Interorganisational networks and transgovernmental cooperation

The EDI MILCON program is executed through a dense web of US defence actors, host-nation ministries, and implementing agents: an archetypal interorganisational network, rather than a single hierarchy or a spot market transaction (Hafner-Burton et al., 2009). Networks are defined by actors, ties, and interdependencies; the ties of some nodes shape, and are shaped by, others (Kinne, 2013). This perspective directs attention away from isolated organisations (“trees”) and towards the organisation of collective action (“forest”) (Salancik, 1995). It also requires a multilevel lens: operational project teams, embassy liaisons, and strategic leadership interact vertically and horizontally, and explanations that stay at one level risk missing key coordination mechanisms (Moliterno and Mahony, 2011; Podolny and Page, 1998, pp. 58–60).

Two implications follow for this study. First, relationships and the capacity to maintain them are not incidental but constitutive of performance in multinational programs. Second, networks in this domain are often transgovernmental, linking specialised agencies directly across borders with limited mediation by foreign ministries (Raustiala, 2002). This is precisely the structure of EDI MILCON program, where ODC, USAF component commands, and host-nation technical offices interact routinely and repeatedly. Within such networks, formal legal instruments (e.g. DCAs) are not mere background; they shape who may coordinate with whom, on what terms, and with what authorities. That observation motivates the article’s core intent: theorising legal instruments as a resource inside the network, not only a constraint on it.

Resource dependence theory in multinational public programs

The RDT explains how organisations survive by securing and managing critical external resources in settings of uncertainty and interdependence (Pfeffer and Salancik, 1978). Five analytic foci structure the approach: (1) which resources matter; (2) how flows of those resources traverse organisational boundaries; (3) the dependencies and power differentials those flows create; (4) how dependencies constrain action; and (5) how leaders manage dependence (Johnson, 1995). Leaders do so through buffering (protecting and insulating organisational boundaries) and bridging (spanning or shifting boundaries) (Johnson, 1995). Subsequent work catalogues concrete tactics—including political action, formal interorganisational arrangements, and board-like oversight mechanisms among them—to reduce environmental dependence (Hillman et al., 2009).

Three refinements adapt RDT to the context of EDI. First, criticality matters: not all resources are equally important. The theory distinguishes magnitude from criticality (how essential the resource is to core tasks, and how concentrated it is in the environment) (Nienhüser, 2008; Pfeffer and Salancik, 1978). Second, in multinational programs, the line between “environment” and “organisation” is blurred: boundaries are defined by patterns of interlocked or coordinated behaviours, not by formal charts (Pfeffer and Salancik, 1978). Third, in transgovernmental settings, legal instruments (DCAs and IAs) can be conceptualised as first-order organisational resources that stabilise access, authority, and compliance expectations—altering the payoff to buffering/bridging tactics, such as bilingual communications support, liaison protocols, and long-range development planning (Raustiala, 2000, 2002, 2005; Kinne, 2020).

Legal–institutional architecture as a constitutive resource

Compliance and effectiveness research in international cooperation shows that formal instruments shape behaviour not only by prescribing rules but also by structuring the ongoing relationships among specialised agencies (Raustiala, 2000, 2002). In defence cooperation, DCAs delineate rights, obligations, processes, and authorities for basing, construction, environmental remediation, access, and information sharing; IAs localise these commitments for specific projects. In network terms, DCAs reconfigure the infrastructure of ties by legitimating DLA, clarifying role responsibilities, and reducing transaction costs for recurrent coordination. Consequently, DCAs function as structural rigidity within the network: they dampen shocks from domestic political turbulence and help translate strategic intent into operational execution. Whenever such an instrument is absent or immature, other resources, for example communication capacity, ODC staffing, and leader sponsorship, become more fragile and more easily swamped by exogenous pressures. This logic is consistent with both the RDT emphasis on managing dependence through boundary work (buffering/bridging) and with network accounts of how formal and informal rules shape tie formation and durability—albeit with emphasis on informal rules (Podolny and Page, 1998, pp. 60–62; also, Kinne, 2013, 2020; Moliterno and Mahony, 2011).

Alliance politics bridge

Our claim complements alliance-management work on commitment and domestic constraints: DCAs operate as hand-tying instruments that align host-nation decision rights with implementation and raise the domestic political costs of reversal (Fearon, 1997; Morrow, 1994; Putnam, 1988; Snyder, 1997). Legalisation—through obligation, precision, and delegation—allocates authority and structures expectations that stabilise cooperation (Abbott and Snidal, 2000; Raustiala, 2005). In this view, the agreement does not merely add resources; it constitutes the execution network, clarifying decision rights and escalation paths, so that boundary work (buffering/bridging) yields progress, rather than politicisation (Johnson, 1995; Pfeffer and Salancik, 1978). Legalisation—via obligation, precision, and delegation—allocates authority and structures expectations that stabilise cooperation (Abbott and Snidal, 2000, pp. 421–424; Raustiala, 2005). Thus, a DCA/IA bundle constitutes the execution network—clarifying decision rights and escalation paths—so that RDT boundary work (buffering/bridging) yields progress, rather than politicisation (Fearon, 1997, pp. 69–71; Johnson, 1995; Leeds et al., 2000; Pfeffer and Salancik, 1978; Putnam, 1988, pp. 434–436; Rapport and Rathbun, 2021; Snyder, 1997).

From synthesis to expectations

Bringing these strands together yields clear expectations that guide the comparative analysis. First, transgovernmental relationships cannot be fully understood through the lenses of classic RDT, emphasising resources and informal network ties. Presence of DCA—a legal instrument—leads to higher network resilience and execution speed. A mature DCA should correlate with faster problem resolution, more stable coordination across strategic/operational levels, and less susceptibility to politicisation of routine design actions.

Second, capacity of ODC is conceptualised as a pivotal node, because ODCs sit at the US–host-nation interface in a transgovernmental network; their staffing, mission clarity, and bilingual capabilities should mediate the returns to legal scaffolding, thereby amplifying benefits under a DCA and partially compensating (but rarely substituting) when it is absent.

Third, buffering/bridging tactics, such as direct liaison protocols, key-point memos, dual-language deliverables, and long-range development plans, may be only conditionally effective; they all are essential resources and strategies, but they should yield higher payoffs when legal–institutional scaffolding stabilises authority and expectations; without it, these tactics face diminishing returns.

By reframing the literature this way, the article speaks back to three communities. For RDT, it identifies legal instruments as a neglected class of resources in public and multinational programs, and specifies how they condition the effectiveness of buffering/bridging. For network scholarship, it shows how formal agreements alter network structure and affiliate durability in practice. For public administration and security cooperation, it clarifies why some multi-actor defence programs proceed while others stall despite similar funding, timelines, and stated strategic aims.

Theoretical framework

We treat each country’s EDI MILCON execution network (USAF/EUCOM/US Army Corps of Engineers [USACE]/US Air Force Civil Engineer Center [AFCEC]–US embassy/ODC–host-nation ministries and base commands) as the organisation. Following RDT, organisational behaviour reflects how actors secure and manage critical external resources under uncertainty (Johnson, 1995; Pfeffer and Salancik, 1978). We operationalise RDT through Johnson’s five foci and the buffering/bridging repertoire, and we specify how legal instruments (DCAs and IAs) enter the model as the first-order resources that condition the effectiveness of other tactics. We align evidence to Johnson’s (1995, pp. 1, 5–7) five RDT foci.

Resources (what matters)

We define resources as inputs that the execution network cannot autonomously generate but must obtain or stabilise from its environment. For EDI MILCON program, the core set is: (1) legal scaffolding: DCA and IAs (access, environmental, information-sharing, and liaison provisions); (2) communication capacity: bilingual technical communication, dual-language deliverables, shared terminology/protocols, and “key-points” documentation discipline; (3) boundary nodes and staffing: ODC manpower/mission clarity, and availability of in-theater technical liaisons; (4) senior-leader sponsorship: timely engagement from DoD/embassy/Host Nation principals to unblock cross-boundary issues; (5) land/use and permissions: site access, easements/zoning equivalents, environmental constraints; and (6) temporal and fiscal resources: schedule latitude and obligation rules, and stability and timing of appropriations.

We hypothesise the DCA as a constitutive resource: it formalises authorities, legitimates direct ties, lowers transaction costs, and creates structural rigidity in the network.

Resource flows (how they move)

Flows occur through codified interfaces (e.g. DLA, IAs, letters of intent [LOI], long-range development plans [LRDPs]), routine artifacts [dual-language submittals, design review cycles], and convening structures [steering committees, and joint site visits]. Smooth flows mean faster mutual adjustment and fewer escalations; blocked flows manifest as mismatches in expectations, cascading rework, or politicisation of routine steps.

Dependencies and power differentials (what that creates)

Whenever access to a critical resource is concentrated (e.g. only the Ministry of Foreign Affairs/Ministry of Defence [MFA/MoD] can confer land/sovereignty permissions; only State/OSD can finalise DCA text), the network becomes dependent on that locus. Absence of DCA heightens dependence on ad hoc substitutes (personal ties, and NATO habits of cooperation), raising vulnerability to domestic politics. Thin ODC staffing increases dependence on overextended project managers or HN translators, amplifying miscommunication risk.

Constraints (how dependency manifests)

Constraints show up as: (a) legal ambiguity (e.g. lack of a shared rulebook may result in contested authority); (b) communication loss (terminology drift, translation error); (c) procedural friction (unclear land/environ responsibilities); and (d) politicisation shocks (media or party framing of technical steps as sovereignty threats). These tighten schedule windows and trigger risk-averse slowdowns or halts.

Leadership actions (how actors manage dependency)

Leaders deploy the classic RDT toolkit, buffering and bridging. The first, buffering (protect/insulate boundaries), may manifest in strategies, such as embedding linguists in contracts; requiring dual-language deliverables; standardising “key-points” memos signed at meeting close; shoring up ODC with dedicated MILCON liaison capacity; pre-clearing terminology (Automatic Data Processing [ADP]/Management Development Programs [MDP]/LRDP), that is strategies to insulate core activities of the program from external dependencies. The second, bridging (span/reshape boundaries), may manifest informalising DLA; co-producing LRDPs to align US/HN roadmaps and land-use plans; issuing LOI to clarify roles; creating steering committees (board-like oversight); and, when needed, political action (principal-level statements/op-edits/ministerials) to reset expectations. Critically, we propose that the payoff to these tactics is conditional on DCA status: with legal scaffolding, buffering/bridging compounds; without it, returns diminish and may even invert under politicisation.

Propositions

We derive three key propositions that structure both comparison and coding:

P1 (constitutive legality): The presence of a DCA/IA increases the system’s capacity to absorb shocks by clarifying authorities, sequencing, and escalation paths.

P2 (conditional payoff to boundary work): Under a DCA/IA, buffering (linguists, bilingual deliverables, and “key-points” memos) and bridging (principal engagement and steering routines) yield increasing marginal returns; without it, returns diminish as politicisation increases.

P3 (network chokepoint): Embassy/ODC capacity mediates P1–P2: with a legal scaffold, extra ODC bandwidth translates into execution speed; without it, the node saturates and becomes a visible program liability (thin capacity invokes frictions even with a DCA).

These propositions are implemented as a deductive codebook with sub-codes for specific tactics, followed by pattern coding; interview and documentary evidence is then matched to the propositions in the findings. Operationalisation and coding details are provided in the below-mentioned section, “Methodology.”

Comparative case design and unit of analysis

We employ the most-similar systems design: two NATO allies received the EDI MILCON investments on a comparable timeline and under shared US/NATO governance, yet diverged in outcomes. The unit of analysis is each country-level execution network linking US Air Force actors (USAF, EUCOM, and US Air Forces in Europe [USAFE]), the US embassy/ODC, host-nation ministries and agencies, and implementing contractors.

Data and sources

We triangulate two streams of evidence. (a) Documentary: unclassified official correspondence, schedules, fact sheets, and public statements (2017–2020). (b) Semi-structured interviews: ten practitioner interviews (e.g. officers and civil engineers) conducted in 2021–2022 by phone or encrypted email; all interviews were conducted on a non-attribution basis and anonymised. Interview prompts focused on authorities, sequencing, and boundary-work practices (liaison, translation, and artifacts).

Analytical approach and coding

Analysis is theory-guided. We developed a deductive codebook anchored in Johnson’s (1995, pp. 2, 5–8) five RDT foci—resources, resource flows, dependencies and power, constraints, and leadership actions—and the buffering/bridging repertoire. We coded interview memos and documents in a first pass, recoded in a second pass to consolidate themes, and built a cross-case matrix (Poland vs. Slovakia) and event timeline for process tracing and pattern matching. We maintained an audit trail of decisions, produced case matrices and timelines (Miles et al., 2020), and conducted pattern matching and process tracing across sources.

Operationalisation

The outcome, execution effectiveness, is judged by (i) attainment and continuity of planned milestones on the documented timeline (e.g. senior leader meeting [SLM] and initial engagements, site access, iterative design reviews, award, and start), (ii) time-to-first-award, and (iii) continuity (absence of a formal ministerial halt). Key conditions include legal architecture (presence/absence and maturity of a DCA and any IAs), ODC capacity (staffing, bilingual competence, and mandate clarity), communication discipline (embedded linguists, dual-language deliverables, and signed “key-points” memos), leadership escalation (number/timing of principal interventions), and land/environmental responsibility clarity.

We also operationalise two families of mechanisms. First, bridging and governance tactics—including DLA, co-produced LRDPs, and steering/board-like routines—are coded as sub-codes linked to legal scaffolding and leadership escalation. Second, politicisation episodes (media frames, parliamentary salience, and ministerial statements) are coded as constraints that compress schedule latitude and reconfigure authority claims. Finally, resource flows are captured via the presence, timeliness, and completeness of bilingual submittals, signed “key-points” memoranda, and documented design–review cycles, with blockages recorded as returned submittals or formal requests for clarification.

Use of interview material

Interview evidence appears in the section, “Findings,” through attributed roles (e.g. ODC engineer, host-nation planner) and is tied to specific coded themes in the analysis matrices. To preserve non-attribution, we cite interview material in-text as “Interview, role” and cross reference it to coding tables themes. Interview evidence is presented through role-attributed paraphrases and selective verbatim quotations. See Appendix A for the interview protocol and Appendix B for the respondent index (on role-based non-attribution basis).

Data availability and ethics

Redacted interview memos, the codebook, and the cross-case matrix are stored on an encrypted university drive; access is provided to editors/reviewers under non-attribution assumption.

Audio recordings were not retained. The project involved professional experiences and organisational processes, and collected neither the sensitive personal data nor the classified or otherwise sensitive organisational data.

Constitutive legality

Across documents and interviews, respondents emphasised that the legal architecture did not merely constrain activity but created the authority needed to act. In Poland, the prior DCA formalised DLA, clarified roles and responsibilities (including information sharing), and provided predictable jurisdictional pathways. This legal scaffolding stabilised coordination and accelerated routine design steps consistent with work on how agreement form shapes organisational behaviour (Johnson, 1995, pp. 5–8; Raustiala, 2005, pp. 581–584). Respondents noted that, even when responsibilities occasionally had to be assumed across organisational boundaries, Poland’s ODC could absorb and route those tasks within the rule set and resist politicisation pressures. By contrast, in Slovakia, ordinary coordination lacked a shared rulebook and became contestable, pulling senior officials into decisions that would otherwise have remained a routine.

Politicisation cascade (Slovakia)

An early public mischaracterisation of a contracting/design step was amplified by parliamentary debate and ministerial turnover, reframing technical design actions as a sovereignty threat. Without a DCA to anchor interpretations, escalations consumed managerial bandwidth and interrupted contracting. Respondents from the Slovak side described the political climate during the period as less receptive to the program and noted that EDI became a vehicle for domestic contestation. The environment further deteriorated following the widely reported killing of a journalist and subsequent political upheaval (Sirotnikova, 2020); opposition actors increased pressure to reject EDI as sovereignty infringement. Public statements—including an ambassadorial op-edit underscoring principles of defence cooperation—did not reverse this trajectory. As one AFCEC project manager put it in official correspondence, accelerated design was used as a scare tactic to halt progression and “had a DCA been in place, the opposing political party would’ve had less ‘ammo’ to work with.”²

Communication discipline and ODC capacity

Poland’s network invested in bilingual artifacts, signed “key-points” after meetings, and embedded linguists at critical interfaces—classic buffering tactics—while a staffed ODC transmitted credibility across organisational boundaries. LRDPs functioned as bridging devices that aligned host-nation and US horizons. In Slovakia, ODC staffing was thinner and translation practices more ad hoc; these efforts were diligent but could not offset legal ambiguity. Differences in land-use regimes (e.g. zoning versus easements and fragmented ownership) compounded the problem: Poland’s legal scaffolding offered tractable routes through permissions and environmental responsibilities, whereas ambiguity in Slovakia raised transaction costs and stall risk.

Synthesis

Whenever a DCA is present, buffering (bilingual artifacts, embedded linguists, and structured meeting discipline) and bridging (DLA, LRDPs, and steering/board-like routines) compound to stabilise resource flows and insulate projects from politicisation. ODC capacity moderates these effects: thin staffing and unclear mandates increase reliance on improvised ties, especially under weak legal architecture. Consistent with this pattern, Poland moved to completion of the first project in spring 2019, whereas Slovakia formally halted negotiations in March 2019 (Spectator Staff, 2019).

Implications for resource dependence theory

Our analysis suggested that legal instruments are not just resources but first-order resources. The comparative evidence underscores that DCAs are not mere background constraints but constitutive resources that shape the dependence structure of multinational public programs. In Poland, EDI program implementation was not seamless, but the DCA—iteratively supported by IAs—reduced environmental uncertainty, legitimised DLA, and stabilised expectations about roles (e.g. environmental responsibilities), enabling the network to convert strategic intent into concrete milestones. In Slovakia, the absence of a suitable DCA magnified dependence on ad hoc ties and left routine design actions vulnerable to sovereignty framings and politicisation, thereby culminating in a ministerial halt.

The analysis also suggested that, as important as buffering and bridging are, the return on them is conditional. Classic RDT tactics, such as buffering (in this case, strategies such as bilingual documentation, signed “key-points,” and embedded linguists) and bridging (in this instance: DLA, LRDPs, and steering/board-like oversight), paid off in Poland because legal scaffolding conferred structural rigidity; the same tactics showed sharply diminishing returns in Slovakia in the absence of a DCA. This conditionality refines RDT’s generic prescriptions by specifying when boundary work scales and when it stalls in transgovernmental settings.

Thus, the findings sharpen RDT’s distinction between magnitude and criticality: resources such as communication funds and staff were important in both cases, but the critical resource—the one whose absence reconfigures all other dependencies—was the DCA. In effect, legality functioned as a meta-resource that determined the payoff to all other inputs (e.g. communication capacity, ODC staffing, and principal engagement): high in both criticality and magnitude, but not constitutive.

Implications for alliance politics and defence cooperation

The EDI MILCON program turns alliance promises into usable infrastructure in a deteriorating European security environment. Our comparison isolates a fixable lever—a DCA with clear IAs—that prevents routine design/permits from being derailed by domestic politics. Thus, the study highlighted the gaps that can form between treaty context and operational authority. Alliance politics scholarship often treats formal agreements as macro backdrops; however, the cases also show they operate as micro-level operating authorities that enable recurrent, low-friction collaboration among specialised agencies (USAF, ODCs, and host-nation technical ministries). Where a DCA exists, it routinises cross-border coordination (e.g. direct liaison), reduces transaction costs for design reviews, and clarifies sensitive issues (land use and environmental remediation). Where it does not exist, even well-intentioned steps (e.g. soliciting design support) can be reframed domestically as sovereignty incursions.

The study also illustrated the variability of politicisation thresholds. The Slovak sequence demonstrates how much the lack of legal scaffolding lowers the threshold for politicisation cascades: an early media mischaracterisation, followed by a national shock and elite contestation, swamped operational relationships, and frozen progress. In the words of one of the respondents, “Opposing HN politicians used the accelerated designing as a scare tactic… Had a DCA been in place, the opposing political party would’ve had less ‘ammo’ to work with” (AFCEC PM, official correspondence). By contrast, Poland’s legal architecture dampened similar shocks through predictable escalation channels and implementing arrangement updates, preserving momentum. For alliance managers, the lesson is that sequencing matters: secure or mature the DCA before compressing design and award timelines.

Finally, key nodes in transgovernmental relationships serve not only as implementers but also as credibility transmitters. ODCs guard at the interface of diplomacy and implementation. Their effectiveness in both is contingent on both resources and legal backing. Their staffing, bilingual competence, and mandate clarity condition whether allied commitments propagate intact through host-nation bureaucracies and publics. Under-resourced ODCs may become chokepoints, especially in the absence of a legal backing, such as DCAs; adequately resourced ODCs act as credibility transmitters that keep alliance promises legible and actionable.

Robustness

Three conceiveable alternative explanations do not overturn the observed mechanism. ODC capacity alone could be conceptualised as a driver of success or failure; our observations suggest that ODC capacity alone is conditional—without DCA/IA it saturates and does not reverse the stall. The outcomes cannot be explained with local politics alone either: routine turnover occurred in both settings; only Slovakia crossed a politicisation threshold. Finally, it is not likely that the outcomes were affected by funding cadence: both cases drew from the same EDI lines/windows and timing does not co-vary with divergence.