The Spectrum of Strategic Autonomy in EU Defence Supply Chains

ABSTRACT

The Strategic Compass aims to strengthen the resilience of defence-industrial supply chains and complements the long-term strategic ambition of enhancing the EDTIB. Impetus for restructuring the EU’s supply chains may be provided by the 2022 Russian invasion of Ukraine, the associated increase in defence budgets, and the promotion of joint projects and cross-border M&A. Since the prospect of achieving strategic autonomy also depends on how ‘local’ current supply chains are, we offer an in-depth investigation of three dimensions related to foreign dependency (company ownership, tender success, supply chain participation). Despite mean scores of non-EU ownership in the range of 25-30% for the EU’s largest defence firms, foreign ownership does not necessarily threaten the EU’s security of supply. Nevertheless, the participation of non-EU firms concerning M&A with high relevance for the EU defence industry increases, as does the share of defence-related EU tenders that are won by non-EU firms. An assessment of four multinational military aircraft programmes further exemplifies that the EU’s defence R&D and production capabilities may critically depend on outside suppliers. We thus find that the importance of non-EU actors for the EU’s defence industry is large and may not easily be mitigated, particularly if commodity imports are considered.

KEYWORDS:

• Defence industry
• EU strategic compass
• European Defence Technological and Industrial Base (EDTIB)
• supply chain dependency
• industrial policy

JEL CLASSIFICATION:

• L60
• L64
• F50

Acknowledgments

We thank the participants of the 25th Annual International Conference on Economics and Security, Universita degli Studi di Perugia, 29^th June 1^st-July 2022, as well as the two anonymous peer reviewers for their constructive comments and suggestions on earlier drafts of this paper. The usual disclaimer applies.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Authors’ analysis of the revenue of the EU’s largest civilian and defence firms, based on data compiled from Forbes (2022) and SIPRI (2021).

2. Airbus, Arquus, BAE Systems Bofors, BAE Systems Hägglunds, Chantiers d’Atlantique, Constructions Mécaniques de Normandie, Czechoslovak Group, Damen Shipyards Group, Dassault Aviation, Diehl, Dynamit Nobel Defence, Elettronica, Eurenco, ESG Elektroniksystem und Logistik, EXPAL Systems, FFG Flensburg, Fincantieri, FN Herstal, Fokker Technologies (now part of GKN Aerospace), Framatome, GDELS, German Naval Yards, Hensoldt, Heckler & Koch, Indra Sistemas, Iveco Defence Vehicles, John Cockerill, KNDS, Leonardo, Lürssen, MBDA, MTU Aero Engines, MTU Friedrichshafen (now part of Rolls-Royce Power Systems), Naval Group, Navantia, OGMA, Patria Group, Polska Grupa Zbrojeniowa, Ratier-Figeac, Renk, Rheinmetall, Rohde & Schwarz, Saab, SABCA, Safran, Sonaca, Thales Group, ThyssenKrupp, Vodochody Holdings, ZF Friedrichshafen. These ‘leading’ defence firms are selected based on their defence revenue. Several other EU defence firms could have been included instead of the smallest firms in our sample, but as we assign weights based on the firms’ defence revenues, the impact on our main results would be negligible.

3. For example: governmental, private firm, financial institute, bank, investment fund, family-owned.

4. Belin and Masson (2017) and Belin, Fawaz, and Masson (2019) cover 61 and 63 firms, distributed as follows. Trans-European: 3/3, UK: 9/9, other EU: 15/17, US: 34/34. While we do not cover UK and US firms, our analysis applies a larger EU sample (50 firms).

5. Defined by Bureau van Dijk as having a shareholding of 0,01% or higher.

6. It is highly probable that state- and family-owned firms in our sample were at least partly owned by non-EU shareholders if such investment were possible. By contrast, EU ownership scores of 0% indicate that a firm has become a wholly-owned subsidiary of a non-EU firm (e.g. GDELS, BAE Systems Bofors & Hägglunds).

7. These explanatory factors for varying degrees of national ownership (state- and family-ownership, cross-shareholdings) are also discussed by Belin and Masson (2017) and Belin, Fawaz, and Masson (2019).

8. However, there have been exceptions. Cevian’s lobbying for a streamlining of ThyssenKrupp’s activities, for example, nearly resulted in a divestment of ThyssenKrupp Marine Systems.

9. Our analysis is restricted to larger M&A with an EU dimension (i.e. those that met the turnover thresholds set by the EU Merger Regulation) and to those that were referred to the Commission by national competition authorities. While defence M&A investigated by national competition authorities are not covered, we consider our dataset as sufficiently large and representative for the most important defence M&A within the EU.

10. This subsample only covers cases in which the acquiring firm(s) are located exclusively outside the EU27. Were cases with an additional involvement of an EU27 acquirer considered (e.g. a joint venture between an EU and a US firm), the size of the subsample would increase from 42 to 71 cases.

11. Such as the 2020 acquisition of the Asteelflash Group, Europe’s second largest electronic manufacturing firm with defence and aerospace applications, by a Shanghai-based subsidiary of the semiconductor producer ASE.

12. Logistic modelling is suited to deal with binary dependent variables (Cameron and Trivedi 2005).

13. Based on 9,596 observations (for 3,008 tenders, value data was unavailable).

14. For the remaining 7,376 tenders, data was lacking (notably for the variables Tender Value_i and Award Criteria_i).

15. Based on data gathered from the SIPRI Arms Transfers Database (2022), we determined that the portion of defence imports by the EU27 that is sourced in non-EU27 countries increased from 42.5% in 2016 to 74.6% in 2020.

16. To test the robustness of our findings for extremely protectionist policies, we ran the logistic model on a subsample of tenders (n = 1,586) which are covered by the Agreement on Government Procurement by the World Trade Organization. The results demonstrate that the significance of the variables of interest (tender size and number of bids) remains, and coefficient sizes increase.

17. Braddon and Hartley (2013) and the NAO (2004, 2001) claim higher delays and cost overruns in collaborative EU defence projects, but Hartley and Braddon (2014) find no inverse relationship between the number of partner nations and development time.

18. Concerning additional explanations for EU defence collaboration, an overview of the various research strands is given by Calcara (2018).

19. See Matthews and Al-Saadi (2021) for additional information on the organisation of the Typhoon consortium.

20. See Gutierrez (2013) for a presentation of the global supply chain of the A400M. This subcontracting process may also apply to R&D: Bellouard and Fonfría (2018) state that about 70% of all R&D in military aircraft is typically subcontracted. According to Hayward (2001), much of the globalisation of modern defence supply chains takes place at lower-tier levels, where it involves commercial technology and is difficult to observe.

21. Disagreement over industrial participation can be a major reason behind project withdrawals (Calcara 2020b).

22. See Keohane (2002); Barrinha (2010); Kleczka, Buts, and Jegers (2020); Matthews and Al-Saadi (2021).

23. According to Airbus data presented by Gutierrez (2013), the equipment list of the A400M covers more than 4,000 individual parts, of which only about 60% are installed at the sites of Airbus and its subsidiaries.

24. Several decisions had to be made during data collection. First, if firms are listed multiple times as suppliers for the same programme, they are counted repeatedly if their contributions can be clearly differentiated. For example, Thales is a tier-1 supplier for five separate avionics and airframe systems of the Tiger. Second, as will be presented in more detail later, the distinction between EU and non-EU suppliers is not always clear. EU firms are not counted as non-EU suppliers if they were acquired only recently (for example: CESA, part of Héroux-Devtek of Canada since 2018, is counted as a Spanish supplier).

25. This data is also evaluated by Linnenkamp (2017). Certain variations between our findings and Gutierrez' (2013) distribution of suppliers within Europe could be explained by the time difference between the separate analyses and by varying treatments of foreign ownership.

26. Based on our evaluation, the approximate distribution of approved Airbus suppliers by origin is as follows. EU27: 57%, U.S.A: 18%, U.K: 12%, other countries: 13%. While this approval list also includes civilian Airbus programmes, its distribution resembles our findings for the A400M.

27. Vertical integration may lessen dependencies for specific parts and materials. Most leading EU defence firms are vertically integrated in certain areas. For example, Naval Group and TKMS have acquired the naval electronics and defence businesses of Thales Naval and Atlas Electronics, respectively.

28. For example, it has been argued that Western European defence industries are at a comparative disadvantage vis-à-vis the U.S in terms of military R&D (Hartley 2017; James 2008; Kapstein 1994; Chiang 1992).

29. See footnote 15.

30. For constructing this list of critical resources for the EU’s defence-industrial supply chain, we combine the findings of Pavel and Tzimas (2016) with the latest edition of the European Commission’s triennial communication on critical raw materials (CRMs) (European Commission 2020), with background documents prepared by Latunussa et al. (2020) and Blengini et al. (2020), and with our own research.

31. Metals: aluminium, barium/baryte, bauxite, beryllium, bismuth, cadmium, chromium, cobalt, copper, gallium, germanium, indium, iron ore, lead, lithium, magnesite, manganese, molybdenum, nickel, niobium, rhenium, tantalum, thorium, tin, titanium, tungsten, vanadium, zinc, zirconium. Precious metals: gold, palladium, platinum, silver. Non-metals: boron, fluorspar, phosphate, selenium. Mineral fuels: crude oil/petroleum, natural gas. REE include, inter alia, dysprosium, neodymium, praseodymium, samarium, and yttrium. This selection is largely based on Pavel and Tzimas (2016), with some alterations to account for the European Commission’s list (2020) and our own findings.

32. Production data is taken from the World Mining Data series of the Austrian Federal Ministry of Agriculture, Regions and Tourism (BMLRT, 2022) and the Mineral Commodity Summaries published by the US Geological Survey (USGS, 2022). Import reliance stems from Blengini et al. (2020). Import data is taken from Eurostat (2022) and Latunussa et al. (2020). The import data should be treated with caution: first, it does not always clearly separate raw from refined materials. Sometimes, imports are not reported due to commercial confidentiality rules (e.g. Austria’s tungsten imports, see Latunussa et al. 2020). Furthermore, the data is not corrected for the Rotterdam effect, which may distort the distinction between imports from outside and within the EU.

33. Following Blengini et al. (2017), import reliance reflects the relation of net imports to apparent consumption for a given year t and commodity i as follows.

$\mathit{import}\mathit{relianc}{e}_{\mathit{i},\mathit{t}}=\left(\mathit{impor}{t}_{\mathit{i},\mathit{t}}-\mathit{expor}{t}_{\mathit{i},\mathit{t}}\right)/\left(\mathit{domestic}\mathit{productio}{n}_{\mathit{i},\mathit{t}}+\mathit{impor}{t}_{\mathit{i},\mathit{t}}-\mathit{expor}{t}_{\mathit{i},\mathit{t}}\right)$.

34. We investigate concentration in the commodity markets via the Herfindahl-Hirschman-Index, defined as $\mathit{HH}{I}_{\mathit{i},\mathit{t}}={\sum }_j^J{\left({\mathit{s}}_{\mathit{i},\mathit{j},\mathit{t}}\right)}^2$ for a given year t and commodity i, with $s_j$ denoting the share of country j in global production.