Fighting for the (Step)Motherland?: Predictors of Defense Willingness in Estonia’s Post-Soviet Generation

Christofer Berglund*

Södertörn University and Malmö University

Kairi Kasearu

University of Tartu

Juhan Kivirähk

University of Tartu

What makes individuals willing to defend their (adopted) homeland as their own? This is an essential question for all diverse societies. We turn to the case of Estonia, which inherited a sizable Russian-speaking population after the fall of the Soviet Union. Using recent polling data, we test demographic and attitudinal predictors of defense willingness among the first generation of males that have been raised in the republic since the restoration of independence. The results enable us to unpack differences between Estonian-speakers and Russian-speakers, as well as disagreements among the latter, which shed light on the state of social cohesion in Estonia’s national fabric.

Introduction

An actor’s relative resolve in a confrontation is crucial to its prospects of emerging victorious (Mack 1975). In fact, weak actors possessing greater determination often win wars against more powerful adversaries (Arreguin-Toft 2005). Signaling a willingness to defend core national interests is therefore a vital resource of small states adjacent to regional powers. This realization motivated Aleksander Einseln, Estonia’s first post-Soviet chief of defense in the 1990s, to issue standing order number one, calling for immediate and active resistance in the event of foreign troops ever crossing the border. However, because Estonia mobilizes rank-and-file soldiers through conscription, the strength of this signal hinges on the population’s defense willingness.

A deficit of social cohesion among the population, or in segments thereof, could thus turn into an Achilles’ heel. When the republic reasserted its statehood in 1991, over a third of its residents were Russian-speakers. Most had settled in Estonia during Soviet rule and had never learned Estonian. As the state restricted citizenship to those who were citizens in 1940—that is, prior to the occupation—and to their descendants, these Russian-speakers became aliens in the republic (Taagepera 1992; Kirch, Kirch, and Tuisk 1993). Their integration remains an existential challenge for Estonia, despite progress made over the last three decades (Lauristin and Heidmets 2002; Pettai 2021). Intergenerational naturalization has turned most Russian-speakers raised after the restoration of independence into citizens. Since male citizens are obliged to participate in national defense, it is essential to understand the willingness with which Estonia’s post-Soviet generation commits to this task.

In this article, we ask the following question: What factors predict defense willingness among the first generation of Estonian and Estonian-Russian males raised in Estonia since the dissolution of the Soviet Union? To answer this question, we utilize data from the Public Opinion and National Defense polls administrated between 2016 and 2020. Our sample includes 1,020 male respondents, born between 1987 and 2005, who were between the ages of fifteen and twenty-nine when participating in the poll. The data are representative of the first generation of male residents raised in post-Soviet Estonia. Their readiness to fight for the republic is a litmus test of their integration into their (adopted) homeland. We test whether a series of demographic and attitudinal predictors that are identified in the extant literature can explain variations in their defense willingness.

As expected, Estonians report a greater resolve to defend the state that bears their name as compared to individuals from liminal groups. Noncitizens, Russian-speakers, and residents in Ida-Viru County, a segregated borderland next to Russia, are least motivated to defend the republic. However, these differences are not static. Experience with and perceptions of conscription, trust in core defense institutions, and national pride are more powerful predictors of defense willingness. These findings contribute to earlier research on the integration of Estonian-Russians and elucidate the conditions that prime people to commit to the cause of national defense.

In what follows, we review the general factors believed to make individuals willing to fight for their homeland. A presentation of the Estonian case and our research design ensues. Next, we examine the demographic and attitudinal determinants of defense willingness among the first generation of males raised in post-Soviet Estonia. We then summarize our conclusions and spell out their implications for researchers and for decision makers.

In Search of the Roots of Patriotism

Obliging people to sacrifice their time and, if need be, their lives for the sake of national defense is perhaps the most demanding request a state can make. Thus, among those liable for conscription, it can evoke strong opinions. In France, which pioneered the introduction of the levée en masse in 1793, the most visible reactions consisted of “draft evasion or desertion wherever and however it could be contrived” (Weber 1976, 295). Indeed, as Levi (1997) argues, getting self-interested individuals to accept concrete personal sacrifices for the sake of an abstract collective good has proven to be an enduring problem for states reliant on conscription. The problem takes on additional proportions in multiethnic states, where appeals to nationalism can trigger a backlash among certain constituents (Peled 1994).

What, then, makes people willing to participate in national defense? The existing literature alerts us to two different kinds of predictors. First, there are demographic predictors that tell us something about the background of those (un)willing to fight for their homeland, such as their language skills, settlement patterns, educational attainment, and citizenship status. The second set of predictors are attitudinal in nature and tell us something about the mind-set of those (un)willing to fight for their homeland, such as their stance toward conscription, trust in government institutions, beliefs about the risks of free-riding, national pride, trust in defense institutions, and threat perceptions. Taken together, these factors ought to be able to account for a good chunk of the variation in a population’s defense willingness.

Let us start with the demographic factors. Our first hypothesis (H1) is: those more fluent in the state language will report a higher defense willingness. This prediction is derived from Marquardt (2018), who argues that individuals who do not speak the state language face linguistic hurdles which make them more prone to protest, since the hurdles are often perceived as discrimination. Our second hypothesis (H2) is: those residing outside segregated borderlands will report a higher defense willingness. This prediction is derived from Toft (2002), who argues that minorities residing as concentrated majorities adjacent to their kin state are prone to challenge the state in which they reside. Our third hypothesis (H3) is: those less educated will report a higher defense willingness. This prediction harks back to Inglehart, Puranen, and Welzel’s (2015) claim that increasing life opportunities make people less willing to fight. Our fourth hypothesis (H4) is: those who are citizens of the republic will report a higher defense willingness. We trace this prediction to Krebs, who argues that minorities holding citizenship often use active service as a venue for winning recognition, thus “fighting for rights” (2006, title page).

Turning to the attitudinal factors, our fifth hypothesis (H5) is: those socialized into the citizen-soldier ideal will report a higher defense willingness. This prediction stems from the notion that conscription sensitizes recruits to the importance of national defense. Yet, as Krebs (2004) cautions, the correlation could also arise from self-selection, since those who are less motivated to undergo the draft may seek an exemption. Our sixth hypothesis (H6) is: those who are more trusting of government institutions will report a higher defense willingness. Levi (1997) put this idea forward when she argued that citizens who perceive the government as fair and who believe that others are not free-riding at their expense are more likely to defend the state. To test the latter prediction, our seventh hypothesis (H7) is: those who are more confident that others will pull their weight will report a higher defense willingness. Our eighth hypothesis (H8) is: those who feel more pride in their nation will report a higher defense willingness. This prediction stems from Torgler’s (2003) claim that pride generates attachment to the in-group. Torgler also theorizes that trust in the defense establishment reduces the cost of participation. Hence, our ninth hypothesis (H9): those who are more trusting of defense institutions will report a higher defense willingness. Our tenth and final hypothesis (H10) is: those who are more fearful of neighboring powers will report a higher defense willingness. This prediction comes from Williams’s (2005) observation that territorial defense appears less important if the principal threats are seen as distant or irregular.

These ten hypotheses inject theoretical structure into our investigation. We distinguish between demographic and attitudinal factors, because it is important to understand whether Estonia’s protectors come from certain societal strata or whether it is their mind-sets that prime them to commit to the republic’s defense. Before we present the data and methods used to test the above propositions, we turn to the challenges faced by Estonia and its Russian-speakers after the status reversal in 1991.

Estonia Reborn in a Different Form

In 1988, the Estonian Soviet Socialist Republic (SSR) declared itself sovereign, thus setting a precedent for other republics wishing to escape the Soviet Union (Taagepera 1993; Raun 1994). It proceeded to make Estonian the official language and restored the blue-black-white national tricolor flag. Eighty-three percent of residents participated in the March 1991 independence referendum, in which 78 percent voted in favor. Then, as the Soviet Union descended into the throes of a coup d’état later that summer, the Estonian Supreme Council seized the momentum to proclaim restored statehood (Riigi Teataja 1991). Thus, Estonia does not see itself as a successor to the Estonian SSR (1940–90) but as a continuation of the Republic of Estonia (1918–40). It therefore limited birthright citizenship to those who were citizens, or descendents of citizens, prior to Soviet occupation. Estonia also made an indirect claim to the territories of the interwar republic, as it existed prior to occupation (Pettai 2007).

Estonian officials regarded this doctrine of legal restorationism as crucial to salvaging their imperiled nation-state after five decades of Soviet repression and Russification (Kangilaski et al. 2005). Prior to annexation, Estonia had been a homogeneous republic in which Estonians comprised 88 percent of the population. However, Russians migrated to Estonia en masse after the 1940s and seldom gave a thought to learning Estonian. As a result, Estonia became bi-national. In 1989 its population included 62 percent Estonians and 30 percent Russians; most of the remaining 8 percent were Slavic Russian-speakers. In an attempt to return to the status quo ante and “guarantee the preservation of the Estonian people, the Estonian language and the Estonian culture through the ages,” the resurrected republic excluded Soviet-era settlers from automatic citizenship (Riigi Teataja 1992). Estonia’s half million Russian-speakers had to choose between naturalization, taking up Russian citizenship, or retaining their liminal status through the use of an alien’s passport (UNHCR 2016).

This turn of events produced a fair amount of discontent among the republic’s non-Estonian population. Utilizing polling data from the spring of 1991, Taagepera (1992, 126) estimates that half of non-Estonians resisted the impending status reversal, one-fourth supported it, and the rest abstained from taking a stand. The loudest protests came from the pro-Soviet Intermovement, an organization advocating for the preservation of the USSR, and from officials in Russian-populated towns in the northeast, who organized a plebiscite on self-determination in 1993. While their call did not gain traction, polarization between Estonians and Russians ran deep, and the latter were detached, or at best divided, in their attitudes toward the Estonian state (Kirch, Kirch, and Tuisk 1993; Evans 1998). In connection to the withdrawal of Russian troops in the first half of the 1990s, close to 100,000 Soviet-era settlers left Estonia for Russia. Others sought and obtained residence permits, thus voting with their feet to remain. In the second half of the 1990s, as officials put together more concerted integration policies,1 the prospect of intergenerational assimilation appeared on the horizon (Laitin 1998).

However, Russian-speakers faced prolonged political exclusion, since those holding Russian or alien’s passports continue to be ineligible to vote or to stand in national elections (Pettai and Hallik 2002; EPR Atlas 2021). Full political rights are reserved for citizens—that is, those capable of passing Estonia’s naturalization criteria. As the naturalization criteria have been relaxed over time, the share of citizens has risen to 85 percent of the population, with the remaining population possessing either Russian or alien’s passports in equal proportion (Järve and Poleshchuk 2019; Pettai 2021). Also, the share of Russian-speakers who consider themselves as competent Estonian-speakers has increased, from 14 to 42 percent, since the restoration of independence (EIM 2020). While knowledge of the state language helps Russians advance within the Estonian labor market, perceptions of an ethnic wage gap endure (Lindemann and Kogan 2013). Furthermore, on each side of the language barrier, media landscapes reinforce different memories and attitudes, as riots over the Bronze Soldier statue illustrated in 2007 (Korts 2009; Kiviräkh 2014; NCDSA 2015, 2019, 2020).2

In addition to these internal difficulties, the state also confronts external problems. Russia controls territories that formed part of the interwar Republic of Estonia but were transferred from the Estonian SSR to the Russian Soviet Federative Socialist Republic (SFSR) in 1945. Estonia has in effect given up on regaining these lands,3 Jaanilinn in the northeast and Petseri in the southeast, which are referenced in article 122 of the constitution (Aalto 2000).4 Russia claims these districts on the basis that Estonia is a successor to the Estonian SSR, thus rejecting Estonia’s stance that processes transpiring under Soviet occupation are illegitimate. Given the absence of a common historical understanding, the parties have failed to demarcate their inter-state border (Mälksoo 2005; Stoicescu 2020). Russia is thus capable of pressuring Estonia with reference both to its borders and to the need to protect “stranded compatriots,” as it has earlier done in Georgia since 2008 and in Ukraine since 2014 (CNA 2015; Kallas 2016). Recent rhetoric from Kremlin ideologues holds that neither the Eastern Slavic nations nor the Baltic states should be allowed to exist outside of Russia’s dominion (Putin 2021; Surkov 2022).

Given this situation, we maintain that it is important to investigate which residents raised in Estonia since the restoration of independence are willing to defend it. “Citizens in arms” are the first line of defense tasked with resisting enemies long enough for NATO reinforcements to arrive.5 This is a formidable challenge considering the posture of Russia’s armed forces, Estonia’s lack of strategic depth, and the risk of slow decision making within the NATO alliance (RAND 2016). Estonia’s Defense Forces were created anew in the fall of 1991 and organized along the lines of a reserve force. Rather than opting for a small, all-volunteer force useful for out-of-area operations, as most European countries had at the time, Estonia prioritized raising manpower for territorial defense at home. Its active personnel consist of some seven thousand service members, half of whom are conscripts. Another twenty thousand are part of Estonia’s Defense League (IISS 2021, 98). In total, 230,000 people—18 percent of the population—can be mobilized in the event of war. As such, societal support is essential for Estonian defense planners (Kepe and Osburg 2017; Riigikantselei 2017).

Data and Methods

Our investigation builds on data from the Public Opinion and National Defense polls. Since 2000, Estonia’s Ministry of Defense has contracted Turu-uuringute AS, a firm specializing in market research, to monitor attitudes toward a range of defense issues. We utilize their polls from 2016 through 2020, which are representative of Estonia’s permanent residents aged fifteen or older. However, we omit females and those aged thirty or above from our sample. Thus, it consists of 1,020 male respondents aged fifteen to twenty-nine and born between 1987 and 2005. Among them, 67 percent were Estonian-speakers (n = 682) and 33 percent were Russian-speakers (n = 338). In the latter group, two-thirds were Estonian citizens (n = 228) and one-third were not (n = 110). Respondents were selected using a stratified random sampling design with proportional allocation and interviewed in person at their home address.6 These rigorous collection procedures ensure that the dataset can be used to make generalizations about the first generation of men raised in post-Soviet Estonia.

No other studies have investigated the predictors of defense willingness in this critical subset of the population. Yet, Estonian officials see cohesion as a precursor to deterrence—a concept that is once again at the forefront of NATO’s policies toward its eastern flank. Interest in the topic is therefore mounting among scholars. Rutkauskas (2018) utilizes the World Values Survey and the European Values Survey to probe the roots of defense willingness among the Baltic populations. However, these datasets do not allow him to test the demographic predictors that are thought to be important (Andzans and Spruds 2020; Berzina and Zupa 2021). Estonia’s Public Opinion and National Defense polls contain more information about respondents and are better suited for the task at hand.7 However, their potential has not been tapped. Reports from Turu-uuringute AS (2019, 2020) describe the attitudes of the general population, as well as differences between Estonian-speakers and Russian-speakers, but refrain from examining the possible causes behind their varied levels of defense willingness.

We take this step. Our dependent variable is operationalized using the question: “If Estonia is attacked, are you ready to participate in defense efforts to the best of your abilities and skills?” Answers were rated on a scale ranging from “definitely no” and “probably not” to “probably yes” and “definitely yes.” Respondents could also choose “hard to say.” For the purposes of the ensuing logistic regression, these were turned into a dummy variable: “will not defend” and “will defend.” We included fence-sitters in the former, since social pressures might induce those disinclined to fight to avoid taking a stand.8 As expected, those answering in Estonian report higher defense will than those answering in Russian. Seventy-seven percent of Estonian-speakers reported being prepared to defend the state compared to 58 percent of Russian-speakers—a difference of 19 percent (χ2 = 39.8, p ≤ .001). Compared to other NATO member states, both percentages are quite high, which is notable considering that we tapped into respondents’ personal readiness to partake in active defense efforts.9

But what explains individual variations in defense willingness? In order to solve this puzzle, we needed to operationalize the demographic and attitudinal predictors presented earlier. We measured the first independent variable using the question: “At what level do you speak Estonian?” The question was posed to respondents who answered the questionnaire in Russian. Their replies range from “I understand well and speak fluently” to “I understand and speak a little” and “I understand a little, but do not speak” to “I do not understand and do not speak.” We coded the first answer as “good Estonian skills” and the last three answers as “poor Estonian skills.” We then compared each to a reference group that consisted of respondents who answered in Estonian. This enabled us to test whether those more capable in Estonian also reported a higher defense willingness (H1).

The second independent variable centers on the role of geographic location. To capture it, we filtered for respondents who live in Ida-Viru, a region that borders Russia. Out of Estonia’s fifteen counties, Ida-Viru is the only county where ethnic Russians dominate demographically, representing three-quarters of residents. This reference group was then compared to those settled in the capital, Tallinn, where ethnic Russians constitute one-third of the population, and to those living in the rest of Estonia, where ethnic Estonians dominate. This comparison allows us to test whether individuals residing outside the segregated borderland report a higher defense willingness (H2).

The third independent variable is operationalized using the question: “What kind of education do you have?” Respondents’ answers include basic education (grades 1–9), vocational training, high school (grades 10–12), and higher education. Using this ordinal scale, we test if those with less education report a higher defense willingness (H3).

The fourth independent variable is citizenship. It is operationalized using a question that asks whether one’s passport is Estonian, Russian, something else, or undefined (stateless). We bundled the latter three categories together under the rubric of noncitizens and then compared them to Estonian citizens to test whether the latter report a higher defense willingness (H4).

The fifth independent variable is measured using two questions. The first asks whether respondents have experienced being drafted into Estonia’s Defense Forces. The second asks if respondents consider it important for young men to complete conscription. We combined these to create four categories: Individuals responding no to the first item and “unimportant” to the second were coded as dissenters. Those answering the opposite were labeled as citizen-soldiers. Individuals with experience of conscription but who assessed it as “unimportant” were branded as consenters. Those who lacked experience of the draft but rated it as “important” were categorized as anticipators. We then compared these four groups to one another to explore if the experience of conscription—or its perceived importance—influences defense willingness (H5).

The sixth independent variable turns to the issue of trust in government institutions. It was appraised through the question: “To what extent do you trust the following … institutions?” The parliament, cabinet, president, and prime minister were chosen as institutions representing the government. Answers ranged from “I do not trust at all” and “rather not trust” to “rather trust” and “I have full trust.” Those choosing “hard to say” were placed in the middle of this scale. Respondents’ ratings of the four institutions were averaged to form a composite index that exhibited strong internal coherence (Cronbach’s alpha = .85). This allowed us to test whether individuals who place greater trust in government institutions also exhibit a higher defense willingness (H6).

The seventh independent variable was measured using the question: “In the event of a foreign armed attack, would it in your opinion be possible to defend Estonia until Allied help arrives?” Responses ranged from “definitely no” and “probably not” to “probably yes” and “definitely yes.” We placed those answering “hard to say” in the middle. This scale enabled us to test whether individuals who are confident that others will stand up to aggression are more willing to fight themselves (H7).

The eighth independent variable taps into the importance of national pride. It was captured by the question: “How often have you been proud or happy to live in the Republic of Estonia?” Answers spanned from “never” and “sometimes” to “quite often” and “all of the time.” In this case, individuals choosing “hard to say” were excluded, since this response is difficult to rank on an ordinal scale. This item helped us to test whether respondents feeling greater pride in their nation report a higher defense willingness (H8).

The ninth independent variable focuses on trust in defense institutions. It was calculated using the question: “To what extent do you trust the following … institutions?” Estonia’s Defense Forces, Estonia’s Defense League, and NATO were chosen as entities representing the defense establishment. Respondents rated each of these organizations on a scale spanning from “I do not trust at all” and “rather not trust” to “rather trust” and “I have full trust.” Those answering “hard to say” were, again, placed the middle. We then created a composite index that showed good internal coherence (Cronbach’s alpha = .78). This allowed us to examine if respondents who are more trusting of the defense sector also report a higher willingness to protect the republic (H9).

The tenth independent variable revolved around threat perceptions. We measured them using the item: “To what extent do you consider [Russia’s attempts to regain influence in neighboring countries] a threat to peace and security in the world?” Answers ranged from “definitely no” and “to some extent” to “definitely yes.” Those responding “hard to say” were folded into the unresolved middle stratum. This operationalization enabled us to investigate whether individuals who are more fearful of neighboring powers do, in fact, report a higher defense willingness (H10).

Table 1 summarizes these independent variables and the mechanisms thought to explain their impact on defense willingness. Which of them help us recognize Estonia’s prospective defenders?

We first explored bivariate associations between each of the independent variables and our dependent variable among Estonian-speakers and Russian-speakers. Significant attitudinal differences emerged across the subgroups and between those (un)willing to defend (see Table A1). However, several correlations faded after we proceeded to the logistic regressions, which hold constant the influence of other predictors included in the models. The next sections unpack these findings.

Demographic Predictors of Defense Willingness

We first performed a logistic regression to profile the individuals (un)willing to fight for the republic (see Model 1 in Table 2). Language, location, and citizenship emerged as significant; education did not. Taken together, these demographic predictors explain a meager 7.4 percent of the variance in defense willingness in the sample. Let us describe these results in greater detail.

The first predictor is language. Marquardt (2018) claims that a lack of knowledge of the state language marks those who do not speak that language as outsiders, since it restricts their life opportunities. He claims that protests against the state arise from this linguistic barrier and not from ethnic differences per se. Our findings lend superficial support for this interpretation. Seventy-seven percent of Estonian-speakers are prepared to defend the republic compared to 55 percent of Russian-speakers not capable in Estonian and 63 percent of Russian-speakers capable in Estonian. The intergroup difference is significant, but the intragroup difference is not. That is, when we compare Estonian-speakers to Russian-speakers, the likelihood of the latter defending the republic is 0.56–0.60 times lower. However, Russian-speakers capable in Estonian are no more eager to fight than Russian-speakers incapable in Estonian. Since their (un)willingness to fight for the (step)motherland is unrelated to their aptitude in the state language, we infer that linguistic integration does not generate defense willingness among Russian-speakers.

The second predictor is location. Toft (2002) argues that residents of segregated borderlands are prone to challenge the state. Our data support her thesis. Fifty-three percent of respondents in Ida-Viru are prepared to defend the republic compared to 71 percent in Tallinn and 75 percent in the rest of Estonia. The likelihood of the two latter groups fighting for the republic is 1.64 and 1.73 times higher, respectively. An intergroup breakdown reveals that these patterns are the product of regional differences among Russian-speakers: 48 percent in Ida-Viru, 59 percent in Tallinn, and 70 percent in other parts of Estonia reported readiness to defend the (step)motherland. Regional differences among Estonian-speakers are negligible.

The third predictor is education. Inglehart, Puranen, and Welzel (2015) posit that education increases life opportunities, which leads people to focus on self-realization and makes them less willing to sacrifice their lives in war. However, our data do not support this contention. The likelihood of reporting willingness to defend the republic increases 1.07 times for each step up in education level (ordinal scale 1–4). The effect runs counter to the prediction but is too slight to be significant.

The fourth predictor is citizenship. As Krebs (2006) illustrates, once a legal bond between state and minorities has been established, the latter can “fight for their rights” through service in uniform, thus forcing officials using civic rhetoric into accepting them as equals in practice. This path is closed to noncitizens. Our initial screening offers some support for this line of reasoning. Seventy-four percent of citizens report readiness to defend the republic compared to 49 percent of noncitizens. In Model 1, the former are 2.01 times more probable to fight for the Estonian state in case of external aggression.

This analysis enables us to profile potential deserters and defenders in the event of an attack against the republic. Russian-speaking noncitizens in Ida-Viru are overrepresented in the former group and Estonian-speaking citizens in the latter group. Yet, considering that the Nagelkerke R2 for Model 1 is .074, these caricatures are poor predictors of defense willingness. Nor does the logistic regression shed light on the mechanisms that produce these demographic findings. To overcome these shortcomings, we expand our investigation in the following section.

Attitudinal Predictors of Defense Willingness

We conducted another logistic regression to examine the mind-sets of individuals who are (un)willing to fight (see Model 2 in Table 2). While conscription, national pride, and defense trustworthiness came out as significant, government trustworthiness, ethical reciprocation, and threat perceptions did not. After adding all these attitudinal predictors to our expanded model, 25.6 percent of the variance in defense willingness was explained.

The fifth predictor is conscription. Militaries are often considered a “school for the nation,” capable of turning reluctant recruits into dedicated patriots, but as Krebs (2004) cautions, it is difficult to peer into the black box of socialization in the armed forces. Our data enabled us to test the effect of conscription experience while controlling for attitudes toward it, and vice versa. We found that both aspects matter. Among individuals rating the draft as “important,” former conscripts are 2.08 times more probable to report defense willingness. At the same time, among individuals lacking experience of conscription, those rating the draft as “unimportant” are 0.36 times less probable to report readiness to fight. Moreover, former conscripts who rate the draft as “unimportant” are 0.43 times less probable to defend compared to those rating the draft as “important” despite lacking experience thereof. When searching for intergroup differences it becomes clear that the perceived importance of the draft has more significant effects than the experience of being drafted (see Table 3). Yet, we still cannot determine whether the positive association between conscription and defense willingness is a consequence of socialization during the draft or of self-selection prior to it.

The sixth and seventh predictors—government trustworthiness and ethical reciprocation—are both foundational in Levi’s (1997) idea of “contingent consent.” Levi argues that people will fight for their homeland if they trust their government and if they perceive that other citizens are doing their fair share. Our data offer scant support for her claims. The likelihood of defense willingness declines 0.89 times for each incremental increase in trust in government (ordinal scale 1–5), but the change is non-significant. When probing for intergroup differences, we find that Estonian-speakers and Russian-speakers report similar levels of trust in government.10 It has no association to defense willingness in the former subgroup and—despite a significant bivariate association in the latter subgroup (see Table A1)—the predictor (operationalized as a composite index, consisting of trust in parliament, cabinet, president, and prime minister) fails to forecast readiness to fight in the logistic regression.

Ethical reciprocation also has slight effect. The likelihood of reporting readiness to fight for the republic increases 1.13 times for each incremental rise in the belief that Estonia can be defended until Allied help arrives (ordinal scale 1–5). However, this effect is non-significant. Thus, even though there is a bivariate association between ethical reciprocation and defense willingness in both language subgroups (see Table A1), the former remains an imprecise predictor of the latter.

The eighth and ninth predictors—national pride and trust in the defense establishment—fare better. Torgler (2003) posits that national pride creates attachment to the in-group, thus making individuals more prone to defend it. He also asserts that trust in the defense sector reduces the costs of participation, thus making individuals more willing to fight. Our data affirm the relevance of his claims. The likelihood of defense willingness increases 1.31 times for each incremental increase in national pride (ordinal scale 1–4). Fifty-two percent of Estonian-speakers and 33 percent of Russian-speakers report feeling frequent or constant pride in living in the republic, and there is an association to defense willingness in both subgroups (see Table A1).

Trust in the defense establishment (operationalized as a composite index, consisting of trust in Estonia’s Defense Forces, Estonia’s Defense League, and NATO) also has a strong positive effect. The likelihood of defense willingness increases 1.59 times for each unit added to the predictor (ordinal scale 1–5). Estonian-speakers report higher trust in the defense sector than Russian-speakers,11 but the association to the will to fight is significant in both subgroups (see Table A1). These findings add weight to Torgler’s propositions, although the mechanisms explaining these patterns still remain open to interpretation.

The tenth predictor is threat perceptions. Williams (2005) leads us to expect that individuals who see Russia’s attempts to regain influence in the “near abroad” as a threat will report a greater resolve to fight. Our data offer partial support for this idea. The likelihood of defense willingness increases 1.25 times for each incremental rise in the predictor (ordinal scale 1–3), but the change is non-significant. When exploring intergroup differences, we see a rift between Estonian-speakers and Russian-speakers. Forty-four percent of the former and 15 percent of the latter rate Russia’s policies toward neighboring countries as a definite threat.12 These attitudes are related to defense willingness among Russian-speakers but not among Estonian-speakers (see Table A1).

After factoring in these attitudinal predictors, Nagelkerke R2 for Model 2 rises to .256, and the demographic predictors, studied as part of Model 1, lose significance. Language, citizenship, and location matter, but as proxies. Readiness to fight for the (step)motherland appears to be a product of individual mind-sets rather than of demographic characteristics. Among males raised in post-Soviet Estonia, our strongest predictors of defense willingness are trust in the defense sector, national pride, and an admixture of experiences from and the perceived importance of conscription.

Conclusions and Implications

What factors predict defense willingness among males raised in post-Soviet Estonia? Getting self-interested individuals to make personal sacrifices for the sake of a collective good is a challenge faced by all militaries. However, in Estonia it is a challenge of exceptional proportions. This once-homogeneous nation-state emerged from the Soviet Union as a bi-national republic, with Russian-speakers comprising 38 percent of its population. Soviet-era settlers did not gain automatic citizenship, and hence faced restricted access to state power, but many have, over time, become naturalized citizens. Individuals from this liminal group are now being conscripted and are expected to defend their (step)motherland. As a small state in need of deterring its great-power neighbor, Estonia cannot afford a glaring deficit of social cohesion.

Prior research on the integration of Estonia’s Russian-speakers has seldom focused on their defense willingness. While this topic has emerged in reports that monitor public attitudes, these reports fall short of investigating the potential causes behind variations in defense willingness. We fill this gap. Our article marshalled data from Estonia’s Public Opinion and National Defense polls to test demographic and attitudinal predictors thought to make people willing to fight for their homeland. Using logistic regression, we estimated the constant effect of each predictor on the likelihood of defense willingness in a representative sample of the first generation of males raised in post-Soviet Estonia. Their commitment to the state is a test of Estonia’s nation-building process.

Our first model studied the background of those (un)willing to fight. It generated three significant effects. The first concerns language. Compared to Estonian-speakers, Russian-speakers are less probable to fight for the republic. But this is not a result of frustrations arising from their lack of state language skills. Russian-speakers capable in Estonian are no more eager to fight than Russian-speakers incapable in Estonian. For reasons to be determined, Russian native speakers report lower defense willingness and do so irrespective of their degree of linguistic integration. The second significant effect concerns location. Compared to residents from Ida-Viru, a segregated Russian-populated borderland, those from Tallinn and the rest of Estonia are more prone to defend the republic. The third significant effect concerns citizenship. Compared to noncitizens, those holding an Estonian passport are more likely to fight for the state. These demographic predictors confirm prevailing caricatures of deserters and defenders but, in fact, explain a meager 7.4 percent of the variance in defense willingness.

Our second model factored in the mind-set of those (un)willing to fight and explained a larger share of variance: 25.6 percent. It generated three significant effects. The first concerns conscription. Both the experience of conscription and its perceived importance have a positive effect on defense willingness. Among those rating the draft as “important,” former conscripts are more likely to fight. At the same time, among those not exposed to the draft, individuals rating conscription as “unimportant” are less likely to fight. The second significant effect concerns national pride. The likelihood of defense willingness increases as pride in living in Estonia rises. The third significant effect concerns defense trustworthiness. The likelihood of fighting for the republic increases as trust in defense institutions rises. The addition of attitudinal predictors in the second model causes our demographic predictors to lose significance, suggesting that the latter are proxies for other reasons of defense willingness.

These empirical findings are the main result of this paper; however, this paper also has important theoretical ramifications. Four of ten predictors failed to generate significant effects. The first forecast to fail concerns education. Inglehart, Puranen, and Welzel (2015) argue that education expands life opportunities, inducing people to focus more on self-realization and making them less willing to fight. We find no evidence of this. The second and third expectations to misfire are derived from Levi’s (1997) idea of “contingent consent.” Levi posits that rational individuals will fight for the state if they trust its government and if other citizens are seen as doing their fair share. Yet, neither government trustworthiness nor ethical reciprocation helps us predict defense willingness. The fourth and final conjecture to flounder concerns threat perceptions (Williams 2005). Fear of Russia’s policies toward the “near abroad” falls short of predicting readiness to fight for the republic. These mistaken prognostications suggest that the aforementioned theories can be rejected or need to have their scope conditions revised.

We see several fruitful avenues for future research. One is to engage in middle-range theorizing and explore if different theories explain defense willingness in different population segments. Education might turn out to have an effect on readiness to fight in other age cohorts. It might also be that transactional calculations about trust in government are tangential to those in the dominant group but essential to earn the “contingent consent” of minorities. Investigating these possibilities will require researchers to explore the scope conditions under which predictors operate. Another promising path forward is to focus on the mechanisms that explain the correlates of defense willingness. What is it about conscription, national pride, and trust in defense institutions that make them good predictors of readiness to fight for one’s state? We need more precise longitudinal data to determine if the theorized mechanisms are behind the observed effects.

Despite the uncertainties surrounding our findings, this article holds valuable lessons for decision makers. Individual attitudes supersede demographic characteristics as predictors of defense willingness in the first generation of males raised in post-Soviet Estonia. “Ethnic profiling,” as some far-right politicians advocate, is hence an ineffectual tool for screening soldiers (ERR 2018). Instead, readiness to fight can be engineered through conscription and policies that nurture national pride and trust in defense institutions. Making manpower decisions on the basis of merit, rather than origin, is critical to this end (Peled 1998; Lyall 2020). Officials are also well advised to monitor information channels that reinforce attitudes detrimental to defense willingness.

References

Aalto, Pami. 2000. “Beyond Restoration: The Construction of Post-Soviet Geopolitics in Estonia.” Cooperation and Conflict 35 (1): 65–88.

Andzans, Maris, and Andris Spruds. 2020. “Willingness to Defend One’s Own Country and to Resist in the Baltic States.” Security & Defense Quarterly 30 (3): 15–30.

Arreguin-Toft, Ivan. 2005. How the Weak Win Wars: A Theory of Asymmetric Conflict. Cambridge: Cambridge University Press.

Berzina, Ieva, and Uldis Zupa. 2021. “Factors Affecting Willingness to Fight for a Country in the Latvian and Russian-Speaking Communities in Latvia.” National Identities 23 (3): 239–52.

CNA. 2015. Mobilizing Compatriots: Russia’s Strategy, Tactics, and Influence in the Former Soviet Union. Arlington, VA: CNA Corporation.

Cronberg, Tarja. 2006. “The Will to Defend: A Nordic Divide over Security and Defense Policy.” In The Nordic Countries and the European Security and Defense Policy, edited by Alyson Bailes, Gunilla Herolf, and Bengt Sundelius, 315–22. Oxford: Oxford University Press.

EIM. 2020. “Estonian Integration Monitoring: 20 Years of Estonian Integration Studies.” https://kul.ee/media/3152/download.

EPR Atlas. 2021. “Ethnicity in Estonia.” Ethnic Power Relations GrowUp Dataset Atlas. https://growup.ethz.ch/atlas/Estonia.

ERR. 2018. “Blue Awakening Wants Ethnic Profiling of Those with Access to State Secrets.” Eesti Rahvusringhääling, September 6.

Evans, Geoffrey. 1998. “Ethnic Schism and the Consolidation of Post-Communist Democracies: The Case of Estonia.” Communist and Post-Communist Studies 31 (1): 57–78.

IISS [International Institute for Strategic Studies]. 2021. “Chapter Four: Europe.” The Military Balance 121 (1): 66–163.

Inglehart, Ronald, Bi Puranen, and Christian Welzel. 2015. “Declining Willingness to Fight for One’s Country: The Individual-Level Basis of the Long Peace.” Journal of Peace Research 52 (4): 418–34.

Järve, Priit, and Vadim Poleshchuk. 2019. “Report on Citizenship Law: Estonia.” Florence: EUI GLOBALCIT.

Kallas, Kristina. 2016. “Claiming the Diaspora: Russia’s Compatriot Policy and Its Reception by Estonian-Russian Population.” Journal on Ethnopolitics and Minority Issues in Europe 15 (3): 1–25.

Kangilaski, Jaak, Virve Kask, Kalev Kukk, Jaan Laas, Heino Noor, Aigi Rahi-Tamm, Rein Raatas, Anto Raukas, Enn Sarv, and Peep Varju. 2005. The White Book: Losses Inflicted on the Estonian Nation by Occupation Regimes, 1940–1991. Tallinn: Estonian State Commission on Examination of the Policies of Repression.

Kepe, Marta, and Jan Osburg. 2017. “Total Defense: How the Baltic States Are Integrating Citizenry into Their National Security Strategies.” Small Wars Journal, September 24.

Kirch, Aksel, Marika Kirch, and Tarmo Tuisk. 1993. “Russians in the Baltic States: To Be or Not to Be?” Journal of Baltic Studies 24 (2): 173–88.

Kivirähk, Juhan. 2014. Integrating Estonia’s Russian-Speaking Population: Findings of National Defense Opinion Surveys. Tallinn: ICDS.

Korts, Külliki. 2009. “Inter-ethnic Attitudes and Contacts between Ethnic Groups in Estonia.” Journal of Baltic Studies 40 (1): 121–37.

Krebs, Ronald. 2004. “A School for the Nation? How Military Service Does Not Build Nations, and How It Might.” International Security 28 (4): 85–124.

———. 2006. Fighting for Rights: Military Service and the Politics of Citizenship. Ithaca: Cornell University Press.

Laitin, David. 1998. Identity in Formation: The Russian-Speaking Populations in the Near Abroad. Ithaca: Cornell University Press.

Lauristin, Marju, and Mati Heidmets. 2002. The Challenge of the Russian Minority: Emerging Multicultural Democracy in Estonia. Tartu: Tartu University Press.

Levi, Margaret. 1997. Consent, Dissent, and Patriotism. Cambridge: Cambridge University Press.

Lindemann, Kristina, and Irena Kogan. 2013. “The Role of Language Resources in Labor Market Entry: Comparing Estonia and Ukraine.” Journal of Ethnic and Migration Studies 39 (1): 105–23.

Lyall, Jason. 2020. Divided Armies: Inequality and Battlefield Performance in Modern War. Princeton: Princeton University Press.

Mack, Andrew. 1975. “Why Big Nations Lose Small Wars: The Politics of Asymmetric Conflict.” World Politics 27 (2): 175–200.

Mälksoo, Lauri. 2005. “Which Continuity: The Tartu Peace Treaty of 2 February 1920, the Estonian–Russian Border Treaties of 18 May 2005, and the Legal Debate about Estonia’s Status in International Law.” Archiv des Völkerrechts 43 (4): 513–24.

Marquardt, Kyle. 2018. “Identity, Social Mobility, and Ethnic Mobilization: Language and the Disintegration of the Soviet Union.” Comparative Political Studies 51 (7): 831–67.

NCDSA. 2015. “Исследование Отношения К Государственной Обороне Среди Русской Молодежи Ида-виру И Таллинна” [Research on attitudes toward state defense among Russian youth in Ida-Viru and Tallinn]. Tallinn: National Centre of Defense & Security Awareness.

———. 2019. “Primary World-View Characteristics of Russian-Speaking Young Adults in Estonia.” Tallinn: National Centre of Defense & Security Awareness.

———. 2020. “Profile of Russian-Speaking Young People in Estonia, 2012–2019.” Tallinn: National Centre of Defense & Security Awareness.

Peled, Alon. 1994. “Force, Ideology, and Contract: The History of Ethnic Conscription.” Ethnic and Racial Studies 17 (1): 61–78.

———. 1998. A Question of Loyalty: Military Manpower Policy in Multiethnic States. Ithaca: Cornell University Press.

Pettai, Vello. 2007. “The Construction of State Identity and Its Legacies: Legal Restorationism in Estonia.” Ab Imperio 3: 403–26.

———. 2021. “National Identity and Re-identity in Post-Soviet Estonia.” Journal of Baltic Studies 52 (3): 425–36.

Pettai, Vello, and Klara Hallik. 2002. “Understanding Processes of Ethnic Control: Segmentation, Dependency and Co-optation in Post-Communist Estonia.” Nations and Nationalism 8 (4): 505–29.

Putin, Vladimir. 2021. “On the Historical Unity of Russians and Ukrainians.” Kremlin, July 12.

RAND. 2016. “Reinforcing Deterrence on NATO’s Eastern Flank.” https://www.rand.org/pubs/research_reports/RR1253.html.

Raun, Toivo. 1994. “Post-Soviet Estonia, 1991–1993.” Journal of Baltic Studies 25 (1): 73–80.

Riigikantselei. 2017. “National Security Concept.” Parliament of Estonia. https://riigikantselei.ee/en/media/376/download.

Riigi Teataja. 1991. “Eesti riiklikust iseseisvusest” [Estonia’s national independence]. https://www.riigiteataja.ee/akt/13071519.

———. 1992. “Eesti Vabariigi põhiseadus” [The Constitution of the Republic of Estonia]. https://www.riigiteataja.ee/akt/115052015002.

Rutkauskas, Virgilijus. 2018. “Factors Affecting Willingness to Fight for One’s Own Country: The Case of Baltic States.” Special Operations Journal 4 (1): 48–62.

Stoicescu, Kalev. 2020. “The Tartu Peace Treaty and Estonia’s Eastern Border.” ICDS Diplomaatia, February 20.

Surkov, Vladislav. 2022. “Туманное будущее похабного мира” [Nebulous future of the disgraceful peace]. Актуальные комментарии, February 15.

Taagepera, Rein. 1992. “Ethnic Relations in Estonia, 1991.” Journal of Baltic Studies 23 (2): 121–32.

———. 1993. Estonia: Return to Independence. New York: Routledge.

Toft, Monica. 2002. “Indivisible Territory, Geographic Concentration, and Ethnic War.” Security Studies 12 (2): 82–119.

Torgler, Benno. 2003. “Why Do People Go to War?” Defense and Peace Economics 14 (4): 261–80.

Turu-uuringute AS. 2019. “Public Opinion and National Defense.” https://kaitseministeerium.ee/sites/default/files/elfinder/article_files/public_opinion_and_national_defence_2019_autumn.pdf.

———. 2020. “Public Opinion and National Defense.” https://kaitseministeerium.ee/sites/default/files/elfinder/article_files/public_opinion_on_national_defence_2020_autumn.pdf.

UNHCR. 2016. Mapping Statelessness in Estonia. Stockholm: UNHCR Regional Representation for Northern Europe.

Weber, Eugen. 1976. Peasants into Frenchmen: The Modernization of Rural France, 1870–1914. Stanford: Stanford University Press.

Williams, Cindy. 2005. “From Conscripts to Volunteers: NATO’s Transitions to All-Volunteer Forces.” Naval War College Review 58 (1): 35–62.

1.   A Minister of Population Affairs began this work in 1997 and oversaw the adoption of the first state integration plan in 2000. In 2009 the Ministry of Culture inherited this mandate.

2.   To Estonians, the statue is associated with the onset of Soviet occupation, but Estonian-Russians often consider it a tribute to Estonia’s liberation from Nazi occupation in 1944.

3.   In 1990, Estonia’s Defense League, a volunteer militia, attempted to place border posts along the frontiers of the interwar republic.

4.   It reads: “The land border of Estonia is determined by the Tartu Peace Treaty of 2 February 1920 and by other international border agreements” (Riigi Teataja 1992).

5.   NATO’s present footprint is light. It helps police Estonia’s airspace and provides a multinational battalion that consists of rotating soldiers through the Enhanced Forward Presence initiative.

6.   A partial exception to this occurred in the last of the nine polls conducted, when a subset of respondents had to be interviewed over the phone because of the COVID-19 pandemic in 2020.

7.   The National Center of Defense and Security Awareness has organized an interesting but less representative set of polls, targeting Russian-speaking students (NCDSA 2015, 2019, 2020).

8.   Almost one-fifth of Russian-speaking respondents replied “hard to say.”

9.   Defense willingness is sometimes operationalized using more abstract questions, which avoid stressing the personal risks involved, thus inflating affirmative responses (Cronberg 2006).

10.   Estonian-speakers’ mean trust in government is 3.19 (SD = .97). Russian-speakers’ mean trust in government is 3.23 (SD = .98). The difference is not significant: t(1018) = 0.58, p = .563.

11.   Estonian-speakers’ mean trust in the defense sector is 3.99 (SD = .85). For Russian-speakers it is 3.31 (SD = .96). The difference is significant: t(1018) = 11.5, p = .001.

12.   Among the two groups, support for NATO membership stands at 90 and 52 percent, respectively.

*   We wish to thank Linda Hart, Miina Kaarkoski, Teemu Tallberg, participants at the 2021 ERGOMAS conference, and the journal’s peer reviewers for their comments. Berglund acknowledges funding from the Foundation for Baltic and East European Studies (S2-20-0011) and is grateful to IOS Regensburg for hosting him in April 2022. The project “Attitudes and Stances towards National Defense among the Russian-Speaking Population” (SSVUH22136) enabled Kasearu’s and Kivirähk’s work on this article. The idea behind it came from our shared interest in the armed forces as an arena for nation-building. Kivirähk supplied the data and the empirical background information, Berglund positioned the argument vis-à-vis prior scholarship, and Kasearu conducted the logistic regressions. While each author drafted different portions of the manuscript and made substantial contributions, Berglund led the editing process. Of course, all three authors are equally responsible for the article’s strengths and shortcomings.

Appendix

Table A1. Means of Attitudinal Predictors among Estonian-speakers and Russian-speakers Willing or Unwilling to Defend.