Do Foreign Military Deployments Provide Assurance? Unpacking the Micro-Mechanisms of Burden Sharing in Alliances

Abstract

How do US foreign military deployments impact the defense policies of host states? Dominant scholarship holds that these deployments play a pivotal role in assuring allies that their security is guaranteed, which in turn leads host countries to neglect their national defense contributions. In this research note, we examine the micro-foundations of this conventional wisdom, investigating how nuclear and conventional troop deployments impact attitudes toward defense policies in (potential) host states. We highlight that the presumed linkage between assurance and free-riding critically implies that foreign military deployments must positively affect perceptions of security among host nations. We test this core logic, alongside some alternative pathways, at the micro level through two survey experiments that randomize hypothetical withdrawals (Germany) and deployments (Czech Republic). Although we find some evidence that foreign military deployments can decrease citizens’ subjective need for defense, the survey experiments also reveal that citizens hardly feel protected by these deployments. Thus, our results cast doubt on the core logic underlying the theory of free-riding in alliances.

¿Cómo afectan los despliegues militares estadounidenses en el extranjero a las políticas de defensa de los Estados anfitriones? El conocimiento académico imperante argumenta que estos despliegues desempeñan un papel fundamental para asegurar a los aliados que su seguridad está garantizada, lo que a su vez lleva a los países anfitriones a descuidar sus contribuciones a nivel de defensa nacional. En esta nota de investigación, examinamos los microfundamentos de esta sabiduría convencional, investigando cómo los despliegues de tropas, tanto nucleares como convencionales, influyen sobre las actitudes con relación a las políticas de defensa en los (potenciales) Estados anfitriones. Destacamos que el presunto vínculo existente entre la garantía y la dependencia implica, fundamentalmente, que los despliegues militares extranjeros deben afectar positivamente a las percepciones de seguridad entre las naciones anfitrionas. Ponemos a prueba esta importante lógica, junto con algunas vías alternativas, a nivel micro a través de dos experimentos de encuesta que aleatorizan hipotéticas retiradas (Alemania) y despliegues (República Checa). Aunque encontramos algunas pruebas de que los despliegues militares en el extranjero pueden llegar a disminuir la necesidad subjetiva de defensa de los ciudadanos, los experimentos de la encuesta también revelan que los ciudadanos apenas se sienten protegidos por estos despliegues. Por lo tanto, nuestros resultados ponen en duda la lógica central que subyace a la teoría de la dependencia en las alianzas.

Quelle est l'incidence des déploiements militaires à l’étranger des Américains sur les politiques de défense des pays cibles ? Les travaux de recherche dominants affirment que ces déploiements jouent un rôle essentiel quand il s'agit d'assurer aux alliés que leur sécurité est garantie et que, par la suite, ces mêmes pays négligent leur contribution à la défense nationale. Dans cette note de recherche, nous examinons les microfondations de cette explication traditionnelle, en nous intéressant à l'incidence du déploiement de l'arme nucléaire et de troupes conventionnelles sur les attitudes à l’égard des politiques de défense dans les États cibles (potentiels). Nous mettons en évidence que le lien présumé entre l'assurance et le parasitisme implique que les déploiements militaires étrangers aient une incidence positive sur les perceptions de sécurité dans les nations cibles. Nous testons cette logique fondamentale, mais aussi quelques alternatives, au niveau micro par le biais de deux expériences de sondage qui rendent aléatoires des retraits (Allemagne) et déploiements (République tchèque) hypothétiques. Bien que nous trouvions certains éléments pour venir étayer le fait que les déploiements militaires étrangers puissent réduire le besoin subjectif des citoyens en matière de défense, les expériences de sondage révèlent également que les citoyens ressentent très peu la protection de ces déploiements. Aussi nos résultats jettent-ils le doute sur la logique fondamentale qui sous-tend la théorie du parasitisme au sein des alliances.

Introduction

How do foreign military deployments impact the defense policies of host states? While deployments of both US troops and nuclear weapons on the soil of North Atlantic Treaty Organization (NATO) allies have been a reality for decades, debate about allies’ defense contributions continues to be a source of friction. Exemplified by former US President Trump’s announcement in 2020 to withdraw significant numbers of military personnel from Germany as a response to alleged free-riding, insinuations persist that allies take for granted that their security is guaranteed by the United States.

Indeed, threatening to withdraw military deployments, or even demanding financial compensation for troop deployments, has been a prominent tool to put pressure on allies since the beginning of the Cold War (e.g., Gavin 2012, 41–3; Blankenship 2021, 719–20). The consequences of foreign military deployments recently attracted renewed attention after the Russian attack on Ukraine, which triggered an increase in US military deployments in Europe (Powell 2022). Critics, such as Posen (2022), were quick to warn that an enhanced presence of US military forces in Europe would undermine recent pledges by European allies to step up their own defense efforts, and instead foster a familiar pattern of free-riding by allowing them to hide behind US protection.

Do foreign military deployments induce a sense of security that leads host countries to discount their own defense? Existing scholarship has claimed that these deployments play a pivotal role in assuring allies (e.g., Yost 2009, 767). While some maintain that this sense of security is crucial for realizing US strategic objectives, such as non-proliferation (e.g., Lanoszka 2018), others contend that a substantial US military presence undermines collective burden sharing.¹ Both perspectives take for granted that US military deployments generate assurance. Although recent studies have explored this issue at the elite level (Blankenship and Lin-Greenberg 2022), there remains a lack of studies investigating the extent to which citizens feel assured by military deployments. This is astounding given that virtually all NATO member states are democracies in which political decisions are influenced by public sentiment (Reiter and Stam 2002; Kertzer and Zeitzoff 2017).

Research examining alleged allied free-riding as a result of these deployments is characterized by mixed findings, possibly due to a neglect of micro-mechanisms (e.g., Martinez Machain and Morgan 2012; Allen, VanDusky-Allen, and Flynn 2016, Allen, Flynn, and VanDusky-Allen 2017; Jakobsen 2018; Jakobsen and Jakobsen 2019). In fact, free-riding critically hinges on the idea that allies are assured. Thus, contributing to a more complete understanding of the consequences of foreign deployed military, this research note provides micro-level evidence on (1) whether foreign military deployments assure and (2) how (non-)assurance affects attitudes toward national defense.

As initial tests of existing theories, we run two survey experiments that examine the effects (1) of withdrawals (in Germany) and (2) deployments (in Czechia). Consistent with prior studies on the theory of free-riding in alliances, the findings indicate that in the German survey experiment, the withdrawal of US military deployments increases citizens’ subjective need for national defense efforts. However, our results also indicate a more complex relationship once assurance is brought into the picture: In both Germany and Czechia, we fail to find evidence that the presence of military deployments actually reduces host state citizens’ threat perceptions, i.e., provides assurance. Consequently, while our findings are characterized by mixed results, they provide preliminary evidence that diverges from the conventional perspective. This evidence questions the widely held belief that foreign military deployments lead to free-riding by creating a sense of protection, at least among the general population.

The Consequences of Foreign Military Deployments

The United States has deployed both its conventional forces and nuclear weapons across the globe since the end of World War II. Today, within NATO, US troops are deployed to at least fifteen allies. Nuclear weapon deployments remain in five states, all of which are members of NATO: Belgium, Germany, Italy, the Netherlands, and Turkey. Nuclear deployments tend to coincide with troop deployments (see figure 1). However, as exemplified by the United Kingdom, a former nuclear host state, this need not be the case.

Military deployments are commonly viewed as tools of extended deterrence, which is principally regarded to suffer from a credibility problem that arises because fighting a costly war on behalf of the protégé may not be in the guardian's immediate interest. To render this promise of protection more credible, and thereby deter the adversary, Fearon (1997) argues that guardian states can engage in (1) hand-tying, which creates unavoidable costs once a conflict breaks out, and/or (2) sunk costs expensed ex-ante, in order to credibly signal to allies and adversaries that they will follow through with their promise to protect another state. Military deployments fulfill both roles. Upfront (sunk) costs represent a costly, and thus credible, signal to adversaries. Moreover, these deployments constitute hand-tying because of their potential tripwire effect that would “blow up” if the US military abroad came under attack (Schelling 1966).²

Assurance–Free-Riding

By lending credibility to security guarantees, foreign military deployments have often been said to assure allies (e.g., Fuhrmann and Sechser 2014a, 461–2). For instance, Lanoszka (2018) argues that the United States has strategically used both conventional and nuclear forces to strengthen non-proliferation efforts. The presence of US military forces is widely seen as a direct and tangible indicator of the security alliance between the United States and its allies, a factor deemed irreplaceable by forces positioned outside of allied nations (Yost 2009, 764). Both foreign troops and nuclear weapons can act as a powerful deterrent against potential aggressors; their physical presence signifies a high level of commitment from the guardian state, suggesting that any attack on the protégé will be met with a strong response (Schelling 1966). Moreover, military deployments increase warfighting capabilities should deterrence fail. In particular, nuclear weapons are unmatched by anything the non-nuclear ally can add.

Existing research has largely taken for granted the view that military deployments generate a sense of security among host nations. Yet, the concept of assurance remains understudied (Knopf 2012, 376–8; Blankenship 2020, 1017). This is puzzling given a vast literature that focuses squarely on the consequences of assurance, debating whether US security guarantees in general undermine collective defense within the alliance, commonly referred to as the theory of free-riding in alliances (e.g., Olson and Zeckhauser 1966; Sandler and Forbes 1980; Lake 2009; Brands and Feaver 2017; Jakobsen 2018). This view holds that allies piggyback on US protection and therefore neglect national efforts to invest in defense and deterrence. Accordingly, Posen (2018) and others have warned explicitly against increasing deployments of the US military to allied states, most recently after the Russian invasion of Ukraine in 2022.

A similar rationale underlies threats to withdraw deployments put forward by many US presidents: withdrawals reduce assurance and increase insecurity, which in turn leads allies to step up defense efforts. Abandonment and entrapment represent the primary challenges within alliances, as outlined by Snyder (1984). Consequently, the prospect of abandonment underpins every alliance interaction. This is especially relevant in asymmetric alliances, where the guardian can use it to pressure its protégés (Blankenship 2021). Notably, allied security policy elites have historically been more sensitive to concerns of abandonment than to possible domination by their guardian (Schuessler and Shifrinson 2019, 40). Research has shown that the removal of US military forces from allied nations can incentivize these countries to invest in their own armaments (Avey, Markowitz, and Reardon 2018). However, the impact may be moderated by the allies’ vulnerability to attacks (Blankenship 2021). In short, existing theory suggests that the rationale is bidirectional and pertains to both deployments and withdrawals. We therefore state the following hypothesis, visualized in figure 2:

 

In line with this conjecture, Martinez Machain and Morgan (2012) find that an increase in US troops lowers host state troop levels. In addition, relying on cross-national survey data, Jakobsen (2018) show that among European NATO allies, the willingness to fight for one's own country has been remarkably low since the end of the Cold War, especially so in countries that host US military bases or significant numbers of US troops (Jakobsen and Jakobsen 2019). However, Allen, VanDusky-Allen, and Flynn (2016) do not find evidence of free-riding induced by US troop deployments within NATO. Moreover, research has shown that defense spending is substantially influenced by perceptions of regional threats which can mitigate the effect of US troop deployments (Allen, Flynn, and VanDuskyAllen 2017; Ma and Kang 2023). These mixed results match the controversy over when free-riding within NATO is or was a reality (e.g., Sandler and Forbes 1980; Bogers and Beeres 2013; Zyla 2016; Jakobsen 2018; Cooper and Stiles 2021).

In sum, the evidence pertaining to free-riding as a result of US military deployments remains inconclusive. Moreover, there is a notable gap in research on the extent to which these deployments provide assurance to allies, especially at the micro level. This is significant given the implicit role of assurance in research linking foreign military deployments to national defense outcomes.

Alternative Pathways: Resolving Insecurity

Motivated by the mixed findings in previous research, we also consider literature postulating that these deployments may not decrease, but actually increase threat perceptions, i.e., induce insecurity. Host citizens could fundamentally reject the use of force as a means to solve political problems altogether, or suspect the guardian to engage in self-serving behavior, e.g., use military deployments to protect itself rather than the host, project power, or pursue other geo-strategic objectives that do not align with those of the host. Howard (1982, 316) argues that citizens in NATO states during the Cold War believed that Washington was pushing a war onto them “in which the Europeans have no interest and from which they will be the first to suffer.” More recently, Sukin (2020) suggests that credible nuclear security guarantees by a guardian can scare the protégé's population. This is because demonstrations of strength can not only deter but also fuel escalation.

Acknowledging that US military deployments may fail to assure allies or even backfire and induce insecurity calls for alternative causal pathways connecting military deployments to citizens’ subjective need for national defense. Drawing on extant research, and without claims to completeness, we here highlight two such pathways, tentatively entitled insecurity–de-escalation and insecurity–self-insurance. While these pathways take seriously the possibility that foreign military deployments may actually increase threat perceptions, they differ in terms of how such insecurity translates into subjective needs for national defense.

Insecurity–De-Escalation

The assurance–free-riding logic hinges on the assumption that guardian states offer protection, typically by means of deterring adversaries. However, it remains unclear whether citizens actually embrace the very concept of deterrence, which is far from trivial. If citizens reject the logic of deterrence—or the use of force more generally—then military deployments may be viewed as contributing to dangerous arms races and a spiral of violence. As Rathbun et al. (2016, 126) point out, doves often believe that displays of power “risk inciting fear on the other side and escalating hostilities in such a way that leaves both sides worse off.” This view focuses squarely on the offensive potential of military deployments: how would an adversary distinguish between genuine deterrence measures and covert preparations for an attack (Tetlock, McGuire, and Mitchell 1991, 246; Knopf 2012, 384)?

Thus, critics of military deployments claim that “they are provocative to adversaries” (Avey 2017, 89–90). Many scholars and commentators argue that both nuclear (Kristensen 2018; Countryman and Reif 2019) and conventional deployments (Poast and Reiter 2021, 52; Kühn 2018, 25) could spur, rather than deter, conflict. Accordingly, foreign military deployments may not decrease, but rather increase threat perceptions, thereby inducing a sense of insecurity. From a dovish viewpoint, the same holds true for national defense, which also contributes to fear, provocation, and escalation. Consequently, to counter any externally induced insecurity and in an effort to de-escalate, states could reduce their own national defense. Accordingly, we state the following hypothesis, also visualized in figure 3 :

 

Insecurity–Self-Insurance

De-escalation by means of lessened national defense is far from the only plausible response to insecurity induced by the foreign deployed military. Citizens may also wish to counter perceived threats by increasing national defense efforts. Sukin (2020, 1038) argues that credible security guarantees may induce fears, leading citizens to place more emphasis on their own military “to remove responsibility over security crises from the nuclear ally and to regain domestic control over nuclear decision-making.”

If foreign deployments are generally viewed as fostering the risk of entrapment, then national security fundamentally hinges on national military. Such isolationism “disapproves of both foreign military engagements and collaborative multilateral efforts to solve global problems,” but recognizes “the importance, effectiveness, and desirability of using force to reach foreign-policy objectives” (Rathbun et al. 2016, 125–6). This stance is common among far-right parties across Europe.³ Thus, unlike in the insecurity–de-escalation pathway, the response to insecurity in this pathway is not principally dovish. Rather, defensive alliances, and especially foreign-deployed military, are perceived as a risk that is best mitigated by taking back control through an increased investment in national defense. We can therefore derive a third insecurity–self-insurance hypothesis, visualized in figure 4:

 

In summary, existing research suggests different pathways connecting foreign-deployed military and the subjective need for national defense. The conventional view assumes that these deployments assure host states, and thereby render national defense superfluous due to guardian protection, leading to free-riding. Contrasting views highlight that military deployments could fail to assure citizens, and instead make them feel less secure. This insecurity could be countered either by means of de-escalation (disarmament) or self-insurance (bolstering national defense).

Further Considerations

So far, our expectations hardly distinguish deployments by type. In principle, nuclear deployments could be expected to carry larger effects than their conventional counterparts. They are particularly costly signals and thus might induce greater feelings of security. Alternatively, such deployments could also entail the risk of nuclear escalation, which should heighten threat perceptions. Indeed, while nuclear strategists view nuclear weapons as deterrence tools, public fears of nuclear escalation are unlikely to be appeased by such theoretical deliberations. On the flip side, research has shown that the impact of nuclear deployments is less significant than often perceived (e.g., Sechser 2017, 451; Fuhrmann and Sechser 2014b; Avey 2017). Consequently, we do not further delineate between conventional and nuclear deployments in our hypotheses.

For similar reasons, we also do not distinguish between “offensive” and “defensive” systems. An extensive debate about offensive and defensive aspects of military systems in international relations (e.g., Quester 1977; Glaser and Kaufmann 1998; Lieber 2000) shows that most individual systems can be used for both offensive and defensive operations (Lynn-Jones 1995, 672–7). Moreover, applied to nuclear weapons, it seems questionable if citizens would feel less threatened by nuclear anti-aircraft deployments (used against attacking aircraft) than nuclear bombs.

Survey Experiments

Ultimately, whether foreign deployed military induces assurance or insecurity, and how this translates into national defense efforts, is an empirical question—for which this research note aims to contribute preliminary evidence. Thus, in this section, we provide initial experimental evidence from two survey experiments on the consequences of US foreign military deployments to allied states, focusing squarely on micro-level threat perceptions and attitudes on national defense. The goal is to test the hypotheses laid out above. To this effect, we experimentally manipulated changes in US military deployments, both nuclear and conventional. In a nutshell, we treated parts of our samples by informing them about different types of (hypothetical) withdrawals (Germany; May/June 2022) and deployments (Czech Republic; September 2022). We then measured threat perceptions and attitudes toward national defense.

This approach has several key advantages. By contrasting the responses of the treated with those of the control groups, we are not only able to scrutinize current theories linking foreign-deployed military and national defense but also to open the black box and examine the underlying pathways: do the effects stem from perceived assurance or insecurity? Moreover, we are able to circumvent the problem of endogeneity that arises if the US systematically deploys or withdraws its military from allies that either contribute sufficiently or underprovided toward common defense (e.g., Blankenship 2021, 697). Although the survey experiments per se cannot solve the problem of non-random assignment of foreign military deployments between countries, experimentally manipulating the treatment of foreign military deployments within a given country allows us to study the underlying mechanisms more closely. By randomly assigning scenarios in our experiments, we are able to observe how the same population in a host and a non-host state react to the presence or absence of US military deployments.

Underlying Assumptions and Scope Conditions

Our approach firmly rests on the assumption that public opinion influences foreign and security policies, at least in democracies (Kertzer and Zeitzoff 2017). Research has found that citizens hold informed views about foreign policy, that these views are congruent with political decisions, and that the causal pathway leads from public opinion to political outcome, rather than the other way around (Page and Shapiro 1983; Aldrich et al. 2006). Moreover, voters care deeply about foreign policy issues, while political leaders and non-elected officials avoid taking foreign policy decisions that are not supported by the public (Tomz and Weeks 2013; Lin-Greenberg 2021, 850). Even in situations where citizens possess limited knowledge about the subject at hand, they still form coherent foreign policy views based on their basic values (Rathbun et al. 2016). Finally, historical evidence suggests that organized public sentiment in the form of protests can influence a state's nuclear policy (e.g., Cooper 2023).

For these reasons, the evidence presented here speaks primarily—if not exclusively—to (potential) host countries that feature a democratic political system. Moreover, especially with regard to nuclear sharing, shared alliance status with the United States (i.e., within NATO) is another real-world scope condition, given that no other nuclear power in the Western world deploys its arsenal abroad, and that US conventional capabilities are unmatched by any other potential guardian.

Conversely, the findings are less applicable to military deployments carried out by the Soviet Union (Fuhrmann and Sechser 2014a), or the ongoing Russian operations in Belarus (Karmanau 2023). These differ significantly from contemporary US deployments in several distinct areas, including the degree of institutionalization, or the degree of influence held by host states (e.g., Nielsen 1998). Moreover, we assume a level of responsiveness of political decision-making to public opinion, which is not the case in authoritarian states—that is, countries that have received Soviet and Russian deployments (e.g., Gandhi and Lust-Okar 2009; Svolik 2012; King, Pan, and Roberts 2013). Fittingly, Conybeare (1994) and Alley (2021) fail to find evidence of free-riding in terms of military spending within the Warsaw Pact.

Case Selection

Given the lack of previous research, our approach employs a diverse case research design, selecting Germany and the Czech Republic based on their contrasting attributes (Seawright and Gerring 2008, 300–01). In theory, this approach carries high external validity by representing the full range of the population of interest, i.e., US NATO allies. While it would have been desirable to conduct experiments in every NATO member state, resource constraints prevented us from doing so. We therefore opted for this exploratory approach, hoping to gain initial insights on the effects of US foreign military deployments within countries, as well as between them.

We focus on three dimensions of diversity between countries: (1) historical US military presence, (2) culture, and (3) economic and military capabilities. First, historical experience of military deployments in one's country arguably affects the perception of such deployments. For instance, hosting US military forces for extended periods might lead to normalization effects, as citizens engage with foreign forces. Conversely, negative sentiment could also emerge over time. Second, cultural dynamics are crucial. Citizens’ views on US military deployments often reflect their trust in the United States and perceived intentions. Intergovernmental relations, cultural exchanges, and personal interactions significantly impact these perceptions. Moreover, past experiences with Russia are likely to influence these attitudes as well. Some NATO members have closer ties with Russia than with the West, at least historically. Third, economic and military strengths are vital in assessing military deployments. Deployments primarily aim to protect host states, but some are more capable of self-defense due to their resources.

Germany boasts one of the largest enduring US military presences, both conventional and nuclear. As a member of NATO since 1955, its cultural identity has been significantly shaped by Western Europe and the United States. Furthermore, Germany's Gross Domestic Product (GDP) and population size are notably high relative to other NATO members, which in theory renders Germany an unlikely candidate for free-riding. In contrast, Czechia as a former Warsaw Pact country, is culturally and historically rooted in Central and Eastern Europe. Having joined NATO in 1999, it has never hosted a US military presence. Additionally, Czechia's GDP and population size are modest compared to other NATO members.

Admittedly, several countries share a socio-cultural history comparable to that of Czechia. However, among these, Poland and Bulgaria have been hosts to US troops, and the Baltic states have a robust NATO military presence—attributes crucial to our diverse case research design. Moreover, Romania is notable for its US missile defense system. Thus, Czechia is an ideal case to examine how history with minimal US military deployments, in this instance none, influences the impact of potential future military deployments.⁴ Hungary, another possible Central and Eastern European case has unique ties with Russia, leaving us with Slovakia and Slovenia. Neither appeared decidedly superior. Given our team's specific expertise and other practical considerations, Czechia emerged as the most fitting candidate for our research.

In summary, our diverse case approach has the potential to offer preliminary evidence that reflects patterns across the entire alliance, especially where findings are consistent between countries (Seawright and Gerring 2008, 300–01). At the same time, a clear downside of our approach is that conflicting or contradictory findings across cases make it difficult to generalize and draw clear conclusions. Indeed, differences in our findings between the two countries suggest that further research is needed to understand these variations more clearly. As such, we view our findings as primarily descriptive.

Treatments: Foreign-Deployed Military

A key challenge in our research design is to exogenously randomize the treatment. The difficulty arises because the treatment, foreign military deployments, is fixed depending on the respondent's country of residence. We therefore split the samples into four groups, a control group in addition to three experimental treatment groups. More specifically, we remind all participants about the factual status quo in their home country pertaining to NATO membership, the level of US troop deployments (GER: 36,000; CZ: 0) and US nuclear weapon deployments (GER: 20; CZ: 0), including a brief explanation that the latter would be delivered only after a joint decision (GER only).⁵

The treatment prompt for Germany states: Now there continue to be political actors in the United States, who advocate to reduce the military presence in Germany. Let us assume that the United States withdraws

• Eighteen thousand US soldiers from Germany.

• All nuclear bombs from Germany.

• Eighteen thousand US soldiers and all nuclear bombs from Germany.

The treatment prompt for the Czech Republic states: Now there continue to be political actors in the United States, who advocate to expand the military presence in Europe. Suppose, then, that the United States were to deploy

• Four thousand and five hundred soldiers in the Czech Republic.

• Twenty nuclear bombs in the Czech Republic, which could be used in the event of an emergency after a joint decision had been taken.

• Four thousand and five hundred soldiers and twenty nuclear bombs in the Czech Republic, which could be used in the event of an emergency after a joint decision had been taken.

Importantly, in both scenarios, our survey does not link military deployment decisions to the threat level. On the contrary, our wording rather implies that certain factions within the US political spectrum might advocate for withdrawal (Germany) or deployment (Czechia), without consideration of the current threat level.

Threat Perceptions

To test whether a foreign deployed military induces assurance or insecurity, we measure the respondent's assessment of the threat. While we ask several questions that broadly pertain to citizens’ threat perceptions in our survey,⁶ we here focus on the most direct items, asking respondents to evaluate the likelihood of (1) their country being attacked and (2) allies being attacked in the medium term.⁷ According to the assurance–free-riding pathway future withdrawals [deployments] should result in heightened threat perceptions [and the other way around in the Czech survey]. By contrast, the insecurity–de-escalation or insecurity–self-insurance pathways predict that withdrawals [deployments] of foreign military deployments decrease perceived levels of threat relative to the baseline condition [and the other way around in the Czech survey].

Attitudes toward National Defense

Finally, to measure the subjective need for national defense, we employed the following questions:⁸

• Of course, we all hope that there will not be another war involving [Country], but if it were to come to that, would you be willing to fight for your country? (Yes/No).⁹

• [Country] should acquire its own nuclear weapons (4 point scale).

• As you know, NATO countries in 2014 pledged to invest 2 percent of their GDP on defense. In 2021 [Country] invested [GER: 1.53, CZ: 1.4] percent. How much do you think [Country] should invest in the future? (continuous scale).

We also use Principle Component Analysis (PCA) to summarize our correlated outcome variables onto a linear coordinate system, thereby reducing the dimensionality of the original data into a single scale.¹⁰

Method

Across all analyses, we use Ordinary Least Squares regressions with robust standard errors. We control for age, region, gender, education, political interest, political orientation, knowledge about nuclear weapons and foreign policy and the day participants took the survey.

Results

The survey experiments allow us to gauge into the black box of causal mechanisms by examining whether foreign-deployed military impacts attitudes toward national defense as an externality of host state assurance or insecurity. Thus, our analysis proceeds in two steps, corresponding to two distinct (sets of) dependent variables: First, we investigate whether foreign military deployments generate assurance among citizens, or whether they induce insecurity. Second, we examine how foreign deployed military and threat perceptions are related to citizens’ subjective need for national defense.

The results largely confirm the suspicion that accusations of free-riding overly simplify the complex nature of the micro-mechanisms of foreign and defense policies, as these deployments might not actually assure allies. However, overall the findings are not entirely conclusive and challenge us to think beyond the theoretical frameworks.

Effects of Foreign Military Deployments on Threat Perceptions

We begin by investigating the relationship between US military deployments/withdrawals and threat perceptions in figure 5.¹¹

The left panel (Czechia) depicts the effects of foreign military deployments on threat perceptions. Here, negative coefficients would be expected if deployments induce assurance (H1; see figure 2). By contrast, the right panel (Germany) depicts the effects of withdrawals, so positive coefficients would be consistent with the assurance–free-riding hypothesis.

Strikingly, in both experiments, we find no evidence that US deployments assure allies. More precisely, if the United States were to deploy its military to Czechia, the results suggest that this would hardly alter threat perceptions among Czech citizens. Conversely, there is some (mild) indication that German respondents regard it as less likely that their country will be attacked or that allies will be attacked if the United States withdraws its military, although only one treatment is statistically significant. Overall, there is no systematic evidence that US military deployments make citizens feel more protected, as would be expected by the “assurance–free-riding hypothesis.” If anything, these results support the expectations of the “insecurity–de-escalation” and “insecurity–self-insurance” hypotheses.

Effects of Foreign Military Deployments on Attitudes toward National Defense

Given that foreign military deployments may not create a sense of assurance among host citizens, and withdrawals could even heighten insecurity, our next step is to ask how this is mirrored by attitudes toward national defense. Figure 6 visualizes the effects of the treatment conditions on the subjective need for national defense.

In the Czech experiment featuring US deployments (left panel) we again hardly detect any effects as most coefficients are close to 0 and not statistically significant. The only exception is the surprising finding that deployments of both nuclear weapons and troops raise the demand for nuclear proliferation. Again, if anything these findings contradict the conventional wisdom that foreign deployed military allow host states to piggyback on US security guarantees and thereby discount national defense (see H1). However, they also do not yield support for the alternative hypotheses, H2 and H3.

The right panel depicts the effects of US withdrawals from Germany. The results broadly indicate that citizens support tougher defense measures under reduced or fully withdrawn US military deployments, as indicated by positive coefficients. While the effect of the different treatment levels is not entirely consistent, all statistically significant results point in this direction. For example, we observe that a withdrawal of troops increases the willingness to fight among Germans. Moreover, US nuclear weapons withdrawal results in an increased preference for nuclear proliferation. Furthermore, we find that German survey respondents support an increase in defense spending after the withdrawal of troops. Consequently, the combined defense scale (PCA) also indicates that both conventional troop and nuclear weapon withdrawals increase the outcome. Overall, the results of the German survey featuring withdrawals of the US military raise the demand for national defense, consistent with the assurance–free-riding hypothesis, but also the insecurity–deescalation hypothesis.

Why do we observe these divergent findings between deployments and withdrawals? One possibility could be that scenarios of withdrawal affect citizens more substantively than scenarios of deployments. Indeed, this would be in line with prospect theory¹² (Kahneman and Tversky 1979) which predicts larger effects for losses compared to gains. Alternatively, it is possible that the German and Czech public differ in how they react to perceptions of threat. We therefore also examine how insecurity in general affects attitudes toward national defense.

Effects of Threat Perceptions on Attitudes toward National Defense

Figure 7 visualizes the independent effects of (non-randomized) threat perceptions — net of deployments/withdrawals — on attitudes toward national defense. We again focus on the two most pertinent threat variables which ask about the likelihood of an attack on the country in question or its allies.¹³ The results are strikingly similar across both countries: perceptions of threat are generally associated with an increased subjective need for national defense, as indicated by a series of positive and statistically significant coefficients. Substantively, the higher the likelihood that the respondent's country or a close ally will be attacked, the higher the demand for higher defense spending, national nuclear proliferation, as well as the combined defense scale (PCA). The only exception is a null finding for respondents’ own willingness to fight for their country. Overall, these findings are consistent with both the assurance–free-riding and insecurity–self-insurance pathways, since insecurity is countered by national defense efforts.¹⁴

Discussion

In summary, the survey experiments highlight the complex nature of the effect of US deployments on citizen's foreign policy attitudes. For the experiment conducted in Germany, on the one hand, we find evidence that hypothetical US military withdrawals increase the subjective need for national defense. On the other hand, we fail to find any evidence that citizens feel protected by such deployments. If anything, citizens seem to feel safer in the absence of US military. In the Czech experiment, it appears that foreign deployed military deployments would neither reduce Czech defense efforts nor would these deployments make the Czech public feel safer. However, for both countries, we find that heightened threat perceptions tend to increase the subjective need for defense.

Overall, our survey experiment is not able to fully disentangle the nexus between foreign-deployed military, assurance/insecurity, and burden sharing. Yet, there is substantial evidence that reevaluating the degree of assurance provided to allies by foreign military deployments is essential. In fact, common free-riding allegations might be misguided, since we fail to find evidence that military deployments reduce threat perceptions, while withdrawals might even foster a sense of security. In general, our evidence presents more substantial findings for Germany than for the Czech Republic across most of the observed effects. Future research should try to dig deeper into differences in the subjective need for national defense among different European states. Particularly, it would be interesting to examine the underlying factors for the divide from different angles: Historical, geographical, cultural.

Conclusion

In line with previous studies, our analysis shows that withdrawing US military deployments can heighten the subjective need for national defense of host state citizens. This applies to both conventional forces and nuclear weapon deployments, which have not been tested before. Conventional wisdom would ascribe such findings to free-riding: Military deployments assure allies by offering protection.

Consequently, their withdrawal diminishes this sense of security. However, based on existing literature, we argued that this is not a given. For example, if citizens carry skeptical views of the guardian, they might experience an increase in threat perception stemming from such deployments. Thus, to further unpack these mechanisms, we conducted two survey experiments, one examining withdrawals (in Germany) and one examining deployments (in Czechia), hoping to get deeper insights into the relationship between US military deployments, threat perceptions and the subjective need for national defense. Our findings partly confirm these expectations: We find little evidence that citizens feel protected by these deployments. In fact, our results in Germany even indicate that the withdrawal of US conventional and nuclear forces lowers citizens’ threat perceptions. By contrast, the Czech experiment suggests that concerns about free-riding as a result of new deployments might be overstated.

The survey experiments unveil important between-country nuances. By and large, treatment effects are less pronounced in the Czech Republic than in Germany, especially with regard to variables measuring respondents’ subjective need for national defense. One possibility is that deployments and withdrawals are distinct phenomena. This constitutes an interesting avenue for future research. Another possibility is that there exist inherent differences between countries that joined NATO after the end of the Cold War and those that joined before. US military deployments might lower the subjective need for national defense in some countries, but not others. Perhaps a natural next step would be to divide NATO member states along these historic dimensions.

Together, the findings suggest a complex relationship between military deployments, threat perception and citizens’ subjective need for national defense. Earlier studies have not taken these nuances into account. However, we were also not able to fully disentangle this conundrum, potentially calling for a new theoretical framework. Theoretically, it is possible that threat perception and subjective need for national defense are more separated than is expected by previous literature, i.e., threat perception is not the main driver of changes in the subjective need for national defense. This would require novel theoretical contributions as the basis for future empirical investigations.

US military deployments have been put in the spotlight once again by former President Trump. Given that Russia's attack on Ukraine has caused an increase in these deployments, they will likely remain a topic of controversy for the foreseeable future. Yet, there is a considerable lack of understanding with regard to the micro-mechanisms underlying the relationship between military deployments and national defense policies. This study indicates that decision-makers should more carefully think about the effects of these deployments and withdrawals. Importantly, citizens may feel different from foreign policy elites. This in turn can cause resentment, and through democratic channels, underinvestment in national defense. Thus, we hope to be the starting point for a new research agenda on the relationship between security guarantees, threat perceptions and citizens’ subjective need for national defense.

Author Biography

Alexander Sorg is a Post-doc at Sciences Po Paris and a Research Fellow at the Hertie School's Centre for International Security.

Julian Wucherpfennig is a Professor of International Affairs and Security at the Hertie School's Centre for International Security.

Notes

Authors’ note: We thank Jonathan Markowitz, David Backovsky, audiences at the Hertie School's Nuclear Security Workshop and Harvard University's Managing the Atoms Colloquium, as well as the editors and two reviewers for helpful comments and suggestions on earlier versions of this research note. Milan Schröder provided excellent research assistance. This research was funded by the Stanton Foundation. The data underlying this article are available on the ISQ Dataverse at https://dataverse-harvard-edu.libproxy.ucl.ac.uk/dataverse/isq.

Footnotes

References