NATO’s Two Percent Guideline: A Demand for Military Expenditure Perspective

ABSTRACT

One of the outcomes of the 2014 North Atlantic Treaty Organization (NATO) Summit held in Wales was the recommendation that allies should strive towards spending at least 2% of their Gross Domestic Product (GDP) on military expenditure by 2024. This has in recent years put the debate on burden sharing within the Alliance in the limelight. Within the Alliance, the interpretation of the guidelines is different. This is partly because the U.S.A, especially under President Trump, has consistently viewed the 2-% guideline as a binding agreement. Some allies have, however, indicated that their military expenditure will not increase to the recommended threshold by the set deadline. Critics of the guideline have also identified several shortfalls associated with it and consequently dismissed it. Those who support it emphasize for instance that it contributes to addressing grievances about free riding within the Alliance. This paper contributes to this discussion and literature on demand for military expenditure by assessing how NATO’s two percent guideline can be viewed from a demand for military expenditure perspective. It also proposes and estimates a dynamic panel model for this purpose. Empirical evidence presented suggests that fiscal conditions require attention in the debate on the two percent guideline.

KEYWORDS:

• Defense investment pledge
• demand for military expenditure
• European Union
• NATO
• two percent guideline

JEL CLASSIFICATION:

• D74
• H56
• H60
• N44

Introduction and Background

Disagreement within the North Atlantic Treaty Organization (NATO) on burden sharing, and especially the weight that should be pulled by European allies, has existed since the creation of the Alliance in 1949. The contest around burden sharing has generally focused on financial aspects, especially military expenditure, mainly because it is easier to measure than for instance defense capabilities (Lunn and Williams 2017; Viggo and Ringsmose 2017).

During the 2014 Wales Summit, NATO formalized how allies are expected to contribute more fairly to financial burden sharing through the Defense Investment Pledge. One of its key recommendations urges allies to strive towards committing at least 2% of their Gross Domestic Product (GDP) to military expenditure by 2024 (NATO 2014). This recommendation has since given the debate on financial burden sharing more impetus, especially under the Trump administration (Viggo and Ringsmose 2017). The Trump administration, for instance, has interpreted the two percent of GDP guideline¹ as a binding commitment rather than a recommendation. In 2018, President Trump argued that allies that did not meet the guidelines owed the U.S.A. money for underinvesting in their defense (Sanders 2018).

The aim of this paper is twofold, namely, to contribute to the debate on the two percent guideline and to contribute to the literature on demand for military expenditure. This is done by approaching the guideline from a demand for military expenditure perspective and by estimating a dynamic panel model using the generalized method of moments (GMM). The paper also integrates public debt as a measure of income in a neoclassical model of military expenditure demand. In this regard, it aims to especially emphasize the importance of the European Union’s fiscal regulations in constraining development in military expenditure of NATO allies in the European Union.

The European Union’s Fiscal Rules

The Maastricht Treaty, signed in 1992, formalized the European Economic and Monetary Union. Among the goals of the Treaty included the coordination of fiscal policies, especially by restricting Member States’ public debts and fiscal deficits. It set the limit for general government consolidated gross debt at 60% of GDP² and for fiscal deficits at 3% of GDP. The Stability and Growth Pact (SGP) that followed in 1997 was intended to strengthen the monitoring and coordination of Members States’ economic and fiscal policies to ensure compliance with existing fiscal regulations. In that regard, the SGP can be described as an instrument intended to support sound public finances and coordinate fiscal policy among European Union Member States. Consequently, the European Union expects its Member States to make budgetary decisions while at the same time taking into consideration the existing debt and deficit conditions (European Commission 2021a, 2021b).

Whereas the SGP applies to all European Union Member States, its terms are stricter for those in the Eurozone and those that are part of the Exchange Rate Mechanism (ERM II).³ Eurozone Member States annually present a stability program that describes the status of their public finances. Non-Eurozone Member States on their part present convergence programs, which, inter alia, include information about their public finances (Angerer 2019). One of the key successes of the SGP is its integration into domestic laws of European Union Member States (Ayuso-i-casals et al. 2009). However, the extent to which it has managed to regulate public debts and fiscal deficits is debatable. On one hand, it is considered a failure due to weak implementation of its requirements (De Haan et al. 2004). Today, many Member States neither fulfil the conditions on the public debt nor on the budget deficit. The debt-to-GDP ratio of countries such as Greece, Italy, and Portugal was for instance above 100% in 2019 (Figure 1).

Nevertheless, Italy is one case to have faced the wrath of the European Union when in 2018 it had its 2019 budget proposal rejected for violating the bloc’s budgetary rules. Nor did it contribute to lowering the country’s public debt. Consequently, the country needed to revise the proposed budget in accordance with the recommendations of the European Union before being accepted (Fonte and Scherer 2018).

Relevance in Studying Demand for Military Expenditure

NATO allies in the European Union form a good case study for demands for military expenditure. Firstly, demand for military expenditure literature explains military expenditure as a function of, inter alia, economic factors. Second, countries strive to optimize utility subject to budget constraints.

Additionally, the SGP requires Member States that face significant deficits to indicate how they intend to reverse the situation (Angerer 2019). In this regard, Becker (2019) argues that the European Union’s fiscal rules, which stipulate measures for sanctions upon non-compliance, could constrain military expenditure.

Furthermore, one can argue that decision makers would find it difficult to convince the electorate about the need for rapid increases in military expenditure but at the same time convince it about the need for austerity in other areas of public expenditure. France, for example, announced in 2018 intentions to increase its military expenditure by 40% between 2018 and 2025 in a bid to achieve the two percent guideline (Cross 2018). The country’s 2018 budget included an initial austerity plan, which was partly intended to decrease the budget deficit in line with the European Union’s regulations (RFI 2017). However, civilian protests forced the government to reverse some of its austerity policies (Viscusi and Fouquet 2018).

Another argument to portray the relevance of the European Union’s fiscal rules bases on Germany’s federal budget plan for 2016–2020, whose priority was to avoid new debt and deficits during any given year under the period. The stated aim was to decrease the public debt to below 60% of GDP (Ministry of Finance 2016). The country’s financial plan for 2019-2022 also aimed to decrease the debt-to-GDP ratio to below 60 percent (Ministry of Finance, 2018). Before Germany made it clear in 2018 that it did not intend to achieve the two percent guideline by 2024, its 2016 Defense White Paper had indicated that attainment of the guideline would depend on the country’s financial potential and resources (Federal Government, 2016).

Research Questions

This paper examines two research questions: i) How can we assess NATO’s two percent guideline from a demand for military expenditure perspective? ii) Has public debt hindered development in military expenditure of the 21 NATO allies in the European Union during 2014–2019?

The rest of the paper proceeds as follows: section two gives an overview of the background of the two percent guideline, describes the Defense Investment Pledge, and reviews some of the assessments concerning the two percent guideline. Section three gives an overview of military expenditure in Sweden and related debates. Section four presents the empirical methodology applied and the data used. Section five assesses the empirical estimations and discusses, whereas section five concludes.

Tracing the Two Percent Guideline Roots and the Motive for Its Choice

Intentions to introduce some form of financial guideline within NATO existed before the 2014 Wales Summit. One of the earliest attempts can be traced to as far as the late 1970s when allies were encouraged to increase annual military expenditure by three percent in real terms (Chalmers 2019; Lunn and Williams 2017).

The steps towards officializing the two percent guideline mainly surfaced during the 2000s. At the 2002 Summit in the Czech Republic, allies pledged to achieve the guideline, and repeated the pledge at the 2006 Summit in Latvia (Dowdy 2017; NATO 2005). It is, however, important to underscore that in the Czech Republic and Latvia, the allies did not set official deadlines for when the pledges were to be achieved.

There is a contestation as to why NATO chose two percent of GDP. One view relates the metric to decreasing military expenditure after the end of the Cold War. NATO chose it as half of the allies between the early 1990s and early 2000s had their military expenditure make at least two percent of GDP (Dowdy 2017). Another view is that NATO chose the metric as a guideline for countries that were aspiring to join the Alliance during the early 2000s. Since military expenditure in these countries was roughly 1.7% of GDP, two percent became a reasonable metric (Lunn and Williams 2017).

The Defense Investment Pledge

NATO formalized the two percent guideline during the 2014 Wales Summit.⁴ It is one of the two key goals highlighted by what is known as the Defense Investment Pledge. The second goal urges allies to commit at least 20% of their military expenditure on major equipment, with defense-related research and development considered.

Focusing on the two percent guideline, the Defense Investment Pledge encouraged allies whose military expenditure was at least two percent of GDP in 2014 to continue with the development. On the other hand, allies whose military expenditure was less than two percent of GDP at the time were to stop further reductions in their military expenditure. They were also encouraged to aim towards a real increase in their military expenditure as GDP grows. At the same time, they were encouraged to aim towards increasing their military expenditure to at least two percent of GDP by 2024 (NATO 2014).

An analysis of the wording used to describe the Defense Investment Pledge, such as aim and as GDP grows, offers no basis for the argument that the two percent guideline is binding. This potentially explains why some allies have chosen to approach it as non-binding despite American pressure. Countries such as Belgium (Ministry of Defense 2016), Germany (Buck 2018) and the Netherlands (Ministry of Defense 2018) have for instance indicated that their military expenditure will not make two percent of GDP in the foreseeable future.

Assessment of the Two Percent Guideline

The two percent guideline has mainly drawn criticism as its critics have highlighted several shortfalls it carries, thereby weakening its relevance. In a few cases, it has also won support. Some of these arguments are discussed below.

Since the U.S.A. is the de facto main protagonist in the debate, it is relevant to commence with it. Many sources often assert that the country accounts for a disproportionate share of NATO’s military expenditure (see e.g. Department of State 2017). Some have estimated the share to be as much as 70% of NATO’s total military expenditure (Cordesman 2018; Shelter-Jones 2017). This figure, however, is a comparison of the U.S. military expenditure, in absolute terms, with the aggregated military expenditure of the rest of the allies, which is also acknowledged by NATO (2018). By limiting the comparison to how allies directly contribute to NATO’s common budget, the disproportionate weight that the U.S.A. seemingly carries becomes much lighter (Shelter-Jones 2017). NATO (2018) indicates that the contribution of the U.S.A. to NATO’s common budget for 2018 and 2019 was around 22%, compared to 14.8% by Germany and 10.5% by France and the United Kingdom. Béraud-Sudreau and Giegerich (2018) argue that the direct contribution of the U.S.A. to European security was about USD 31 billion in 2017, which corresponded to 5.1% of the country’s total military expenditure that year. This figure also includes the contribution to NATO’s common budget. Other expenses it considers include the cost of American troops in Europe and military aid. However, the perception that the U.S.A carries a disproportionate financial burden within NATO is also understandable. For instance, European allies significantly decreased their military expenditure following the end of the Cold War (Techau 2015).

Another critic of the two percent guideline is that it favoured the U.S.A as its military expenditure already exceeded two percent of GDP in 2014. This also allowed the country space to decrease its expenditure despite the Wales declaration encouraging allies to halt any further decreases (Chalmers 2015).⁵

The two percent guideline has also faced scrutiny because the military expenditure of allies does not differentiate between national interests and those of NATO. The U.S.A for instance has a global presence unlike many smaller allies. It is also a major donor of military aid, which counts as part of its military expenditure. Hence, one can argue that it is unfair for the U.S.A. to benchmark on its level of military expenditure to push other allies to commit more resources. France and the United Kingdom, whose military expenditure, in absolute terms, follows that of the U.S.A. within NATO, also have military commitments that stretch beyond the Alliance (Béraud-Sudreau and Giegerich 2018; Grundman 2018; Mattelaer 2016).

Another discontent with the two percent guideline emphasizes that while it has put heavy emphasis on input, less attention has been given to output. For instance, it does not consider military spending in absolute terms, nor does it consider military capabilities and quality of the armed forces (Kalnins 2017; Techau 2015). Military expenditure, in absolute terms, of two percent of GDP also depends on the size of GDP and the business cycle. Periods of GDP contraction, for instance, imply that the two percent metric results in lower levels of absolute military expenditure (Lunn and Williams 2017).

Failure to consider output also implies that the two percent guideline does not reflect how allies contribute to NATO’s activities. Demark, for instance, contributed more soldiers than Greece to the Alliance’s presence in Afghanistan between 2010 and 2012. It also contributed more military planes than Greece during the 2011 NATO-led military operation in Libya (Deni 2015). The share of GDP committed to military expenditure also neither reflects the willingness of allies to participate in operations nor the willingness to undertake risks during operations. Canada and Denmark, for instance, showed more willingness to participate in more dangerous operations in Afghanistan (Deni 2015; Kalnins 2017; Techau 2015; Viggo and Ringsmose 2017). Nonetheless, Greece’s military expenditure has always been at least two percent of GDP, whereas that of Denmark has until recently floated below 1.3% of GDP. Canada’s military expenditure, in terms of GDP, has, on the other hand, been among the lowest in the Alliance (NATO 2019).

A further critique of the two percent guideline is that it does not consider efficiency, specifically how allies spend their military expenditure whether they achieve the guideline or not, which has also been recognized by NATO (Béraud-Sudreau and Giegerich 2018; NATO 2018; Shelter-Jones 2017). Personnel expenses, for instance, dominate most European allies’ military expenditure, which for some leaves less room for other key investments such as equipment (Shelter-Jones 2017).⁶ Only pensions accounted for 33% of Belgium and Portugal’s military expenditure in 2017. In France, the pension share was 24% (Béraud-Sudreau and Giegerich 2018). The share of military expenditure left for equipment also faces further pressure from factors such as limited competition and cooperation within the European defense industry. For instance, countries with defense industries tend to make purchases from their domestic markets. Germany’s 2016 Defense White Paper highlighted that among the consequences of this fragmentation is the burden it has placed on countries’ defense budgets (Federal Government 2016).

Finally, another critic of the two percent guideline concerns the unclear methodology that NATO relied on for its choice. The Alliance did not, for instance, rely on any known financial analysis to decide the optimal percentage of GDP that allies would need to commit to military expenditure for it to achieve its strategic goals (Techau 2015).

Views in support of the two percent guideline point out, for instance, that input is a prerequisite to achieve output, and thus an indicator of allies’ future defense capabilities (Becker 2017). Furthermore, after a period of peace dividend, especially among European allies, many armed forces within the Alliance require modernization. This is needed to reduce dependency on U.S.A. and to prepare the Alliance to be able to better defend itself. The guideline also contributes to addressing American grievances about free riding within the Alliance as it acts as one of the measures of burden sharing (Mattelaer 2017).

Additionally, the guideline re-awakened the debate on European security as it acts as a benchmark for assessing who is politically committed to NATO’s key assignment – to guarantee European security (Techau 2015). In this way, it is also used as a reference for accountability at the highest levels of government (Pothier and Vershbow, 2017).

Finally, the guideline is credited for halting decreases in military expenditure and for triggering increases. Since its introduction, the number of allies meeting the two percent guideline has increased from the three that did so in 2014 (Pothier and Vershbow 2017). (Figure 1) shows that seven countries fulfilled the threshold in 2019. Latvia, Lithuania, Poland, and Romania have gone as far as passing national laws or agreements to annually spend at least two percent of GDP on military expenditure (Hickey 2019).

The Case of Non-aligned Sweden

Sweden is part of the European Union but not a Member State of the NATO Alliance. Consequently, the country is only required to adhere to the fiscal regulations of the European Union, but not to NATO’s financial guidelines.

Sweden’s military expenditure, measured as a share of GDP, was constant at 1.1% between 2011 and 2016. It then decreased to 1.0% of GDP in 2017 and 2018 (SAFs 2019).⁷ However, despite Sweden not being part of NATO, the debate on increasing military expenditure to at least two percent of GDP has been gaining momentum in the country, both in the political and public domains (see, inter alia, Bruhn 2020, 2018; DI 2019; SVT, 2020; Swedish Parliament 2018). Of the eight political parties that make the 2018–2022 Parliament, at least five have expressed that military expenditure should make up a minimum of 2% of GDP (Bruhn 2020, 2018).⁸

Furthermore, Sweden’s Defense Commission suggested in 2019 that the defense budget should be gradually increased to SEK 84 billion from 2025. Based on the 2019 price level, this would translate into 1.5% of GDP from 2025 (Ministry of Defense, 2019). According to the government’s 2020 draft budget released in September 2019, official defense related nominal expenditure was projected to gradually increase from SEK 48.5 billion in 2018⁹ to SEK 71.7 billion in 2022 (Ministry of Finance, 2019¹⁰). The projected 2022 expenditure level is likely to translate into 1.29% of GDP,¹¹ putting it on the right trajectory in accordance with the Defense Commission’s recommendation.

Modelling Demand for Military Expenditure in the NATO–EU Context

Seventy percent of NATO allies, or 21 of 30, are also part of the European Union. Consequently, a demand for military expenditure model can be applied in view of the contesting financial demands of the two supranational institutions.

There is no specific economic theory that covers military expenditure as an economic activity (Nikolaidou 2008). However, based on different theories, literature recognizes various models that contribute to understanding decision-making and factors that influence countries’ military expenditure (Dunne, Nikolaidou, and Mylonidis 2003). One of the most common theoretical frameworks used to examine the demand for military expenditure relies on public finance theory and neoclassical models of rational agents maximizing a utility or welfare function – subject to a budget constraint (see, inter alia, Douch and Solomon 2014; Dunne, Nikolaidou, and Mylonidis 2003; Nikolaidou 2008; Smith 1995).

According to Smith (1995), a standard neoclassical model applied to explain the demand for military expenditure presumes that a country aims to maximize its welfare $(W$) – see equation (1)(1) $W=f\left(Y,S,N,Z_n\right)$(1) . This welfare is explained as a function of economic variables ($Y$), security ($S$), population ($N$) and other variables that explain shifts in the welfare function, e.g. policies of the government in power. The welfare maximization problem such as the one in equation (1)(1) $W=f\left(Y,S,N,Z_n\right)$(1) is subject to a budget constraint and the security function – see equation (2)(2) $Y=C{p}_c+M{p}_m$(2) . In equation (2)(2) $Y=C{p}_c+M{p}_m$(2) , $Y$ is nominal aggregate income, $C$ is real consumption, $M$ is real military expenditure, whereas $p_c$ and $p_m$ are the respective prices.

(1) $W=f\left(Y,S,N,Z_n\right)$(1)

(2) $Y=C{p}_c+M{p}_m$(2)

Price deflators have been excluded from previous studies (see, inter alia, Dunne et al. 2003; Nikolaidou 2008) due to lack of separate price deflators for military goods for most countries. Furthermore, previous studies (see, inter alia, Dunne, et al. 2010; Douch and Solomon 2014; Nikolaidou 2008; Wang 2013) have used GDP as a measure for the income variable. Instead, this paper aims to extend and validate the emerging strand of the literature that uses fiscal indicators, instead of GDP, as measures of income in neoclassical models of military expenditure demand. This is in line with Christie (2019), which uses a fiscal space indicator based on public debt to model military spending decisions and progress towards the two percent guideline among NATO allies in the European Union. The study also emphasizes the need to give fiscal indicators more attention in studying demand for military expenditure, particularly for European Union Member States. Other recent studies that have considered fiscal indicators as explanations for demand for military expenditure include Becker (2019), Blum and Potrafke (2020) and Odehnal and Neubauerb (2020).

Dynamic Panel Regressions

This paper suggests the following dynamic panel model(s) to explain demand for military expenditure among NATO allies in the European Union.

(3) $Me{x}_{it}={\beta }_O+B_{1}Me{x}_{it-1}+{\beta }_{2}Deb{t}_{it-1}+B_3{V}_{it-1}+{\epsilon }_{it}$(3)

(4) $Me{x}_{it}-Me{x}_{it-1}={\beta }_O+(B_1-1)Me{x}_{it-1}+B_{2}Deb{t}_{it-1}+B_3{V}_{it-1}+{\epsilon }_{it}$(4)

$\left(i=1,\dots ,N\right),\left(t=2014,\dots ,T\right)$, $|B_1|<1$

$Me{x}_{it}$ refers to military expenditure as a share of GDP in the country $i$ during year $t$. $Deb{t}_{it}$, as a measure of income, is the explanatory variable of interest. $V_{it}$ is a matrix of control variables that can influence demand for military expenditure. Previous studies have for instance controlled for government expenditure minus military expenditure as a share of GDP, with the intention of capturing the opportunity cost of military expenditure. Population is also often included to control for military expenditure as a public good, i.e. military expenditure should not increase as population increases. Additionally, previous studies have considered net exports as a ratio of GDP, or trade balance, to control for the openness of the economy. Threat factors and spillins also feature in models to control for the influence of perceived insecurity and being part of an alliance (see, inter alia, Dunne et al. 2003; Nikolaidou 2008).

The difference between model (3) and model (4) is the subtraction of $Me{x}_{it}$ from both sides of model (3), which results in a dynamic model of military expenditure in the latter. Djurfeldt and Barmark (2011) argue that this allows a return to the mean. In this paper, a possible implication is that whereas there could be volatility in military expenditure in the short run, it gains a more stable status in the long run. Otherwise, Djurfeldt and Barmark (2011) indicate that the two models give similar estimations except for the coefficients of the dynamic term, i.e. $Me{x}_{it-1}$. However, the coefficient of the lag of military expenditure in model (4), given as $\left(B_1-1\right)$, can be derived by knowing $B_1$ from model (3). Since estimating the two models presents similar results, this paper only estimates and assess the results from model (3).

The independent variables enter the models in lag form to control for simultaneity (Djurfeldt and Barmark 2011). For instance, it is possible that military expenditure affects public debt, but the opposite could also be true. By lagging the debt variable, the models control that military expenditure in period $t$ cannot affect public debt in period $t-1$. However, it is plausible that debt status in period $t-1$ can affect decisions concerning military expenditure in period $t$. Furthermore, by allowing time lag, the models enable time span to capture countries’ spending adjustments in line with the European Union’s fiscal rules.

Dynamic modelling also allows controlling for the fact that development in military expenditure among the studied group of countries affects itself over time as allies strive to achieve the two percent guideline. Thus, if since 2014 allies are making efforts to progress towards the two percent guideline, their previous level of military expenditure theoretically influences the speed of change in their military expenditure as they approach the set deadline.

Empirical Model for Estimation and Data

(5) $Me{x}_{it}={\beta }_O+B_{1}Me{x}_{it-1}+{\beta }_{2}Deb{t}_{it-1}+B_{4}RussiaMe{x}_{it-1}+B_{3}USMe{x}_{it-1}+{\beta }_{5}GDPgrowt{h}_{it-1}+{\epsilon }_{it}$(5)

Model (5) considers that the development in military expenditure of the current 21 NATO allies that are part of the European Union can be explained by its previous level and economic conditions, where debt is of most interest in this paper, but also controlling for GDP growth. Controlling for GDP growth is important, especially considering that the Wales declaration encourages allies to increase their military expenditure in line with real GDP growth. Russia’s military expenditure is included to control for the potential impact of external threat, whereas the military expenditure of the U.S.A. is included to capture potential spillin effects. Russia’s military expenditure as a measure for threat measure is explained by the country’s military modernization program which in recent years has been backed by increasing military expenditure (Wezeman 2020). Furthermore, Russia’s military expenditure makes a good measure for threat considering that the Ukraine crises is one of the key reasons why NATO agreed in 2014 to increase defense investment (Pothier and Vershbow 2017; Techau 2015). Finally, the error term, ${\epsilon }_{it}$, captures other possible factors that could explain demand for military expenditure, such as those highlighted in subsection 4.1. (Table 1) shows a summary of the variables and their data sources.

Table 1. Variables and data sources

Label	Variable	Unit	Source
$Mex$	Military expenditure (NATO allies in the EU)	% GDP	NATO (2019)
$Debt$	General government consolidated gross debt (NATO allies in the EU)	% GDP	Eurostat (2020)
$USMex$	U.S military expenditure (measure of spillin effect)	% of GDP	NATO (2019)
$RussiaMex$	Russia’s military expenditure (measure of threat effect)	% of GDP	SIPRI (2020)
$GDPgrowth$	Real GDP growth (NATO allies in the EU)	% real change	IMF (2019)

(Table 2) shows summary statistics of military expenditure and public debt, which are the variables of main interest. (Figure 1) on the other hand, shows the status of the studied group of countries vis-à-vis these variables. The figure, for instance, shows that only six countries or 29% of the studied group of countries, fulfilled the financial goals of both NATO and the European Union in 2019. These include Bulgaria, Estonia, Latvia, Lithuania, Poland, and Romania. Greece, on the other hand, fulfilled NATO’s two percent guideline but with a debt-to-GDP ratio of 178%, it did not meet the debt criterion of the European Union. Furthermore, in 2019 most countries had their military expenditure below 1.55% of GDP. However, with some of them having a debt-to-GDP ratio below 60%, they possibly have fiscal space to enable then increase investment in military expenditure.

Figure 1. Military expenditure and public debt in 2019 (Eurostat 2020; NATO 2019).

Table 2. Summary statistics for military expenditure and public debt

Variable	Mean	Standard deviation	Minimum	Maximum	Observations
Military expenditure, share of GDP	1.38	0.49	0.38	3.25	126
Debt-to-GDP ratio	68.43	41.69	8.40	181.20	126

Note. The sample data covers 21 NATO allies in the European Union, covering the period 2014–2019 (Eurostat, 2020; NATO 2019).

Empirical Estimations

Model (5) is estimated using the generalized method of moments (GMM) suggested by Arellano and Bond (1991), using a balanced panel covering the period 2014–2019. Table 3 reports the empirical estimations performed using a one-step GMM. They show consistency, especially considering that the coefficients of the lag of military expenditure and public debt have the expected signs and are significant.

Table 3. Arellano-Bond (Robust S.E.) one-step GMM estimation

Variable	Model 1	Model 2	Model 3
Lag Mex.	0.663*** (0.196)	0.574* (0.295)	0.579* (0.279)
Lag Debt	−0.028*** (0.361)	−0.024* (0.012)	−0.024* (0.012)
Lag Russia Mex.		−0.034 (0.021)	−0.040 (0.024)
Lag U.S Mex.		−0.053 (0.217)	−0.131 (0.220)
Lag GDP growth			−0.023 (0.014)
Constant	2.444*** (0.847)	2.646* (1.115)	3.015*** (1.161)
Sargan test of overidentifying restrictions	10.155 p = 0.388	8.709 p = 0.465	8.810 p = 0.455
Arellano-Bond test for AR(1)	−1.113 p = 0.266	−1.014 p = 0.311	−1.181 p = 0.238
Arellano-Bond test for AR(2)	1.313 p = 0.189	1.466 p = 0.143	1.381 p = 0.167
No. of observations	84	84	84

Note. ***significant at 1 percent level, * significant at 5 percent level.

The estimations suggest that for the 21 NATO allies in the European Union during 2014–2019, development in military expenditure has been positively and significantly influenced by the previous level of military expenditure. The estimations also suggest that developments in military expenditure have been negatively and significantly affected by public debt. Due to the lack of significant evidence, models (2) and (3) suggest that development in military expenditure of NATO allies in the European Union has not been impacted by the military expenditure of Russia and the U.S.A. Consequently, there is no evidence of threat- and spillins impact from the two countries. Additionally, due to lack of significant evidence, model (3) does not suggest that the real GDP change or real economic growth has affected development in military expenditure of the studied group of countries.

Dynamic panel estimation using GMM requires performing specifications for overidentification and serial correlation (Arellano and Bond 1991; Labra and Torrecillas 2018). According to the Sargan tests presented in (Table 3), there is no evidence to suggest that the three estimated models suffer from overidentification. Additionally, based on the Arellano-Bond tests for AR (2) shown in (Table 3), there is no evidence to suggest that the error terms face serial correlation.

Discussion

The aim of this paper is partly to contribute to the debate on the two percent guideline. Consequently, the most significant implication of the empirical findings presented in the debate is that fiscal ability to increase military expenditure across the Alliance requires consideration, or at least should not be ignored.

Pushing allies to guarantee military expenditure a share of national resources could potentially force some of them to give other sectors of public expenditure less priority depending, inter alia, on the domestic economic cycle. This carries the risk of decreasing public support for NATO and contributing to political rhetoric against the Alliance. It also risks causing friction between the European Union and NATO as some countries that belong to both blocs could blame the failure to meet the former’s fiscal goals on the defense Alliance’s financial demands.

The restrictions imposed by the European Union’s fiscal regulations also risk encouraging some allies to resort to ‘creative forms’ of accounting in a bid to achieve and maintain the two percent guideline. This has already surfaced as the former European Union Member State, the United Kingdom, in 2015 changed the method used to calculate the military expenditure it reports to NATO. By doing so, expenses that previously were not counted as part of military expenditure were added (Defense Committee 2016). This enabled the country to continue achieving the two percent guideline according to NATO but not according to sources such as the Stockholm International Peace Research Institute and the International Institute for Strategic Studies (Bond and Tetlow 2018; SIPRI 2020). Germany’s former Foreign Minister, Sigmar Gabriel, also suggested in 2017 that expenses such as development aid should be considered as part of the guideline (Bartunek and Blenkinsop 2017; Cermak and Steingart 2017).

If NATO allies were to achieve the two percent guideline partly by adjusting their accounting methods, there is a need for reflection within the Alliance if this approach contributes to its key intention of strengthening defense capabilities. Consequently, the Alliance may need to revise ways that give allies incentives to genuinely increase their military expenditure. Alternatively, the Alliance needs to reflect on whether the focus should include guiding allies on how to acquire the right capabilities, without necessarily focusing on the two percent monetary measure. This could, for instance, be through strong encouragement of efficient expenditure means within the Alliance.

Conclusions

The key aim of this paper has been twofold, namely, to contribute to the debate on NATO’s two percent guideline and to contribute to the literature on demand for military expenditure. This has been approached by assessing how NATO’s two percent guideline can be viewed from a demand for military expenditure perspective and whether public debt has hindered development in the military expenditure of NATO allies in the European Union. Consequently, the paper has proposed and estimated a dynamic panel model. By using debt-to-GDP ratio instead of GDP or GDP growth, as is commonly the case, it has added to the growing literature that considers fiscal indicators in neoclassical models of demand for military expenditure. Through one-step GMM estimation, the paper establishes that public debt could have constrained development in military expenditure among the 21 NATO allies in the European Union during 2014–2019. These results suggest that fiscal conditions require attention in the debate on the two percent guideline. This is important especially if the Alliance is to best achieve its intended objectives.

Acknowledgments

I would like to extend my appreciation to the peer reviewers for their suggestions that were key in advancing this paper. I would also like to thank Peter Nordlund, who has coordinated the process, and offered me encouragement throughout.

Disclosure statement

The views expressed herein are those of the author(s) and do not necessarily reflect the views of the United Nations.

Additional information

Funding

This work was supported by the Totalförsvarets Forskningsinstitut.

Notes

1. Henceforth referred to as the two percent guideline.

2. Henceforth referred to as the debt-to-GDP ratio.

3. ERM II aims to achieve a stable exchange rate between the currency of a non-euro-area country and the euro as well as a stable monetary policy, in preparation for acquiring the euro.

4. The Wales Summit is widely considered to have been the official inauguration of the two percent guideline as the final declaration set a time frame for its achievement. However, NATO (2018) argues that the Alliance agreed on the two percent guideline already in 2006.

5. NATO (2019) indicates that the military expenditure of the U.S.A. decreased from 3.73 percent of GDP in 2014 to 3.30 of GDP in 2018.

6. Personnel expenses, for example, on average accounted for more than 70 percent of Greece’s military expenditure between 2012 and 2018 whereas equipment accounted for 11 percent.

7. The Swedish Armed Forces (SAFs) refer to data retrieved from SIPRI.

8. Among the political parties that have expressed support for increasing military expenditure to at least two percent of GDP include the Christian Democrats, Centre Party, Liberals, Moderates, and the Swedish Democrats.

9. According to SIPRI (2020), Sweden’s military expenditure in 2018 was SEK 50 billion, whichis relatively consistent with the official figures.

10. See expenditure sector 6, p. 13.

11. IMF (2019) forecasts Sweden’s GDP in 2022 to be SEK 5,553 billion.