No wings attached? Civil–military relations and agent intrusion in the procurement of fighter jets

Abstract

This article studies civil–military relations in defence procurement. Applying insights from the principal–agent model, we argue that decision-making about defence procurement is inherently vulnerable to agency problems. Given the technical nature of these dossiers, governments and parliaments are often heavily dependent upon military expertise, creating leeway for defence administrations to steer decision-making towards their preferences. By means of a case study of the replacement of the F-16 fighter jets in Belgium, we examine whether and how complex defence procurement dossiers allow for exploitation of expertise through strategic information management from the defence administration to the Minister of Defence. In addition, empirics reveal a to date unexplored phenomenon of agent intrusion. It captures the situation in which an agent takes a prominent formal advisory position within the decision-making apparatus of its political principal, providing additional means to outplay its information advantage over the principal in favour of its own interests.

Introduction

In March 2017, the Belgian government distributed a Request for Government Proposal (RfGP) for 34 fighter jets to five state agencies, in prospect of replacing its ageing F-16 fighter jets, which were considered to be out of use by 2023 (Belgian Ministry of Defence, 2017). By holding an open competition, the government attempted to avoid doubts over its final decision, such as there were in Canada in 2010 (Hoeffler and Mérand, 2016; von Hlatky and Rice, 2018). Yet, in March 2018, a political crisis developed when an opposition party disclosed email exchanges between the military staff, revealing attempts to conceal studies which indicated that the lifetime of the F-16s could potentially be extended beyond 2023. The conversations furthermore showed that several leading military commanders openly favoured the American-built F-35 Joint Strike Fighter as replacement, which undermined the unprejudiced procurement process the government officially committed itself to. The revelations led to accusations that the Minister of Defence did not have control over his administration (De Morgen, 2018: 6). The Belgian case strikingly illustrates the contentious nature of large defence procurement decisions, raising questions about civil–military relations and the capacity of the military to push its preferences.

This article studies civil–military relations in defence procurement decision-making, guided by the question ‘How can we explain agency problems in defence procurement?’ We first highlight why large defence procurement dossiers are inherently contentious and come with a considerable risk of agency problems, or undesired agency behaviour. We then apply the principal–agent model, a core framework in the study of civil–military relations (e.g. Auerswald and Saideman, 2014; Feaver, 1998; Lagassé and Saideman, 2017, 2019), to the case of the F-16 replacement in Belgium. This typical case of controversial defence procurement allows for establishing whether and how complex defence procurement dossiers allow for exploitation of expertise. Empirics show that the defence administration profited from its expertise vis-à-vis the Minister of Defence to strategically withhold information in order to steer the governmental decision-making process towards its own preferences.

Our case study furthermore reveals a to date unexplored phenomenon, which we introduce as agent intrusion. It captures the situation in which an agent (here: military staff) takes a formal advisory position within the decision-making apparatus of its political principal. We argue that instances of agent intrusion increase the risk of agency problems. It gives the agent additional means to outplay its information advantage over the political principal in favour of its own interests. We conclude that this risk is particularly likely in technical and confidential decision-making dossiers wherein the Minister of Defence is highly dependent upon the expertise of its military staff and administration, such as in large defence procurement decisions.

Civil–military relations

The risk of agency problems, defined as behaviour by the agent which is not entirely in line with what the principal originally envisioned (Hawkins and Jacoby, 2006: 212), is a key concern in civil–military relations. It is at the core of a literature which is largely founded on Huntington’s The Soldier and the State (1957) and Janowitz’s sociological The Professional Soldier (1960). At the origins of this civil–military relations scholarship is the tension between creating a military which is strong enough to guarantee security of the civil community by which it was created, and ensuring that it does not become so strong that it can enforce its will on that civil community. While initially developed in a Cold War context with the threat of coups in emerging or consolidating democracies in mind (e.g. Cottey et al., 2005; Nordlinger, 1977; Rouquié, 1987), civil–military scholars gradually expanded their focus to established democracies (Feaver, 1999).

Central to the civil–military relations literature is the hierarchical relationship between the civilian leadership and the military organization. It implies that ‘[r]egardless of how strong the military is, civilians are supposed to remain the political masters’ (Feaver, 1999: 215). The principal–agent model has become a dominant framework within this field as it captures the relationship between such a hierarchical superior (the principal) and a servant (the agent) (Auerswald, 2017; Auerswald and Saideman, 2014; Avant, 1993; Desch, 2001; Feaver, 1999). In civil–military relations, roughly stated, civilian leadership delegates the implementation of (parts of) its defence policy to the military. Through this delegation from a civilian (or political) principal to a military agent, however, the problem of agency arises (Avant, 1993: 413). Questions of democratic control are hence at the core of the civil–military relations literature (Feaver, 1996).

Most civil–military relations studies focus on decisions about military action per se, with democratic control of the use of force being a central area of interest (Auerswald and Saideman, 2014; Feaver and Gelpi, 2003). Over the years, a literature on legislative control of the use of force and military troop deployments has also developed (e.g. Dieterich et al., 2015; Fonck and Reykers, 2018; Peters and Wagner, 2011). This increasingly prominent research agenda often refers to the underlying civil–military relations concept (Auerswald, 2017; Lagassé and Saideman, 2017; Wagner, 2006). Meanwhile, the focus on democratic control of the use of force leaves many aspects of civil–military relations in (Western) democracies unexplored. Bruneau and Matei (2008: 910) highlighted how it impedes ‘understanding the larger and more complex relationships concerning democracy and security forces’, thereby referring to the wide spectrum of military activities. They stressed the need to extend the conceptual breadth of the civil–military relations concept. We add that decision-making about large-scale defence procurement dossiers can equally become victim to agency problems and strategic agency behaviour.

Defence procurement and civil–military tension

Civil–military relations in defence procurement remain largely unexplored. Defence procurement, defined as ‘the process by which states acquire goods and services required by their armed forces’, entails the purchase of both standard civilian products and major weapon systems (Uttley, 2018: 73). The lack of attention is remarkable, as Desch (2001: 6) argued that ‘the issues likely to divide civilian from military leaders are the larger military-political questions, including the use of force, budgets and procurement’. Although the literature on corruption in military spending emphasizes problems of secrecy (Gupta et al., 2001; Rendon and Rendon, 2016), there is little to no research on agency problems in major weapon acquisitions.

According to Uttley (2018: 74), defence procurement generally comes with three challenges: choosing the type of equipment in order not to become inferior to rivals, deciding from whom to buy this equipment to ensure stability of supply, and deciding upon the way in which to acquire this equipment to ensure the timely delivery and quality of the material. Meanwhile, there is an inherent tension between ‘a state’s desire to ensure its security and sovereignty on the one hand, and to deploy its financial resources to greatest domestic effect on the other’ (Uttley, 2018: 77). This balance between guaranteeing a reliable weaponry system and securing sufficient economic return has attracted considerable scholarly attention, mainly from an economics perspective (e.g. De Fraja and Hartley, 1996; Rogerson, 1994, 1995). It forms the slippery slope upon which governments balance in large defence procurement dossiers.

One aspect makes defence procurement dossiers particularly vulnerable to agency problems: the government’s dependence upon military expertise. Similar to decisions about military action (Fonck and Reykers, 2018), defence procurement dossiers are characterized by high specificity and technicality. The government, which should ultimately make the decision, generally lacks the expertise to make a well-informed decision. By consequence, the defence administration’s know-how gives it a considerable information advantage over its political superior. This puts the government, and particularly the Minister of Defence who is politically responsible, in a relationship of dependency vis-à-vis its military agent.

The contentious politics of fighter jet acquisitions

Fighter jet acquisitions have repeatedly proven to be the defence procurement dossiers par excellence. In this section, we argue that these dossiers have a high likelihood of becoming contentious, as illustrated by debates in several Western democracies about the extent to which military alliance preferences hindered a fair and unbiased competition.

Vucetic and Tago (2015) highlighted how arms procurement decisions, and fighter jet acquisitions in particular, are often shaped by hierarchical inter-state relations. They showed that ‘[a]ll else being equal, the more a state is embedded in US security and economic hierarchies, the more it is likely to buy US-made military hardware’ (Vucetic and Tago, 2015: 102). Following Hartley (2007), they emphasized that defence contractors and consortia ‘cannot function without the financial and political support coming from ‘home’ governments’ (Vucetic and Tago, 2015: 105). Referring to the buy-in of several states in the American-led Joint Strike Fighter (JSF) weapons development programme in the 2000s, von Hlatky and Rice (2018: 20) similarly highlighted tension between ‘pressures of being a member (or ally) in “good-standing” and developing the capacity to continue to operate alongside other alliance partners, not just the US’. Arms procurement is hence not just a balancing act between reliability and economic concerns, but also a political process linked with alliance and partnership considerations.

In many of the countries which over the past years decided to replace their ageing fighter jet fleet with the American-built F-35 JSF, discussions arose about the extent to which military alliance preferences may shape the outcome of procurement processes to the detriment of more open and fair competition. The most telling example is the Canada case, which received ample scholarly attention (Byers and Webb, 2011; Hoeffler and Mérand, 2016; von Hlatky and Rice, 2018; Vucetic, 2016). In 2010, the government decided to replace its ageing CF-18 fleet with over sixty F-35 fighter jets. Yet, in March 2011, the minority government faced a vote of no confidence following a report which indicated it had underrepresented the total acquisition costs (Hoeffler and Mérand, 2016). The Auditor General concluded that the government was unable to run a fair competition due to its involvement in the aforementioned JSF programme (Office of Auditor General of Canada, 2012: 2).

Fighter jet dossiers in the Nordics have been subject to similar debates (Ringsmose, 2013). In Norway, an intense debate arose around the 2008 decision to acquire F-35 fighter jets. The credibility of the seemingly open competition initiated by the Stoltenberg government in 2005 was undermined by the release of a set of WikiLeaks cables from the US Department of State in 2010. The cables fed rumours about the competition in fact being ‘a wilting fig leaf for the “real” deal, which was making good on the F-35 precommitment’ (Vucetic and Rydberg, 2015: 62). Not only have there been accusations that its decision was shaped by the earlier participation in the JSF weapons development programme; there were also strong indications of the USA having repeatedly interfered in the decision-making process (Vucetic and Rydberg, 2015). Finally, while the Danish Court of Audit (Rigsrevisionen) concluded that the government’s final financial assessment which led to the acquisition of 27 F-35 fighter jets in June 2016 was made on solid grounds, it also raised concerns about the potential underestimation of the jets’ flight hours and life-cycle costs (O’Dwyer, 2017; Rigsrevisionen, 2017: 3). Meanwhile, debates on the fairness of competition have in many countries been strategically fed by procurement contenders that saw themselves outplayed but sought to uphold their reputation. This was, for instance, the case in the Danish replacement process, where Dassault and Eurofighter argued they were disadvantaged by the government’s procurement prerequisites, although government and private auditing later concluded no such evidence existed (Ringsmose, 2013: 96–97).

While much of the scholarly analyses of these dossiers have looked into governmental preferences (Ringsmose, 2013; von Hlatky and Rice, 2018; Vucetic and Rydberg, 2015; Vucetic and Tago, 2015), strategic agency behaviour by defence administrations has to date not received much attention. This is remarkable, as it is ultimately the national air force which has to operate these newly acquired weapon systems; hence they can be expected to have autonomous preferences. We will highlight how key principal–agent assumptions allow for gaining insights in agent strategies to shape decision-making in large defence procurement dossiers, such as the acquisition of fighter jets.

Principal–agent theorization and defence procurement

We look at the delegation relationship from the Minister of Defence (i.e. the political principal) to the defence administration (i.e. the military agent). As we have argued, implementation of defence policy not only implies the use of military force or the conduct of military activities; the Minister equally delegates responsibilities to its administration in highly technical procurement dossiers. In theory, the civilian leadership, as the political principal, is expected to formulate a strategic vision that sets out the contours of the procurement process. This implies that the selection and evaluation criteria of defence acquisitions are predefined by the Defence Minister or their government.

The actual implementation of the procurement process, in turn, can be delegated to the military administration, again subject to final approval by their political masters. Delegated tasks may include assessing the necessity to acquire new arms or equipment, defining technical criteria for the drafting of a tender framework which fulfil the attainment of the above mentioned political benchmarks, or evaluating (parts of) bids in a competitive tender. This delegation relationship, however, creates tensions. On the one hand, the armed forces possess most expertise on the technical attributes of military warfare and will have operational responsibility over the new equipment, which makes it self-evident that military commanders are tasked to participate in the implementation of the procurement process. On the other hand, the military remains hierarchically subordinate to their political principals and their political preferences, ‘even if civilians are woefully underequipped to understand the technical issues at stake’ (Feaver, 1999: 216).

Principal–agent theorization identifies several factors which increase the risk of agency problems (Hawkins and Jacoby, 2006; Nielson and Tierney, 2003). This article focuses on three factors which equally apply to defence procurement decision-making. First, the preference constellation between a principal and an agent is considered crucial for explaining agency problems (Miller, 2005). As principals and agents are assumed to seek goal attainment, goal conflict raises the risk of opportunistic agency behaviour (Shapiro, 2005). Goal conflict is particularly likely to be present in large-scale defence procurement dossiers, because the government has to take into account considerations about military efficiency and reliability, alliance and partnership considerations, as well as to calculate the financial costs and potential economic returns. Meanwhile, it is directly accountable to parliament, its own political principal (Bergman et al., 2000). Governments and ministers can be expected to keep their hands and eyes off toxic procurement dossiers, as they constantly face the risk of being sanctioned, either electorally or with a motion of no confidence. They will therefore cautiously guard being perceived as objective and neutral. The military, to the contrary, is much less directly accountable. Because the military ultimately has to use these weapon systems, it can be assumed that its preferences in a given dossier are more outspoken, mostly driven by efficiency or reliability concerns. Interoperability concerns, continuing existing partnerships and the ambition to participate in high-end operations are in that sense also likely to shape their preferences.

Second, information between principals and agents is not always symmetrically divided (Pollack, 1997: 108). Situations in which the agent has a considerable information advantage over the principal, resulting from an agent’s expertise, are assumed to create opportunities for strategic information behaviour by that agent (Hawkins and Jacoby, 2006). Such dynamics are particularly likely in defence procurement decision-making, which is often characterized by high degrees of technicality and low levels of transparency.

Third, principal–agent scholars generally assume that the risk of agency problems is greater as the distance between a principal and agent increases (Bergman et al., 2000; Nielson and Tierney, 2003). Although this article focuses on the relationship between a minister and his or her administration, we recognize that this focus only partially captures the democratic delegation process. In parliamentary democracies, there is a so-called ‘chain of delegation’ from the voter (represented by the parliament as the ultimate principal) to the government and its ministers (including their cabinets ¹ ), who at their turn act as the proximate principals of their administration (Bergman et al., 2000; Elsig, 2011; Nielson and Tierney, 2003). In other words, civil–military relations in parliamentary democracies are defined by two connected principal–agent relationships: (a) delegation from parliament to the Minister of Defence, and (b) delegation from the minister to the administration. Such chains of delegation are assumed to create more opportunities for strategic information behaviour by the agent and come with an increased risk of information loss (Elsig, 2011: 5). In countries with a strong executive dominance over defence policy-making, such as Belgium, this risk is even greater, as parliaments have limited powers or instruments to force the Minister of Defence to provide full information (Fonck and Reykers, 2018; Lagassé and Saideman, 2019).

Two expectations about potential information loss can be formulated from these insights. First, in case of diverging preferences, the proximate principal (i.e. the minister) risks receiving partial or biased information, due to strategic information management by the military administration. And second, the ultimate principal (i.e. parliament) also risks receiving only partial or biased information, due to strategic information management by either the military administration or the Minister of Defence. Given the focus on the delegation from a minister to his or her administration, we only address the first expectation. Meanwhile, principal–agent scholars generally agree that a principal can install control mechanisms in order to anticipate these risks (e.g. Hawkins and Jacoby, 2006; Nielson and Tierney, 2003; Pollack, 1997). The mere presence of preference divergences and information asymmetries between a minister and his or her administration, or between parliament and government is hence not always sufficient for explaining agency problems.

Based on the case of the Belgian F-16 replacement, we furthermore introduce the concept of ‘agent intrusion’. It captures the situation in which agent members take a formal advisory position within the decision-making apparatus of the political principal. We will show that agent intrusion was possible by military staff taking a prominent position in the Minister of Defence’s cabinet, which comprises his advisory staff. In this way, the defence administration had the chance to strategically exploit its information advantage. It seems to have created an opportunity to hide information from its minister and steer decision-making in the direction of its own preferences.

The idea of agent intrusion connects, among others, to theorization about bureaucratic politics (Allison, 1969), which highlights bargaining between individuals and groups with different powers and preferences within the government. Also related is Janis’s (1972) work on groupthink, which shows how conformity thinking among bureaucrats within decision-making bodies may have a profound (and pernicious) impact on policy outcomes. Finally, with their concept of cultural contestation Barnett and Finnemore (1999: 724) have equally described a problem of ‘pockets of autonomy and political battles within the bureaucracy’. What these models have in common is their attention for the crucial role of bureaucratic actors within decision-making bodies, emphasizing the importance of political games between subgroups and subcultures in a decision-making process. Yet, they do not look into the specific phenomenon of how actors affiliated to an agent body strategically exploit information asymmetries or may even ‘infiltrate’ the decision-making apparatus of the principal; neither do they benefit from the accountability insights that a principal–agent perspective might bring. This is where the added-value of the principal–agent approach and the newly introduced concept of ‘agent intrusion’ lies.

Case: the F-16 replacement in Belgium

Belgium’s F-16 replacement dossier is a typical case for exploring civil–military relations in defence procurement and explaining how agency problems can occur. First, the replacement of fighter jets is the defence procurement dossier par excellence. Second, the Belgian defence administration is almost entirely composed of military staff, which neatly fits the concept of a military agent. Third, as will be shown, the leaked emails scandal of March 2018 revealed that there clearly was undesired agency behaviour by the defence administration. Before applying the aforementioned principal–agent framework, we briefly outline the history of defence procurement in Belgium and provide a background to the replacement dossier and the leaked emails scandal.

Empirically, we build upon the leaked email conversations as reported in the press and in parliamentary debates and the findings of two audits that were commissioned by the Minister of Defence. We also use the minutes of the parliamentary hearings that were organized in the aftermath of the leaked emails scandal, where both the Minister of Defence and involved actors from the defence administration were summoned to inform the parliament’s Defence Committee. ²

Belgium’s history of controversial defence acquisitions

Much of the public and political suspicion surrounding the tender followed from Belgium’s rather contentious history of defence procurement. A key episode in this history is the so-called Agusta/Dassault affair, which erupted in the early 1990s and revealed corruption in the acquisition of Agusta A109 combat helicopters by the Martens government in 1988 (e.g. Cools and Pashley, 2015; Maesschalck and Van de Walle, 2006). In July 1991, a judicial investigation discovered that several prominent socialist members of government and officials had received bribes from aviation companies Agusta and Dassault to ensure that Belgium would buy the Agusta A109 helicopter (and to get a deal on refitting the F-16 jets). The case shook up Belgian politics and led to a series of forced resignations, including from the Minister of Interior Affairs, the Minister of Defence and then NATO Secretary-General Willy Claes, given his role as Minister of Economic Affairs during the acquisition process.

The Agusta/Dassault affair still shapes public and political debate in Belgium about large defence acquisitions (Cools and Pashley, 2015). These past events put additional pressure on governments to carefully guard objectivity and neutrality in defence procurement decision-making. The public tender for the replacement of the F-16 fighter jet fleet that was distributed by the government, as well as the repeated reassurance that the final decision would be based upon an open competition (see below), should be interpreted against this background.

Background to the F-16 replacement

The case of the Belgian replacement of its F-16 fleet reads as a chronicle of a controversy foretold. Preparations for buying a potential replacement aircraft started in 2014, when the Defence staff sent out a Request For Information (RFI) (entitled ‘Preparation Survey’) to government agencies of several EU and NATO partners (Belgian Ministry of Defence, 2014), while the incoming Michel government confirmed in its coalition agreement ‘to maintain a fighter jet capacity on a long term basis given the replacement of the F-16’ (Belgian Government, 2014: 208). Based on this, a ‘Strategic Vision’ was presented by the Minister of Defence, which served as a guideline for Belgian defence policy and investments until 2030 (Belgian Ministry of Defence, 2016). In the process leading up to the Strategic Vision, the defence administration was tasked to investigate several alternative scenarios, including a life extension of the current F-16 fleet, the use of a mixed fleet of old and new aircraft and the acquisition of second-hand F-16s. These options were presented by the military via an internal study and discussed between the different ministerial cabinets of the federal government (Belgian Ministry of Defence, 2015). As the study concluded that a life extension was not opportune and too costly, the government decided to put these options aside in December 2015 (CRIV 54 COM 862: 33–34). The Strategic Vision, accepted by the government in June 2016 and published in January 2017, therefore announced the full replacement of the F-16 fleet with 34 new fighter jets by 2023. In response, the Minister of Defence delegated to the defence administration the identification of selection criteria, resulting in a Request for Government Proposal (RfGP) which was distributed on 17 March 2017 (Belgian Ministry of Defence, 2017). Crucially, the entire procurement process was built upon the assumption that the F-16 fighter jet fleet would reach its maximum life expectancy of 8000 flying hours in 2023.

As the tender proceeded, three defence consortia (Lockheed Martin, British Aerospace and Dassault) submitted an extensive proposal (VRT News, 2017). ³ Lockheed Martin and British Aerospace responded directly to the RfGP, while Dassault proposed an alternative partnership outside the tender (Belgian Ministry of Defence, 2018b). Importantly, the planning process took place following similar replacement decisions by several of Belgium’s military partners. The Netherlands, a long-standing military partner of Belgium, already in September 2013 decided on the acquisition of Lockheed Martin’s F-35 JSF as a replacement for its F-16s. A similar decision was taken in Denmark (2016) and Norway (2008), the two other partners with Belgium in the F-16 ‘buyers consortium’ in 1979 (Ringsmose, 2013). Military partnership considerations were therefore expected to shape preferences towards the F-35. Also the scholarly literature seemed to share this expectation, with von Hlatky and Rice (2018: 26) suggesting that ‘European countries that purchased the F-16 in the mid-1970s (Belgium, Denmark, Norway and the Netherlands) are top candidates for the F-35’.

The leaked emails scandal

Based on the government’s engagement to replace the F-16 by 2023, the Material Resources department of the defence ministry (DG-MR), charged with the maintenance of the current fleet, commissioned further studies on the life expectancy of the Belgian F-16s in the context of a projected phasing-out of the aircraft in between 2023 and 2028. Against expectations, F-16 manufacturer Lockheed Martin provided two studies, outlining new theoretical models which suggested that the current life expectancy of the structural parts of the Belgian F-16 fleet could hypothetically be extended to the period 2029–2036 by means of a so-called ‘soft SLEP’ (Service Life Extension Program). These projections were the result of a new calculation method, based on equivalent flight hours, which recalculate a plane’s actual flying hours based on the intensity of the usage of the plane. For the Belgian fleet, it was calculated that its moderate usage would allow an extension up to 9500 hours instead of an initial service guarantee of 8000 flight hours. In other words, theoretically, operational availability could be extended by a period of 6–8 years, which questioned the urgency of a replacement (CRIV 54 COM 871: 1–2; De Standaard, 2018c: 2–3; Lockheed Martin, 2017).

This finding was seemingly in dissonance with the decision to replace the fighter jets by the time the fleet was supposed to reach its forecasted 8000 flying hours. Therefore, within the DG-MR, the decision was made not to forward this new information to the Minister or his cabinet, as it was considered irrelevant in light of the earlier-made engagement to replace the fleet in the period 2023–2028 (CRIV 54 COM 871: 1–2). According to later statements, the Director-General of MR asserted he was not aware of the existence of these reports, nor of the decision not to forward these new studies. Despite many questions within Parliament, as to whether the Minister had any information or studies on a potential prolongation of the life expectancy of the Belgian F-16s, the Minister consistently replied he was not aware of any such information.

On 20 March 2018, however, the socialist opposition party (sp.a) leaked email exchanges between military staff, which revealed that attempts were made to conceal the Lockheed Martin studies. The conversations furthermore proved that several leading military commanders favoured the American-built F-35 JSF candidacy, which undermined the neutral procurement process the government officially committed itself to. Considering it ‘a serious judgement error at least’ (CRIV 54 COM 848: 5) the Defence Minister commissioned an internal investigation by the military administration as well as an ‘external audit’ conducted by the Federal Internal Audit (FIA) service ⁴ to determine who was aware of the particular studies, and why the information had not reached him. While both investigations concluded there was no direct proof of deliberate manipulation of the procurement process itself, they nonetheless highlighted problems of strategic information management within the defence administration (Belgian Ministry of Defence, 2018a; FIA, 2018).

In the following sections, we demonstrate how the F-16 replacement procurement dossier allowed for an exploitation of expertise through strategic information management by the military. We first map the preference constellation of the Minister of Defence and the federal government on the one hand, and the military administration on the other. Next, we show how the military embarked on information strategies to deliberately steer the information flows towards the minister in line with their preferences. Finally, we show how agent intrusion enabled the military administration to further outplay its information advantage.

Preference constellation

From the outset of the replacement dossier, it was clear that the Minister of Defence’s primary goal (and by extension that of the whole government) was to avoid political controversy, as he strived for absolute neutrality and transparency. The Minister decided early-on to delegate the procurement procedure entirely to the defence administration, in that way avoiding accusations of political meddling in the procurement process (De Standaard, 2018d: 11). Therefore, a so-called ‘Air Combat Capability Program’ (ACCaP) was established within the defence administration, mandated with selecting and negotiating a replacement aircraft (Belgian Ministry of Defence, 2017: 3). Further indicative of his prudence was the decision not to allow his chief of cabinet (2014–2017) to be involved with any of the political decision-making regarding the replacement process. This was a control mechanism which the Minister installed given that his chief of cabinet was the former head of the armed forces’ Air Component (CRIV 54 COM 876: 32).

A different climate existed within the Air Force. Both the acting and the former chief of the Air Force have made statements indicating their preference for the Lockheed Martin F-35. A leaked email revealed that already in 2013, while debriefing the military leadership (including the Air Force departments and the F-16 replacement unit), then Air Force Commander General Van de Voorde stated:

What is the alternative if it won’t be the F-35? The [Swedish Saab] Gripen, the [French] Rafale [of Dassault], and so on, is that what we want? I hope not? As far as I am concerned, you may also add a new F-16 to that list (CRIV 54 COM 876: 2).

One year later, his successor, Air Force Commander General Vansina, ⁵ made similar statements in favour of the F-35. Meeting his counterpart from the Dutch Air Force at a Jordan airbase, he insinuated that the F-35 would generate strong benefits in terms of maintaining strategic cooperation on maintenance, training and joint operations that Belgium had built up with the Netherlands over the last decades, a statement for which he was later criticized (De Standaard, 2016: 9). According to the newspaper De Standaard (2018a: 4), it was a public secret that Air Force Commander Vansina and other leading staffers of the Air Force have always had a strong preference for the F-35.

In sum, a clear preference divergence was present between the Minister, who wished to avoid direct involvement by stressing his neutrality and support for an open competition, and his defence administration, openly preferring the F-35. This divergence was a breeding ground for undesired agency behaviour.

Strategic information management

The decision to outsource the entire procurement process to the so-called ACCaP unit within the defence department played into the hands of the agent, making possible exploitation of its information advantages to advance its own preferences and goals. The leaked emails scandal indicates that this particularly manifested with regard to the two F-16 structural life extension studies that were withheld from the Minister. At least three elements point towards instances of strategic information management by the defence administration.

A first indication of information management concerns an internal governmental note of November 2015, based upon which the December 2015 governmental decision was made to replace the F-16 (Belgian Ministry of Defence, 2015). As mentioned, the note compared and evaluated several alternatives to a complete replacement, yet it concluded that all of these options were uncertain, costly, and therefore not opportune. The accuracy and objectivity of the data and facts in the note were, however, strongly disputed once its content was made public in the aftermath of the leaked emails scandal (see the debates on 18 and 25 April 2018, CRIV 54 COM 866; CRIV 54 COM 871; CRIV 54 COM 876). Moreover, conclusions regarding the life extension of the F-16 were made conditional upon a feasibility study that was never commissioned. Most problematic was that while officially being redacted by the cabinet of the Minister of Defence, the input for the note was exclusively provided by the defence department’s ACCaP team. In other words, the note was largely written by the special programme unit in charge of replacing the F-16, whereas the unit in charge of the maintenance of the F-16 was not able to contribute to the note, nor to review its findings (CRIV 54 COM 871: 20, 27, 30–31, 44; CRIV 54 COM 876: 10).

A second element concerns a leaked email written by a Lockheed Martin staff member. It revealed that Air Force Commander General Vansina had contacted the producer directly in July 2016, saying he was ‘quite upset’ about the studies they were proposing on the life expectancy of the structural parts of the F-16. The proposals were made in the margins of a so-called ASIP meeting, a bi-annual Steering Committee meeting of Lockheed Martin with military representatives of the different countries that participate in the F-16 programme. In particular, Lockheed Martin notified that a life extension to 9500 flying hours would be possible for the Belgian F-16s, while proposing more detailed projections. Being summoned in parliament, General Vansina admitted he was indeed not pleased with the ‘mixed signals’ they received from Lockheed Martin. He added that he therefore decided to contact them on his own initiative ‘to remind Lockheed about the decision of the Belgian government to acquire new fighter jets’ and stress that a life extension of the F-16s was not what the Belgian armed forces expected (CRIV 54 COM 876: 23–24).

A third and most serious indication of information management concerns a series of attempts by a Lieutenant-Colonel within the armed forces’ DG-MR to communicate the Lockheed Martin projections on a structural life extension of the F-16s to his hierarchy, much to the annoyance of the latter. ⁶ In the aftermath of the above-mentioned May 2016 ASIP meeting, the Lieutenant-Colonel transferred the projections to his hierarchy, indicating the relevance of the findings in the light of the planned replacement of the F-16s, without any response coming from his leadership (CRIV 54 COM 871: 2).

In addition, when a first study reached the DG-MR on 17 April 2017, two further attempts were made to hide the findings on the possibilities for life extension of the F-16 fleet. A first attempt occurred in the context of a conference for which the aforementioned Lieutenant-Colonel prepared a presentation to senior officers, summarizing the findings of Lockheed Martin’s projections. His direct superior, Colonel Letten, interfered by instructing: ‘please do not use the slide with the data on the longer life duration. The purchase programme [of a replacement aircraft] is in a crucial phase. We do not want to disturb this sensitive moment’ (De Standaard, 2018b: 1) (CRIV 54 COM 871: 37).

Another obstruction followed after the Lieutenant-Colonel’s attempt to bring the information about the life extension to the military leadership. On 10 May 2017, he sent an email to his direct superior within DG-MR (Colonel Letten), but this time also to the two leading generals of the DG-MR as well as to the Colonel in charge of the replacement (ACCaP) programme (Colonel Van Pee). In this email, he argued that the Lockheed Martin study of April 2017 is ‘very important, since it confirms previous estimations of Lockheed on the life expectancy of our F-16s. This can have an impact on the timing of the purchase of a replacing aircraft’ (De Morgen, 2018: 6). He added his belief that ‘this needs to be forwarded to the level of the Chief/Minister of Defence, since they need to have the correct information in the framework of the replacement procedure’ (De Morgen, 2018). The email caused much disturbance within the Air Force leadership, who also informed their Commander General Vansina, resulting in considerable pressure being put on the Lieutenant-Colonel to send a new email on 19 May, which nuanced the arguments of his first email (CRIV 54 COM 871: 16, 51). His superior subsequently responded to his colleagues that ‘there’s hence no indication towards an approval for an extension of the lifetime of the F-16 fleet above 8,000 flying hours’ (De Morgen, 2018: 6).

Hearings within parliament on 18 April 2018 indicate precisely that within the ACCaP group the decision was made not to forward the information to the Chief of Defence or the Minister in order to safeguard the replacement process (CRIV 54, COM 866: 12–13). Asked why the head of the F-16 fleet, Colonel Letten, did not transfer the studies, he responded that the ‘memos only gave an idea, but no conclusions on life extension; they do not serve this goal. Therefore, they did not provide any new information, since the document was not useful in the context of a life extension.’ (CRIV 54 COM 871: 35). The Chief of Defence agreed with this interpretation, calling Colonel Letten’s decision ‘justifiable’ (CRIV 54 COM 866: 17).

Interestingly, while the department under scrutiny thus concerned the DG-MR, the internal investigation unit of the Defence Ministry relied on the head of the DG-MR department, General Debaene, to assist in leading the investigation. Since the whistle-blower leaked the documents out of frustration that his direct superiors were not willing to forward the Lockheed Martin studies, it is remarkable that the leadership itself was asked to provide assistance in the investigation. The internal investigation unit motivated the involvement of General Debaene as a ‘matter of efficiency’ given the small internal inspection service of the armed forces, which requires the assistance of other departments and their expertise in investigating specific matters (CRIV 54 COM 866: 19).

In addition, the external audit conducted by the Federal Audit Service, which equally looked into the ramifications of the Lockheed Martin studies on the procurement process, was ‘based on a premise that the governmental decision on replacing the F-16s at the moment they reach 8,000 flying hours would be maintained’ (CRIV 54 COM 862: 16–17). It therefore concluded that ‘the procurement process was not contaminated by information in the memos of Lockheed Martin’ (CRIV 54 COM 862: 26). This is striking, since the content of the revealed studies precisely questioned the desirability of this decision. The seriousness of the investigations could moreover be doubted given the short time span of two and half weeks in which they were conducted (CRIV 54 COM 862: 38). The fact that the screening of internal emails was based on retrieved copies of email archives which were requested by the Federal Audit Service 12 days in advance, further adds to these doubts (CRIV 54 COM 862: 39).

In sum, empirics indicate that highly informed military staff members were repeatedly able to steer information in line with their preferences and to hide information working against those. While the audits did refrain from providing proof of deliberate manipulation, they did however warn that the current organizational structure of the F16 replacement project (ACCaP) is ‘not conducive for sufficient information streams’, and that ‘herein a danger could be identified, which is that only like-minded visions determine the course of action, therefore lacking a healthy and critical reflection, which is alien to the own powers’ (Belgian Ministry of Defence, 2018a: 2). The conclusions of the audits in that sense nevertheless paved the way for a continuation of the procurement process, which was concluded on 25 October 2018 with the announcement of the F-35 as the winning contender of the bidding process (Belgian Ministry of Defence, 2018c).

Agent intrusion

While the previous section clearly identified strategic information management by the defence department and military staff, opportunities for the strategic exploitation of information advantages were seemingly also increased through a process of agent intrusion. The Belgian Minister of Defence’s cabinet, overseeing the procurement dossier, is composed of several members of the military staff, who often take prominent positions. In addition to several military advisers, even the position of so-called ‘Chief of Cabinet’ was twice held by a high-ranking military official. From December 2014 until June 2017 the position of Chief of Cabinet was held by Lieutenant-General Van de Voorde, the former Air Force Commander. Later, in June 2017, he was succeeded by Major-General Claeys, who previously held a position in the ministerial cabinet. This strong military presence in the ministerial cabinet is a striking deviation from previous cabinets of the Minister of Defence, in which this crucial position was held by a civilian.

There are strong indications that in this way, strategic information management was possible not only within agent bodies, but equally within the principal’s own advisory body. Strategic information management through agent intrusion was manifested in at least two instances during the procurement process.

On 7 February 2018, a parliamentary opposition member submitted a question to the Defence Minister asking if his administration ever required Lockheed Martin to analyse whether the life expectancy of the Belgian F-16 type could be extended beyond 8000 hours, and if the defence administration had ever considered proposing an extension based on actual flying hours, taking into account a so-called ‘severity factor’ (CRIV 54 COM 815: 3). The technicality of the question suggested that the opposition party had by then obtained the later revealed Lockheed Martin studies, and attempted to confront the minister with this new information.

In response, a former Air Force official, working as an adviser for the party of the Minister of Defence, directly contacted Air Force Commander General Vansina to warn him ‘this is no innocent question’. Vansina, in turn, sent an email to the head of the ACCaP unit (Colonel Van Pee), mentioning ‘as guidance’ in his subject line (CRIV 54 COM 866: 3; CRIV 54 COM 876: 7, 24, 36, 46). Interestingly, the ACCaP unit in charge of finding a replacement aircraft, and not the DG-MR unit in charge of the F-16 fleet, was asked to provide an answer to the parliamentary question. Asked during a parliamentary hearing why he decided to do so, General Vansina replied that he ‘indeed made a mistake, since in fact it should have been DG-MR that would be entitled to answer such questions, not the ACCaP team’ (CRIV 54 COM 876: 36, 46). The so-called ACCaP replacement team hence drafted the answer and in turn forwarded it to the cabinet of the Minister of Defence. Cleverly, the answer did not provide a response to both questions asked, but instead stated that life extension was not relevant in the light of the earlier made decision of the government not to do so (CRIV 54 COM 815: 4).

Further evidence of steering-through-intrusion can be found in another email sent by Colonel Van Pee on 20 March 2018, instructing his staff ‘don’t let the minister of Defence change the answer [to the parliamentary question] that we wrote for him’ (CRIV COM 866: 3). Even more, on 13 April, the head of the internal investigation committee (General Robberecht) admitted in parliament that a cabinet aid of the Minister ‘was orally informed about the existence of a document with results of a simulation provided by Lockheed Martin’ while adding that ‘the uselessness of the results of the simulations was equally explained’ (CRIV 54 COM 862: 15).

A second instance of agent intrusion occurred when a military official from the Minister’s cabinet interfered in the ongoing internal investigation of the defence department. On 22 March 2018, the Lieutenant-General leading the internal investigation received a phone call from the military assistant of the Minister of Defence. The lead investigator was asked to make the whistle-blowing Lieutenant-Colonel sign a paper that stated no email exchanges took place between him and the cabinet of the Minister, which he refused to do. The military assistant also asked the lead investigator to deliver all email files of the whistle-blower, to verify whether he had contacted anyone in the cabinet (CRIV 54 COM 866: 19, 23).

Conclusion

We started this article by suggesting the broadening of the civil–military relations literature to the domain of defence procurement. Following principal–agent modelling, we argued that the highly technical nature of large defence procurement dossiers, such as the replacement of fighter jets, inherently comes with a risk of agency problems. A government’s, and by extension a parliament’s, dependency upon the expertise of the defence administration creates room for the latter to steer decision-making towards its own preferences.

Through a case study of the replacement of the F-16 fighter jet fleet in Belgium, we showed how large defence procurement dossiers allow for exploitation of expertise through processes of strategic information management. The empirics revealed that the military staff actively concealed information about a potential life extension of the F-16 fleet from the Defence Minister, in an attempt to steer decision-making towards their own preferred aircraft, the American-built F-35. In other words, strategic information management created agency problems in the principal–agent relationship between the Minister (principal) and his administration (agent).

Moreover, we have introduced the concept agent intrusion. Empirics from the Belgian case provide indications of how military staff taking up formal advisory positions in the Defence Minister’s internal decision-making structures might engage in hiding crucial information. In more generic terms, agent intrusion captures the to date unexplored phenomenon of an agent taking up advisory positions in its principal’s decision-making structures. We have provided indications of how this creates additional opportunities for the agent to exploit its information advantage. Ultimately, agent intrusion in the principal structures can increase the risk of the principal being confronted with a runaway agent. Agency problems are particularly likely to occur in decision-making dossiers or policy domains in which the principal is highly dependent upon the expertise of the agent, such as in the acquisition of new fighter jets and defence policy more broadly. From a policy perspective, these findings equally highlight the risk of controversy that comes with ministerial reliance upon military expertise. It raises questions not only about the most appropriate composition of a Minister of Defence’s advisory staff, but also about the representation of sufficient civilian personnel within the defence administration.

Further research on agency problems in civil–military relations in the domain of defence procurement is warranted, first and foremost in other fighter jet replacement dossiers that have raised contention, primarily in Canada, but to a lesser extent also in Norway or Denmark. This should include studying information loss, both in the relationship between the executive and its administration, and between parliament and government. It would also require exploring potential agent intrusion and the effects thereof in similar controversial cases. Meanwhile, the application of the newly introduced concept of agent intrusion should not be limited to the field of defence procurement or to civil–military relations. Agent intrusion can shape principal–agent relationships in all sorts of technical decision-making dossiers.