compliance (Sandholtz, 2008; Van Kersbergen and Verbeek, 2007; Wiener, 2004). The more ambiguous regimes are, the greater is the likelihood that their meaning will shift or wither in response to external events (Wiener, 2009), for instance, through providing a springboard for advocates of ‘counter-norms’ (Bailey, 2008). Lastly, the external environment and internal regime contestation depend on the ability of agency (civil society, key individuals and states) to influence regime adoption and/or change through advocacy and (re)interpretation. Agents may respond by way of promoting new regimes/norms, (re)interpretation and (re)representation of old ones, or by pushing regime/norm substitution or decay (Krook and True, 2012: 105–107; Wunderlich, 2013: 21).

Several studies attribute the regimes that control the export and usage of various conventional weapons (cluster munitions, landmines and small arms) to normative progress and ethical formulas (e.g. Garcia, 2006, 2011; Joachim and Dembinski, 2011; Müller and Wunderlich, 2013; Price, 1998). In compatibility with the aforementioned framework of norm dynamics, these studies typically attribute the adoption and evolution of these regimes to advocacy efforts and ‘gatekeeping’ politicians promoting (through ‘norm cascades’ and by means of ‘shaming’) a ban or limitation on the export of a weapons system. New norms/regimes are institutionalized when these advocates are eventually ‘empowered’ by structural developments (crises, wars), scandals, moral progress and/or norm conflicts. It is both explicit and implicit in these studies that the emerging regimes express ‘nice’ norms and intentions of restraint in arms export. Although some of these studies recognize that ‘interests’ play a role, interests (for instance, material interests in arms export) are treated as relevant only at the implementation stage, not at the stages of regime formation and evolution (e.g. Müller, 2013b). This preoccupation with norm dynamics leaves very little scope for assessing alternative explanations. As the following sections highlight, the exclusive attention on the progress of nice norms misses how norm dynamics and material interests interact in regime formation and evolution. Even more so, it overlooks how material interests are indispensable to explaining the formation, development and design of the EU arms export control regime.

Enhancing security

Arms exports can be both necessary for and damaging to the security of a state. On the one hand, arms exports can help maintain a national defence-industrial base capable of at least partly sustaining military and police forces, thus serving the security of the exporting state. Arms exports can also benefit security indirectly through serving the security of allied states, and providing material for peacekeeping operations. For these reasons, states promote arms exports in foreign markets. On the other hand, arms exports can subject states, their friends/allies and citizens to security threats if armaments are used offensively; thus, restrictive arms export control entails security benefits. Against this backdrop, the export and control of arms is a delicate balancing exercise between the promotion and restraint of arms exports.

It is generally acknowledged that the latter security dimension, more specifically, the enhancing of security for EU member states and their friends/allies (i.e. state security), made up the very core of European states’ calls for arms export control around 1990. ⁸ This connected intimately to the security concerns following the export practices of the 1980s and 1990s, when European arms manufacturers exported weapons to conflict zones, pariah states, repressive regimes and even embargoed destinations, oftentimes with governmental support (Phythian, 2000). More than any other event, the 1990 Iraqi invasion of Kuwait in the First Gulf War put the inadequacies of national export control regulations on display as European weapons boomeranged on Coalition troops fighting Iraqi troops during Operation Desert Storm. This generated massive public outcry for stricter and more transparent export control policies so as to prevent ‘another Iraq’. It also triggered arms export scandals that resulted in revised legislation in several member states (see Anthony et al., 1992: 291; Cornish, 1995: 4; Davis, 2002: 51).

The export revelations had very clearly demonstrated the ease with which rogue states could acquire destabilizing arsenals of weapons from Europe. Therefore, the momentum in the early 1990s’ security rationale lay in finding a way to prevent the supply of armaments and potentially dangerous technologies to unreliable end users. This reveals a highly material rationale for arms export control as a means to enhance security. In the EC, the completion in 1992 of a common market for civilian goods with the Single European Market (SEM) intensified this rationale. In particular, the European Commission (Commission) and the European Parliament (EP) argued that converged arms export control regulations became all the more important with the growing risk that in a single market, weapons and technology could escape the EU through the countries with the laxest export control laws, and hence compromise the security of all (European Commission, 1996; European Parliament, 1992). As described by Davis (2002: 1), the SEM also threatened to undermine unilateral export control efforts through enabling opportunistic manufacturers to produce strategic equipment in and license exports from the state with the most relaxed laws. Broad consensus prevailed among governments, NGOs, the EP and the Commission that only multilateral measures could be effective.

Calls for a level playing field

A second rationale heard in the calls for a common arms export control approach was commercial, and can be understood as an attempt to create a ‘level playing field’ for the European defence industry. This rationale, also highly material, was inextricably connected to the changing structure of the armaments market after the Cold War. Defence industries worldwide had to adapt to an entirely new reality as changed security perceptions led to new procurement and armaments planning priorities, drops in military spending among suppliers and buyers, and a loosening of the previously symbiotic industry–state relationships (Scherpenberg, 1997: 102; Schmitt, 2000: 10). In Europe, the process towards European Monetary Union (EMU) also contributed as the rigorous convergence criteria of the EMU required budget deficit reductions, partly achieved through defence budget cuts (Bauer, 2003: 61). In the early 1990s, overcapacities, stockpiles of surplus weapons and spiralling financial problems were common denominators for industry, and the surviving defence companies relied heavily on exports to offset lower budgets (Bitzinger, 1994; Bjurtoft, 1998). This set the scene for a new age of export-driven proliferation in an armaments market that was more competitive than that of the Cold War, and largely characterized as a ‘buyer’s market’ (Cornish, 1995: 2; Guay, 1998: 12).

Industry was also impacted by rapid technological change and structural difficulties in the European armaments market. Research and development (R&D) costs had spiralled with a gradual shift towards more sophisticated weapons and technologies (Devore, 2012). Meanwhile, the European defence market was characterized by high fragmentation, national division and duplication, and the high costs of R&D disadvantaged smaller companies that had difficulties exploiting economies of scale. Furthermore, European firms were disadvantaged following the dominance of US companies in the high technologies; these were also generally better off after more rapidly consolidating during the early 1990s (Bitzinger, 1994; Scherpenberg, 1997). In comparison, consolidation within the EC and later the EU was bumpier and largely followed national frontiers. Cross-border collaboration typically confined itself to ad hoc joint ventures or consortia, entailing less comprehensive collaboration on single weapons systems (Bitzinger, 1994). This mirrored several factors: national determination to sustain self-sufficiency in defence manufacturing; time-consuming administrative procedures following high trade barriers for defence goods; and divergent arms export control legislation across Europe. With remaining duplication, long production runs and trade barriers, European weapon systems often became more expensive to potential buyers than US alternatives, with US products sometimes preferable even in Europe (Guay, 1998: 12).

Armed with commercial motives and increasingly frustrated over an institutional framework that created uncertainties for joint production projects and impeded the collaboration needed to exploit economies of scale, corporate interests requested EU-level restructuring. On their ‘wish list’ was convergence of legislation governing extra-EU arms transfers so as to facilitate cross-border collaboration, and lower trade barriers on intra-EU transfers (Bauer, 2003: 66; Schmitt, 2000). This implied a ‘level playing field’, where companies could compete on equal grounds (Bauer, 2003: 66; Davis, 2002: 251).

A moral order to the arms trade

A third rationale informing the emerging export control negotiations was that of a moral order to the arms trade. In contrast to the aforementioned rationales, this refers to the non-material, ideational rationale emphasized in constructivist scholarship on arms export control.

Awareness of the association between arms transfers, conflict and human rights abuses had emerged in the 1970s and 1980s, which saw talks about limiting arms transfers to the unstable regions of the Third World for both security and moral reasons (Cornish, 1995). During the Cold War, however, controlling the spread and use of weapons of mass destruction (WMDs) and related technologies received the bulk of the attention (Krause, 2011), as did the state security rationale in the early 1990s. With nuclear war no longer considered probable, and with the numerous civil wars in the early 1990s displaying how conventional weapons intensified conflict and precluded peace talks and post-conflict reparation, arms export control over moral concerns was eventually recast in human security drapery, empowered by the conflict prevention and human security agendas of the mid-1990s. From the mid-1990s, most states and practically all the relevant international organizations engaging in human rights and humanitarian issues began addressing various dimensions of arms export control. Several weapons-specific (landmines and cluster munitions) and more general conventions (small arms) came out of this environment.

In the EU, among the actors promoting a common export control approach for moral and/or human security reasons was especially sections of the EP, NGOs and some (often smaller) member states (Smith, 2001). ⁹ A common denominator in these calls was the request for a multilateral approach to arms export control because uncoordinated and secretive export control policies had previously led to dubious and destabilizing arms transfers.

The tottery steps

As discussions about the prospects for a common approach to arms export control began, the heterogeneity of positions already left little doubt that ambiguity would be hard to avoid, as actors armed with partly competing and partly overlapping rationales entered the negotiations. To the extent that it is possible to distinguish between the impacts of the various rationales in the early 1990s, little doubt remains that it was security concerns following the First Gulf War that triggered and enabled the emerging talks. To varying extents, the other rationales informed the debates through the agents promoting them.

Among the EC institutions promoting a common export control approach, the EP was the most dedicated, reflecting a long tradition of EP engagement on the issue. With the 1978 Klepsch Report, the EP had already suggested abolishing Article 223 and taking steps towards the creation of a rationalized common arms market. The logic was that a more effective market would depend less on exports to the Third World and would therefore be more responsible (Klepsch, 1978). Following the First Gulf War, the EP passed several resolutions urging actions to prevent destabilizing accumulations of weapons through an EC approach (see Anthony et al., 1993; European Parliament, 1992). The EP repeatedly addressed member states’ responsibility for the conditions in the importing countries and regions and the need to establish moral guidelines. It also emphasized a normative ‘sufficiency principle’, which would entail that the exporting states had a responsibility to stop exporting if it was likely that the importing state was acquiring more weapons than necessary (see European Parliament, 1992; Ford Report, 1989: 6; Klepsch, 1978). Hence, both moral and security arguments were visible in the EP’s discourse. However, the EP was not very influential in foreign and security policy matters, and most states considered its calls to be too radical (Cornish, 1995: 20).

The Commission, traditionally a more influential institution in the area of foreign and security policy, had long attempted to bring the trade, production and procurement of armaments in under EC competencies and abolish Article 223 (Koutrakos, 2008). However, in the early 1990s, the Commission primarily concentrated on dual-use export controls following the introduction of the single market (see Davis, 2002: ch. 3). To the extent it focused on conventional weapons, it mainly addressed the potential of a common approach in enhancing the competitiveness of the European defence industry, for instance, vis-a-vis US industry, and that it could become a vital part of the CFSP and a future common defence policy (European Commission, 1996). As such, in contrast to the EP’s moral and security concerns, the Commission echoed the desires of industry in certain ways. NGOs, moreover, had promoted a moral order to the arms trade since the 1980s, and the First Gulf War strengthened these efforts (Stavrianakis, 2010: 6). Nevertheless, NGOs had minor influence on the talks of the early 1990s, presumably because of a lack of access points (Joachim and Dembinski, 2011).

While the EP, the Commission, industry and NGOs pushed for a common export control approach over partly divergent and partly overlapping motives that have been largely consistent since, it was member states’ sovereignty concerns that most clearly came to characterize the actual negotiations in the early 1990s. Importantly, this did not imply the lack of recognition of the utility of a coordinated approach. The First Gulf War had reminded states that closing the loopholes that had compromised security required coordination, and civil society pressure following the export scandals was substantial. Member states were also increasingly confronted with scenarios of lower security of supply in defence material, and with the fact that the existing small-scale cross-border collaboration entailed licensing decisions on joint production and therefore required bilateral or multilateral coordination (Bauer, 2003: 374). For all these reasons, debates about a coordinated EC approach to arms export control became a natural element of the Intergovernmental Conference (IGC) on Political Union, beginning in 1990 and leading to the 1993 Maastricht Treaty establishing the EU and the CFSP (Anthony et al., 1992: 295). The issue of a common approach first appeared in the December 1990 ‘Asolo List’ as one of two areas deemed crucial for the CFSP (the other area being economic and technological cooperation in arms production) (Cornish, 1995: 22–24). It reappeared in draft treaty texts prepared for the IGC, and was supported by a broad coalition of exporters (France, the UK, Luxembourg and the Netherlands), the Commission and the EP. However, the Maastricht Treaty pushed the question of common arms export control into the unforeseeable future and confined itself to deciding that member states under the CFSP had the opportunity — in the future — to reach agreement on the areas in which to take ‘joint action’ in foreign and security policy. The reason was disagreement on the relationship between national and EC competencies, and governments’ fears of the sovereignty costs of drawing this issue area into the process towards supranational integration that Maastricht was all about. Eventually, sensitive issues that could hamper the adoption of the Maastricht Treaty — including arms export control — were excluded from the final stages of negotiation (Anthony et al., 1993: 462).

The first concrete steps towards a common approach were instead taken through the parallel (to the IGC), strictly intergovernmental and presumably less sensitive European Political Cooperation (EPC) process. ¹⁰ The Ad Hoc Working Group on Conventional Arms Export (which later became the Council Working Group on Arms Export (COARM)) was appointed by the EPC Political Committee ¹¹ with the mandate of comparing national positions and practices, and exploring possibilities for further, coordinated action on arms export control (Cornish, 1995: 20–21). Following the secretive nature of this policy area, member states had very little knowledge about each other’s export control frameworks and practices, and comparison was a natural starting point. The outcome of the process was the adoption of seven criteria at the June 1991 Luxembourg European Council through the Declaration on Non-Proliferation and Arms Exports, together with an eighth criterion at the June 1992 Lisbon European Council. The criteria arguably contained elements present in all EC member states’ export control systems, although not necessarily equally emphasized by each member state. ¹² They listed the following principles upon which member states’ arms export control was based (European Council, 1991, 1992):

Later versions of the criteria are very close to the original formulations, although the criteria have become increasingly detailed (for the full versions, see Council, 1998, 2008).

Although noble, the criteria were embedded with ambiguity, inter alia, in the sense that the European Council declarations avoided addressing how they could be applied. At Luxembourg, the European Council was merely hopeful that a ‘common approach’ based on ‘criteria of this nature’ might lead to ‘a harmonization of national policies’ ‘in order to prevent situations of regional instability’ (European Council, 1991). The criteria did not ask member states to abstain from exporting if certain conditions were met, let alone proscribe when conditions in receiving countries would be grave enough for a licence denial. The diffuse application was further compounded by cross-country differences in emphasis, which hinted at likely differences in interpretation even under more precision. The eighth criterion (known as the ‘development’ criterion) is a case in point; while arguably reflecting a comparative process, the criterion was added at the insistence of Germany (and strongly supported by the Netherlands). The criterion was subject to deep controversy due to British and French fears that their long-standing export interests in the Middle East would be negatively impacted. ¹³ As returned to later, these differences influenced the wording of the criterion in the later rounds of regime development.

In some way, this ambiguity and non-obligation was related to the Ad Hoc Working Group not having a policymaking mandate; the adopted criteria were only meant to chart the lowest common denominator upon which a possible common approach could develop. Yet, the mandate was deliberately weak from the outset following some governments’ worries about sovereignty and collaboration costs. Certainly, not all member states signed up to the goal of convergence; some (France and the UK) aimed at a minimalist approach that limited the assessments to a mere comparison of national regulations. Others (Germany and the Netherlands) aimed at a maximalist approach, and saw the process as a step towards deeper integration in foreign, security and defence policy (Davis, 2002: 84). The UK and France, top-ranking European arms exporters with substantial stakes in arms export promotion as a means to maintain defence production (Cooper, 2000; Kolodziej, 1987), repeatedly stressed that the result of the work of the Ad Hoc Working Group ought not to represent a first step towards common export control policies; neither was it a commitment that all criteria were to be applied in all cases (Cornish, 1995: 21). In a Council of the European Union (Council) Decision from 29 December 1994, which weakened the COARM mandate (substantiating the sensitivity of the issue), France insisted on the insertion of the following language: ‘Exports of conventional arms touch upon important and legitimate national interests (foreign policy and other), and no mention in the mandate should be construed as curtailing national decision-making capability’ (quoted in Adam, 1995: 72). This is in line with the RI assumption that states facing higher collaboration costs are likely to opt for ambiguity and weakness. Although also a big exporter, for historical reasons, Germany has traditionally pursued a more restrictive approach to arms exports, and therefore took a maximalist position (Davis, 2002).

Hence, in the tottery steps of the regime, the material interests and cost–benefit calculations of states wary of losing control of an issue area closely intertwined with national sovereignty prevented collective action beyond the comparison undertaken by COARM. Although a certain moral pressure and new norms were emerging following the export scandals, it was material interests — most notably, state security and sovereignty — that set the stage for and determined regime design, including ambiguity, during this stage of the negotiation process. This lends support to the RI interpretation of the roots of ambiguity.

The Code of Conduct: An ambiguous, lowest common denominator

The reluctant sentiments previously described hampered deeper commitment throughout the 1990s. However, in June 1998, member states formally adopted a common approach to arms export control through the politically binding Code of Conduct on Arms Export (Code). Here, they agreed to apply the eight criteria on a case-by-case basis when assessing export requests, in addition to pursuing a number of operative provisions (Council, 1998). As returned to later, the text of the document displayed considerable ambiguity. This was no surprise given persistent sovereignty concerns and competing perceptions of what the Code was a tool for: export restraint or export promotion.

Despite previous UK resistance to a strong regime, the Code was adopted as a UK initiative under the UK presidency of the Council. Several changes in the policy environment, and various agents’ ability to exploit these, facilitated the UK initiative. First, there was a change in the domestic policies of the UK with New Labour entering office in 1997, with ‘ethical arms export’ explicitly on its foreign policy agenda. This shift, in turn, connected to the export scandals dating back to the late 1980s and early 1990s. In the UK, these scandals had led to the launching in 1992 of a judicial inquiry mandated to investigate allegations that Conservative ministers had misled Parliament over arms exports to Iraq during the mid- and late 1980s (Bauer, 2003: 180–181, 369; Miller, 1996). The inquiry’s 1996 conclusions highlighted major shortcomings in the export control legislation and practice of the UK (Scott, 1996). Massive pressure built up to make export decisions more ethical and accountable to the public and Parliament, and the Conservative government was forced to undertake revisions of British export control legislation. Meanwhile, in its election manifesto, New Labour committed to going further by adopting an ‘ethical’ and more transparent arms export control policy and promoting an EU code of conduct (see Labour Party, 1997). This dovetailed with what NGOs were calling for. Upon entering office, Labour had tied substantial political prestige to reforming British export control policies. NGOs, long calling for a moral order to the arms trade and an EU code of conduct, took a watchdog position, prepared to hold the government’s feet to the fire upon deviance from its promises. Once in power, Labour dutifully began pursuing an EU code of conduct.

While the UK shift was very much a legacy of the ‘arms to Iraq’ scandals, with the security concerns and outright embarrassment that the scandals involved, it was also enabled by the broader norm cascade that had gained momentum since the mid-1990s. By now, state security concerns had given way to a stronger human security formula, and a hotchpotch of international processes addressing various aspects of the arms trade were under way, as were discussions on a global code of conduct. Labour, therefore, was surfing an already-crowded wave. Both in the UK, Europe and elsewhere, NGOs have been attributed substantial credit for putting arms export control on the agenda during this period — in the EU, by naming and shaming governments on the basis of the Luxembourg/Lisbon criteria (Joachim and Dembinski, 2011; Stavrianakis, 2010). Furthermore, the UK policy shift has been linked to the personal preferences of Sir Robin Cook, Foreign Secretary for the Labour administration during 1997–2001 and a fervent supporter of an ‘ethical’ EU code of conduct (Bauer, 2003: 370). This supports a constructivist interpretation of regime adoption/change emerging as norm entrepreneurs exploit windows of opportunity stemming from changes in the external environment.

Yet, the UK shift was anything but monocausal, and while Labour’s initiative dispatched rigorous moral fabric, it was closely linked to material interests. First, it has been connected to the political ambitions of Tony Blair in advancing the CFSP, which would require a more efficient defence-industrial base, with the common export regulations this necessitated (Calleo, 2001: 299–336). Even more so, previous studies have shown that economic incentives and a strategy to create a level playing field for British industry were just as important in New Labour’s initiative as norms and restraint, if not more important (see Bauer, 2003: 416; Cooper, 2000). Arguably, the UK attempted to ‘upload’ its revised national guidelines to the EU level in order to minimize potential competitive disadvantages to its arms industry, by subjecting competitors to equally stringent rules. Subsequent British export practices and actual alterations to British export control legislation support this interpretation (Cooper, 2000; Hansen and Marsh, 2015). As Cooper (2013) shows, the ‘ethical’ arms export policy embraced by New Labour was even less restrictive than Labour policy on these issues in the 1980s.

With the other member states on board the initiative for their own, heterogeneous reasons, the success of New Labour’s ambitions ultimately rested on the ability to convince the still-reluctant French government to change its view on a code. The UK strategy was to come to agreement on an Anglo-French draft based on revised UK guidelines, which could then be approved by the other member states (Davis, 2002: 101). This was a calculated tactic to minimize French opposition (Hagelin et al., 1999). Like New Labour, the new French government (a socialist government elected in May 1997) wanted to avoid the embarrassing export scandals of the previous government. However, as shown by Béraud-Sudreau (2014: 24), the ultimately altered French position was not so much connected to a commitment to moral or security concerns as to a belief in the French government that the regime could serve as a useful tool to improve market conditions for the French arms industry and Franco-British collaborative projects, and level the playing field vis-a-vis its main competitor — the UK. Arguably, both the British and French governments also expected that a code could be turned into a useful instrument to curb perceived US attempts to control and dominate the arms export policies of its allies (Calleo, 2001: 320; Joachim and Dembinski, 2011: 1159). This indicates that it was the interest overlap between France and the UK — top-ranking EU arms exporters — in a level playing field that enabled the adoption of the Code. Substantiating this reasoning, the Franco-British draft proposal for a code proposed committing member states ‘to the maintenance of a strong defence industry which is a strategic part of their industrial base as well as their defence effort’ (UNESCO, 1998). This language was later removed following objections from other member states (Bauer, 2003: 104). Through the Code, 15 member states declared themselves ‘determined to set high common standards which should be regarded as the minimum for the management of, and restraint in, conventional arms transfers by all EU member states’, but only while ‘acknowledging the wish of EU Member States to maintain a defence industry as part of their industrial base’ (Council, 1998). The latter point, reformulated from the previously cited initial suggestion, was a crucial addition for both the UK and France. However, with the reformulation and the generally unfavourable French position towards the Code, the result was that the remainder of the agreement was watered down to secure French signature.

Due to French objections, some of the more restrictive and less ambiguous options included in the draft texts following intensive NGO lobbying and supported by several governments (Germany, the Netherlands, Ireland and the Nordic member states) — for instance, tighter criteria on human rights and development (criteria 2 and 8) — were excluded. Some ministers, including the Irish foreign minister, had promoted a ban on arms sales to human rights violators; this was unacceptable to France, which had considerable arms exports to the Middle East (BASIC, 1998). Arguably, in the final document, ‘all the weaker options were chosen’ to get France on board (Davis, 2002: 101), and the adopted Code looked little like that envisaged by NGOs and the most favourable governments (Walker, 1998).

To forge consensus, several issues were intentionally avoided, for example: how much could a receiver import before it overstepped its legitimate security needs (criterion 8)? The Code also got an extremely high risk threshold. The first four criteria were considered the most important, and were meant to trigger licence denials. Yet, two of these, criteria 2 and 4, stated that arms export would be refused only in case of a ‘clear risk’ that the armaments ‘might’ be used for human rights abuses or aggressively against another country (Council, 1998). To many, the use of ‘clear risk’ instead of ‘risk’ implied an unreasonably high risk threshold because ‘the danger of this approach is that countries may be able to authorize arms transfers even if there is some risk of misuse of the weapons to be supplied’ (Yihdego, 2009: 288). In addition, risk assessments are likely to be influenced by who assesses and by the fact that assessing risk is no exact science. Furthermore, the latter four criteria were only to be ‘taken into account’ and ‘considered’, and thus had an even obscurer obligation. This reflected intra-EU divergences; as an example, not all states were equally supportive of criterion 8 due to its potential harm to export interests in the Middle East. Hence, the criterion was formulated to permit an interpretive latitude that satisfied all audiences, and it was symbolically placed among the less obligatory of the criteria. The great interpretive leeway resulting from the Code was exactly the reason COARM started compiling a regularly updated User’s Guide in 2003. The User’s Guide contains agreed ‘best practices’ on the interpretation of the criteria. While the User’s Guide provides guidance and hence some precision, it has consistently been interpreted as a non-binding document. ¹⁴

Competing rationales and continuing worries about sovereignty and collaboration costs were also accommodated by adopting the Code as a politically binding Council Declaration. Through the Code, the Council ‘recognized’, ‘wished’, ‘noted’ and ‘acknowledged’; nowhere did it agree, approve, instruct, mandate or decide. The Code was also exempted from Commission infringement proceedings and European Court of Justice (ECJ) jurisdiction (Lustgarten, 2013). A Council Declaration did not even constitute the strongest tool of the Council under the CFSP; although neither a ‘joint action’ nor a ‘common position’ would have transposed directly into national legislation, they would have entailed a more binding language and implementation requirements. In addition, Article 346, preserved by the Maastricht and Amsterdam treaties, still provided a highly present escape clause.

Notwithstanding unwillingness to establish formal enforcement mechanisms at the EU level, the Code introduced informal operative provisions for bilateral consultations and transparency. Bilateral consultations were to take place when a state considered granting an export licence that was ‘essentially identical’ to a licence that another member state had denied within the past three years (to prevent undercutting), and member states were requested to share data on exports and denials. In addition, informal dialogue and peer pressure through COARM may impact decision-making. Assessment of the merits of these mechanisms is beyond the scope of this article. The critical point for this chronology, however, is that softer and ambiguous arrangements were deliberately preferred to binding and more precise ones due to a configuration of heterogeneity and resistance attributable to material interests.

The Common Position: a final bid at a level playing field

As examples of dubious arms transfers continued to hit the headlines, a sentiment prevailed in NGOs, the EP, national parliaments and academia that the Code was too ambiguous for interpretation to converge and that its non-legal status rendered it inadequate to ascertain compliance (Amnesty International, 2004; Lustgarten, 2013; Saferworld, 2008; Yihdego, 2009). Hence, in the years following its adoption, NGOs and the EP in particular exerted pressure on member states to reduce ambiguity and make the Code legally binding. Following a 2004 EP initiative, in mid-2005, the Council completed a draft document for a legally binding common position aimed at strengthening the EU’s export control policy beyond the capacity of the Code (Council, 2005). The Code was replaced by the legally binding Council Common Position 2008/944/CFSP (CP) in December 2008. The CP advocated a strong discourse of restraint, stating that ‘Member States are determined to set high common standards which shall be regarded as the minimum for the management of, and restraint in, transfers of military technology and equipment’ (Council, 2008). This step, which suggested member states were finally committing to restrictive arms export policies, was warmly welcomed by export control advocates. At face value, the CP certainly echoed the age-old proposals of civil society, parliaments and the more restrictive member states. Yet, the process towards adopting the CP highlights how material interests sat at the core of major arms exporters’ perceptions of the CP.

It took three-and-a-half years to adopt the CP. The long hold-up — and the eventual consent for the CP — was connected to the contested arms embargo on China. Several member states had long promoted lifting the embargo because it prevented exploiting commercial opportunities with China (Erickson, 2013; Gupta, 2013). However, absent consent on lifting the embargo, France conditioned its consent to the CP on other states’ willingness to lift the embargo (Depauw, 2010; Gupta, 2013). Arguably, the adoption of the CP was enabled because France rested its case regarding the China embargo by the end of 2008, partly because domestic developments in China had made lifting the embargo impossible (Depauw, 2010). French NGOs in particular highlight this link, claiming that they were able to exploit this window of opportunity to pressure the government into accepting the CP. ¹⁵

Yet, the course of events, previous research and interviews suggest that the embargo issue and NGOs seizing a window of opportunity cannot alone account for the altered French position. Rather, substantial evidence suggests that the new French position on the CP connected intimately to the European Defence Package (EDP), launched by the Commission in 2007. The EDP included the ‘Directive on intra-Community arms trade’ (ICT Directive), which would lower intra-EU trade barriers for military equipment so as to facilitate intra-EU defence collaboration, and potentially increase the external competitiveness of European defence industries. As such, it played into the model that industry had proposed since the early 1990s, a model now widely supported by export-dependent member states. It also played into long-standing attempts by the Commission to shrink the issues covered by Article 346 and open the defence market to greater competition. At this point, the financial crisis beginning in 2008 had led to defence-industrial austerity measures, intensified government engagement in arms export promotion and renewed pressures to rationalize European defence industries (Heidenkamp et al., 2013: 82–89; Jackson, 2013: 207–208). Last, but not least, France was the only remaining bottleneck to adopting the CP, but with a large defence industry that would benefit from a liberalized defence market, France was also a proponent of the ICT Directive.

The empirical evidence is not conclusive regarding the exact manner in which the ICT Directive triggered French consent on the CP. Depauw (2010) argues that the EP pressured France to accept the CP by conditioning the adoption of the ICT Directive on the adoption of the CP. Indeed, minutes from the meetings in the EP in July–December 2008 do show that the EP, prior to voting in favour of the ICT Directive, pressured the French presidency of the Council on the issue (see European Parliament, 2008). In contrast, Béraud-Sudreau (2014) connects the altered French position to age-old, French sovereignty concerns. The French government and export control authorities arguably worried that the Commission, after acquiring a role in defence procurement and intra-EU trade with the EDP, would build on case law to extend its competencies into extra-EU arms export controls. To sections of the French government and bureaucracy, adopting the CP arguably became a way of enshrining national control over export licensing and leaving no ‘empty spot’ for the Commission to fill.

Whether emphasizing EP pressure, the threat of supranational competence or a combination, ample evidence points to the importance of the EDP and material incentives in the account of the French consent to the CP. Substantiating this interpretation, the French Ministry of Foreign Affairs states that ‘the existence of the [CP] is gradually creating the level playing field between European exporters that French manufacturers have, quite legitimately, been calling for’ (MFA France, 2015). In addition, as explained by a former French export control officer when asked about the French consent to the CP, ‘as a big exporter, France sought to ensure that competitors were subject to the same rules’. ¹⁶ Accordingly, long-standing desires for a level playing field influenced the adoption of the CP.

The preceding has demonstrated that France, at least in part, saw the CP as a means of promoting its interests in the arms trade. In comparison, the Netherlands, long promoting more restrictive export controls, has advocated more of a discourse of restraint. Beyond doubt, cross-country variation on the matter can be spotted. Moreover, accounting for all states’ views of the CP extends beyond the scope of this article. What is worth dwelling on, however, is the fact that the CP is formulated in a way that provides more room for considering material interests in the arms trade than some of the literature on arms trade regulation gives the impression of. According to criterion 5:

Member states shall take into account the potential effect of the military technology or equipment to be exported on their defence and security interests … while recognizing that this factor cannot affect consideration of the criteria on respect for human rights and on regional peace, security and stability. (Council, 2008)

This wording, preserved since the Code, can be read as permitting defence-industrial considerations in licensing decisions, provided criterion 2 (human rights) and criterion 4 (regional peace, security and stability) are not undermined. This is remarkable as it seems to heighten the importance of defence-industrial considerations at the expense of the remaining criteria. Furthermore, Article 10 of the CP states that ‘Member States, where appropriate, may also take into account the effect of proposed exports on their economic, social, commercial and industrial interests’, although ‘these factors shall not affect the application of the [eight] criteria’ (Council, 2008). The UK’s Consolidated Criteria, updated in March 2014, illustrates the room for selective interpretation of the CP and exemplifies a rather permissive interpretation of Article 10. The Consolidated Criteria notes that:

The Government will [following Article 10] continue when considering licence applications to give full weight to the UK’s national interest, including … the potential effect on the UK’s economic, financial and commercial interests … the potential effect on any collaborative defence production or procurement project with allies or EU partners [and] the protection of the UK’s essential strategic industrial base. (Parliament, 2014)

The wording - both in the CP and the UK Consolidated Criteria - surely echoes the disputed formulations once removed from the initial Franco-British draft Code following fellow member states’ objections.

A notable change in vocabulary has certainly taken place with the changed legal status: where the Code ‘recognized’, ‘wished’, ‘noted’ and ‘acknowledged’, member states under the CP ‘intend’ and are ‘determined to’. Furthermore, the CP has a stronger obligation than the Code, as ‘should’ has been replaced with ‘shall’; now, exports of equipment on the common military list ¹⁷ ‘shall’ be refused if violating criteria 1–4, while criteria 5–8 are still to be ‘considered’ or ‘taken into account’. Article 29 of the Treaty of the European Union, which lays out the legal nature of common positions, also states that ‘member states shall [emphasis added] ensure that their national policies conform to the [CP]’. Yet, member states retain decision-making capacity and the discretion to decide the extent to which the CP is transposed into domestic law. The CP also sustains the high risk thresholds and ambiguities of the Code.