ADAPTING TO DEMOCRACY: IDENTITY AND THE POLITICAL DEVELOPMENT OF NORTH KOREAN DEFECTORS

Abstract

Defection from North Korea to South Korea has increased dramatically, but little is known of its political consequences. Do North Korean defectors successfully adopt democratic norms, and if so, what factors aid this process? Through a novel survey of defectors, I find that national identification plays a significant role in motivating their fledgling sense of democratic obligation. Greater feelings of national unity with South Koreans lead to a stronger duty to vote and otherwise contribute to the democratic state. This effect is more powerful than that of conventional contractual factors, on which most state resettlement policies are based, and is surprising given that defectors’ nationalist socialization mostly took place under the authoritarian North. The findings suggest the need to reconsider integration approaches toward North Korean defectors and similarly placed refugees elsewhere.

They say South Korean tax money paid for your house, your clothes, your life here, so you owe your loyalty to this country now. You follow the laws and do your part. Have you ever seen real loyalty come from money? No, never. [ … ] You are not loyal to your family because of what it gives you.

North Korean defector, male, 39 years old Author interview, July 13, 2013

More than 29,000 North Koreans currently reside in South Korea, according to the Ministry of Unification (2017), a number that has grown rapidly since the early 2000s. Numerous studies have examined the economic, psychological, and health related aspects of this population's resettlement. However, surprisingly little is known about their political adaptation from one of the most stringent and closed off authoritarian regimes to a democracy. Yet great variation exists on this score. While some defectors become active participants in South Korea's democracy, others withdraw almost entirely from politics, becoming dead weight in the democratic process (Go 2014).

The democratic adaptability of North Korean defectors has direct implications for political stability in South Korea, but also more generally for host democracies receiving an increasing number of refugees fleeing from authoritarian contexts. The smooth integration of new citizens into the existing polity is an important part of how “democracy becomes routinized and deeply internalized” toward further consolidation (Linz and Stepan 1996, 5). In this context, understanding how North Korean defectors develop into responsible and responsive democrats is intimately tied to stabilizing the region's democratic future, especially as the potential breakdown of North Korea seems to loom larger than ever.

Becoming a democrat is a complicated process. But from a basic functionalist perspective, it is important that new members internalize a sense of responsibility for the roles of the democratic citizen. One of the key features that sets apart democracies from non-democracies is the strict limitation on how they can coerce citizens. Instead, democracies must often rely on the voluntary willingness of citizens to fulfill political roles such as voting, staying informed, or even paying honest taxes when monitoring is limited. For host democracies, it matters that new citizens come to see such roles as their responsibility, even in the absence of austere and authoritarian sanctions. Whether the influx of North Koreans becomes a supportive or disruptive force to South Korea and other receiving democracies will depend significantly on this aspect of their democratic development.

What motivates a sense of democratic obligation? The conventional view has been that political duty in liberal democracies can be fostered through a positive “give-and-take” with the state. As individuals receive satisfactory deliverables or treatment from the state, the argument is that political trust builds, and more are willing to reciprocate by complying with state demands. Indeed, this is the view that implicitly undergirds South Korea's resettlement policies, which focus predominantly on providing material aid through living, housing, and medical subsidies, as well as job and education assistance.

Yet the contractual approach represents only one side of a rich debate among political theorists about the sources of political obligation. The other side of that debate, tracing back to ideas of civic republicanism, claims that a different pathway exists in the intrinsic power of identity. Specifically, liberal national theorists claim that ethical ties to a national community can serve as the foundation for loyalty to a specific democratic state (Tamir 1995). In this framework, it is attachment to a national people—and the sense of collective obligation that it instills—that drives a sense of duty to a democratic state, not necessarily aspects of state performance or treatment. The aim of this article is to empirically assess these different frameworks in the context of North Korean defectors.

Understanding the different bases for democratic duty among defectors has important policy implications, but has received little empirical attention for two reasons. First, explicitly political surveys of the defector population are rare. While surveys on non-controversial topics such as health, employment, or social life are common, it is harder to obtain a large enough sample for more sensitive surveys that ask about political attitudes.

Second, empirically distinguishing between attitudes toward the nation versus state is challenging, since the two are often conflated in practice. For most new citizens, socialization into a new national community and new state happen simultaneously. From an inference perspective, this makes it difficult to assess causal direction, and it is unlikely that most survey respondents reliably separate the two concepts when answering questions.

North Korean defectors in South Korea are a particularly attractive sample in this regard. The politics of Korea's division resulted in two radically different states, but a shared understanding of the Korean nation as a singular, ethnically homogeneous community that spans across both states. Defectors therefore enter South Korea with pre-existing variation in their strength of national identification from socialization in the North. And while the causes of such variation are unknown, they are unlikely to be systematically correlated with attitudes post-entry into South Korea, since successful defection is not selectable. This setup establishes that among most defectors, national identification comes temporally prior to democratic attitudes and in a manner that greatly reduces potential for bias. Thus, a more reliable and accurate estimate of the role of national identification in democratic development is obtainable compared to most other observational contexts.

Using an original, face-to-face political survey of 228 defectors, I find that national identification significantly explains a sense of democratic duty, and more powerfully so than contractual factors such as satisfaction with state aid. Defectors who strongly identify with the Korean nation as a community that includes both North and South Koreans have a 33 percent greater sense of duty to vote in democratic elections—even when that nationalist socialization has mostly taken place in the authoritarian North. The findings demonstrate the highly contextual nature of political obligation based on identity, and suggest the need to diversify government policies toward North Korean defectors and similarly situated refugees elsewhere. A sense of belonging to the national community appears to be more than a desirable side effect of integration policies, but a real driver of successful integration itself.

AN IDENTITY-BASED THEORY OF DEMOCRATIC DEVELOPMENT

From where does a sense of responsibility for the roles of the democratic citizen arise? Questions about the basis of political duty have deeply engaged political theorists for centuries. This article explores a specific strand of the debate that emphasizes the role of group identities—particularly national community—as the source of political obligation to the state.

Individuals belong to many groups, but some groups are special in that they are seen as an integral and defining part of one's identity. Communitarian political theorists have long argued that such group memberships exert a moral force, instilling a sense of obligation to the collective welfare solely “in virtue of those more or less enduring attachments and commitments that, taken together, partly define the person that I am” (Sandel 1984, 90; also Walzer 1990) ¹

I argue that for many modern individuals, the nation is one such special group. The nation is a community of people who, for one reason or another, see themselves to share a history and future as part of the same political collective (Renan 1990 [1882]; Beissinger 1995). In many established states, citizens are socialized to feel as if they are a natural part of the nation—true to the word's root meaning of “to be born” (Verdery 1993):

In so many little ways, the citizenry are daily reminded of their national place in a world of nations. However, this reminding is so familiar, so continual that it is not consciously registered as reminding (Billig 1995, 8).

This perceived naturalness is what grants nations the power to elicit “special duties” from their members (Goodin 1988; Yack 2012). As Anderson (1983) explains, “For most ordinary people of whatever class the whole point of the nation is that it is interestless. Just for that reason, it can ask for sacrifices” (143–144).

For liberal national theorists, the nation's ethical capacity serves as a powerful basis for political obligation in liberal democracies. The connection lies in what Tamir (1995) refers to as the magic pronoun “my.” As Brubaker (2004) explains, “the feeling that this is my country, and my government—can help ground a sense of responsibility for, rather than disengagement from, actions taken by the national government” (121). When the state is seen to represent “my” national people, the collective welfare of one's nation becomes intimately tied to the welfare of that state. In such cases, national identification motivates a sense of obligation to contribute to that state by fulfilling one's citizen roles. In a democracy, such roles include paying taxes and obeying the law, but also voting, staying informed, and otherwise engaging in the democratic process. ² Fulfilling one's citizen roles to the state becomes a “ritualized means of fulfilling moral responsibilities” to one's national community (Wuthnow 1982, 135).

This identity-based model of democratic development departs from conventional approaches that emphasize a contractual relationship with the state. In prior work, citizen responsiveness to the state has been seen as the result of how fairly and transparently the state treats its citizens (Levi 1997; Tyler 2006) or a matter of reciprocal trust (Putnam 1993). Indeed, South Korea's policies toward defectors are mostly focused on providing satisfactory material aid and equal protection under the law. But a contractual framework assumes that the long-term trustworthiness of the state is reliably known. In the case of defectors, where such patterns of trust are still uncertain due to lack of experience with democracy, what complementary role, if any, does national identity play?

The theory is tested in the context of voting, a quintessential role of the democratic citizen. The prediction is straightforward: defectors who identify with the Korean nation as one that includes both North and South Koreans should feel a stronger duty to vote in South Korea's democratic elections, all else equal. Testing this prediction, however, is not as straightforward, as I explain in the next section.

EMPIRICAL STRATEGY

The relationship between national identification and duty to the state is difficult to specify. As Miller (1988) put it, “nationality is to a greater or lesser degree a manufactured item” by the state (654). Because of this circularity, it is possible that both national identification and citizen duty are the result of a strong state with an expansive socialization apparatus. A study that simply takes the difference in national identification across citizens as a determinant of the duty to vote therefore runs the risk of finding a spurious relationship. Priming experiments are a step forward, but since aspects of identity such as gender or national identity cannot truly be manipulated, they remain limited.

What we ideally want is this: a group of individuals, identical in national identity to the political community represented by a democratic state, to be launched de novo into the democracy. While such an experiment is not feasible, the subpopulation of North Korean defectors adapting to South Korean democracy comes very close within an observational context.

First, North Korean defectors enter South Korea, and democracy, for the first time already socialized to varying degrees into putatively the same national community—the result of the unique identity politics of the Korean division. ³ The aftermath of World War II split Korea into two very different states, but the belief of the Korean nation as a single, ethnically homogeneous community that spans across both states survived. Indeed, since the division, both states have claimed that each is the legitimate representative of the entirety of the Korean nation. Thus, North Korean defectors enter South Korea with pre-existing variation in identification with the national community of the new state—one that was solidified outside of a democratic context.

Second, the strong element of unpredictability in successful defection generates a high degree of independence between stable pre-defection characteristics and post-defection attitudes. Defection is a multi-stage process where, despite defectors’ best efforts to secure safe passage through higher pay or selection of quality brokers, luck and simple human mistakes introduce significant ad hoc uncertainty. Successful defection is not a condition into which one can self-select. Because of this feature, the worry that the original causes of variance in national identification in the North are somehow systematically related to a potential to develop democratic attitudes after defection is significantly diminished. ⁴

Outside of an experimental context, the defector population therefore offers one of the strongest empirical tests possible of the role of national identification in democratic development. The claim is not that this population meets all of the stringent criteria for causal inference; extremely few observational cases ever could. Rather, the claim is that the political context of this sample offers rare inferential advantages that shield it from the most prominent set of confounders that plague observational studies on this topic, yielding an estimate that should approximate the true causal relationship better than in most other cases.

POLITICS OF NATION AND DEFECTION BETWEEN THE TWO KOREAS

In the aftermath of World War II, the Korean peninsula found itself caught between the ideological tensions of the United States versus the Soviet Union. As a Soviet-backed Communist government took hold in the North, the United States supported the democratic efforts of the South. A violent civil war reached an armistice in 1953, solidifying the division into two states. Despite territorial separation, the belief of Korea as a single nation remained intact, largely due to the “racialization” of Korean identity during Japanese colonialism. To keep the Korean nation alive even in the absence of political sovereignty, nationalist leaders reimagined Korea as a blood-based collective (Robinson 1984; Shin 2006).

The belief of ethnic homogeneity played a central role in nation-building efforts by both states, with each government claiming to be the legitimate representative of all Korean people (Shin, Freda, and Yi 1999). For example, Lee (2010) finds minjok-jeiljuui—“the supremacy of the Korean race”—to be a core ideological principle taught in North Korean textbooks. Although South Koreans are often portrayed as a degenerate subgroup—in one textbook parable, South Korean children beg to “take us to North Korea with you” (Lee 2010, 365), and the regime routinely portrays South Korea as the bed of racial contamination (Myers 2011)—they are still defined as minjok and part of the Korean national community.

The same belief also survives in South Korea. National oneness with North Koreans is reflected in their exemption from multiculturalist policies that treat other diaspora as foreigners (Kim 2016). Only North Koreans are automatically granted South Korean citizenship upon entry and few South Koreans would dispute co-ethnicity with North Koreans in principle. It is important to note, however, that generational changes are underway. Among South Korean youth, the feeling that North Koreans are “one of us” is weakening as the importance of a shared bloodline to Korean identity is declining (Lee 2010; Sohn and Lee 2012; Kim 2014). Such shifts are not necessarily problematic for this study, since what matters is how defectors perceive South Koreans as part of their national community, not vice versa. But even in light of such trends, ethnicity still remains the single broadest basis of national identification on both sides (Grzelczyk 2014).

The two states obviously contrast in how national identity has been fused with specific ideology to assert each government's superiority over the other (Koh 1970; Hart 1999). But a “strong, almost mythical vision of homogeneity permeates both parts of Korea” (Bleiker 2001, 123; Jager 2003), and there is little doubt that both states aspire to represent the same nation. As one defector described: “We are part of the same village, just uptown and downtown” (Author interview, July 17, 2013). Thus, defectors enter South Korea already socialized, to varying degrees, into the national community represented by the new democratic state. ⁵

The Ministry of Unification (2017) most recently estimates that 29,830 defectors now reside in South Korea, with the first large influx in the late 1990s. From 1994 to 1998, a disastrous famine known as the “Arduous March” killed an estimated 5 to 10 percent of the entire Northern population (Goodkind and West 2001; Oh and Hassig 2004). The crisis forced the state to turn a blind eye to “marketization from below” (Haggard and Noland 2007)—the burgeoning internal and cross-border black markets that seeded the necessary networks for increased defection.

The decision to defect is not random, but successful defection is a multi-step process involving high uncertainty. The standard route begins by crossing the Tumen River to neighboring China. Since China does not recognize defectors as international refugees, however, defectors must find hiding to avoid being repatriated. From here, lack of permanent legal status prompts some defectors to find brokers to smuggle them to nearby Southeast Asian countries such as Laos, Thailand, or Vietnam that have a diplomatic history of repatriating to South Korea instead. The pathway to South Korea is riddled with luck and human mistakes that make eventual entry largely independent of the initial motivations for defection.

Why do most defectors decide to leave? Contrary to popular belief, a desire for political liberation or democracy is seldom the main driver. Defectors who left during and shortly after the great famine overwhelmingly cite economic and livelihood reasons (Oh and Hassig 1999; Lankov 2006). Table 1 shows the breakdown of motives in my sample, where respondents chose from the given response categories or could write in their own reason under “other.” The majority 62 percent chose economic or life-quality-related reasons, such as wanting a better life for their offspring or reuniting with family or friends who have already defected. ⁶ Even among the 37 percent that chose “political freedom,” follow-up qualitative interviews with volunteers from my survey sample reveal that respondents most likely meant something very different from a desire for democracy or free electoral government. Status is strictly monitored in the North based on loyalty to the regime (Hunter 1999), so that what respondents mean by political freedom is usually social mobility or freedom of choice in profession. One indication is that respondents who cited politics tended to be younger and better educated, with two out of three mentioning that an immediate or extended family member had been threatened or punished by the regime. The identity politics of Korea's division, combined with the mechanics of defection and entry into South Korea, therefore offers an unusually strong inferential setup for an observational study. The next section details the empirical measurements and specification.

Table 1 Primary Reason for Defection

DATA AND MEASURES

Sampling

The data are from an original, face-to-face survey of 228 defectors conducted in South Korea in July, 2013. To locate defectors, I established an initial connection with an instructor at a defector job training center. From there, I used snowball sampling—a method frequently used in qualitative studies to identify sensitive or non-obvious subjects—to branch out to other centers in the Seoul, Incheon, and Gyeonggi provinces, where the Ministry of Unification (2017) estimates that 64 percent of the defector population lives.

The head of each center was shown the full questionnaire before he or she decided whether the center would participate. Out of ten contacted centers, six participated and four declined due to the political nature of the survey. To cooperating centers, I made a $100 donation to improving facilities and course materials. While economic and life satisfaction surveys are fairly common, explicitly political surveys that ask about democratic attitudes and behaviors are unusual for this population.

The survey was administered after one to three classes at each center, with classes selected so that there would be no repeat subjects. The number of respondents per class ranged from 5 to 25. The instructor, usually a defector himself or herself, would introduce me to establish a base level of trust. I described the content and motive of the survey in Korean and gave students the option to leave if they did not wish to participate. For those remaining, all questions were reviewed and clarified in person before I left the room and waited outside while subjects completed the survey. The survey typically took about 30 minutes. Because subjects were identified through job training centers, Appendix 1 shows that the sample is slightly older and more educated than the national defector population. Since inferential leverage comes from variation within the sample itself, however, this representative skew is not an issue.

Measures

Duty to vote is the dependent variable. In addition to the usual social desirability bias toward participation in democracies, defectors can feel added pressure from being habituated to mandatory voting in confirmatory elections in the North. To minimize over-report, I adapt the wording developed in Blais and Achen (2010) that offers “voting as a choice” as a socially acceptable and particularly appealing alternative for defectors experiencing free elections for the first time:

Different people have different opinions about voting. Some say that voting is a responsibility and you should vote even if you do not like any of the candidates or parties. Others say voting is a choice and you should only vote if you like a candidate or party.

For you personally, is voting more of a responsibility or choice? [Voting is a responsibility, Voting is a choice]

How strongly do you think that way? [Very strongly, somewhat strongly, not very strongly]

Strength of national identification is the main independent variable. Communitarian theorists argue that it is not simply classification into a group, but the psychological or affective attachment to being part of the group that generates moral force. Specifically, it is a defector's attachment to the Korean nation as one that includes both North and South Koreans—the feeling of truly being one with all Koreans—that is the causal lever. Almost all respondents—92 percent of the sample—agree with the statement that North and South Koreans are both “my national people.” Indeed, national unity is one of the core principles emphasized at Hanawon, the government resettlement facility that all defectors are required to stay in for 12 weeks after being cleared for security (Cho and Kim 2011). ⁷ What is needed is a measure that taps how much this belief is actually felt, something that is not easily captured in standard measures of identification.

To develop an accurate measure, I conducted multiple focus groups with recent defectors who have been in South Korea for less than a year—those who are closest to their pre-defection nationalist socialization. Participants in the focus groups were also recruited from the job training centers, but as part of a separate qualitative project; they therefore did not participate in the actual survey. ⁸

I asked participants to focus on specific instances in their daily lives in which they truly felt national oneness with South Koreans. Each story was different, but a common thread emerged: the sharing of similar responses or emotions to stimuli. For instance, one defector mentioned how South Koreans share the particular satisfaction from enjoying a spicy Korean soup on a cold day; another mentioned cheering alongside South Koreans for figure skater Kim Yuna, who won the Olympic gold over Japan. Indeed, social psychological research shows that a strong proxy for the strength of interpersonal attachments is the empathic sharing of positive emotions (Gable et al. 2004). Therefore the following measure was developed:

When there is a joyful event for South Koreans, do you feel happy as if it were your own? [Strongly

agree, somewhat agree, somewhat disagree, strongly disagree]

The estimation model also included motivations to vote that are commonly seen as political incentives to engage with the state, so as to capture the contractual approach. The first measure is satisfaction with state aid. Defectors receive financial assistance from the South Korean government for housing, medical, and life expenses, as well as non-monetary aid in education and vocational training. I asked the level of satisfaction with each type of aid and calculated the average for each respondent. The prediction is that greater satisfaction with state services motivates a stronger duty to vote out of reciprocity, gratitude, or to secure the continuation of such aid. The second measure is perceived fairness of the state, based on the literature on procedural fairness and compliance with the state. The question asks how much defectors trust the government to protect them as equals to South Koreans in the event of a problem.

As controls, I include measures of political socialization, both before and after defection, that likely affect one's duty to vote. Collective values—the belief that the individual should always sacrifice for the group—addresses the possibility that the duty to vote is a shadow of North Korea's collectivist and compliant ideology (Koh 1970). Level of political interest accounts for variation in the duty to vote that is attributable to a penchant for politics. The proportion of South Koreans in one's personal network proxies the degree of social integration.

Finally, a range of demographic and socioeconomic variables related to voting is included. Years since defection to South Korea counts the length of democratic exposure and familiarity with the democratic process. Female accounts for gender differences in the adaptation experience in general (Jeon et al. 2005; Jeon, Yu, and Lee 2011), as well as the higher sense of duty to vote among women found in comparative studies elsewhere (Blais 2000; Campbell 2006). Age accounts for the rigidity of one's prior non-democratic beliefs based on socialization theory (Greenstein 1965; Krosnick and Alwin 1989). Party member is a binary indicator for registered member status in the North, which accounts for differential exposure to authoritarian values as well as participatory experience prior to defection. I also include highest level of education in the North. Appendix 2 shows the descriptive statistics.

EMPIRICAL RESULTS

How does strength of national identification affect defectors’ fledgling sense of democratic duty, specifically the duty to vote? The starting assumption is that the strength of national identification socialized in the North is a stable, slow-moving part of self-identity and therefore largely independent of the immediate post-defection experiences with the South Korean state or society. Appendix 3 shows evidence that is consistent with this assumption.

The estimation model is as follows. Individual i’s sense of duty to vote is modeled as a function of her strength of national identification, contractual motivations for feeling a responsibility to vote, and a set of demographic and attitudinal covariates, with standard errors clustered by job training centers in which the survey was administered:

$$\eqalign{duty\,\, to \,\, vote_{\rm i} =\; &\beta _{\rm 0} + \beta _{\rm 1} (national\,\,identification)_{\rm i} \cr & + \beta _{\rm 2} (incentives)_{\rm i} + \beta _{\rm 3} (controls)_{\rm i} + \varepsilon} $$

The way that the duty to vote variable was asked lends itself to different functional forms: a binary measure (duty to vote vs. choice) or a continuous measure of the strength of duty to vote. Table 2 shows results for both functional forms. Across both models, strength of national identification has a positive and significant effect on the duty to vote. ⁹ In fact, it is consistently the most powerful variable. Substantively, going from weak to strong national identification leads to 15 times higher odds of seeing voting as a duty (Model 1) or a 33 percent stronger duty to vote (Model 2), all else equal. Surprisingly, neither perceived fairness of the state nor satisfaction with state aid has any meaningful effect on the duty to vote once national identification is included.

Table 2 National Identification with the State and the Duty to Vote

***p < .01, **p < .05, *p < .10. All variables rescaled 0–1.

Figure 1 shows the predicted marginal effect of national identification on the duty to vote based on Model 2. As a defector moves from weak to strong national identification with the South Korean state, her predicted duty to vote increases from zero to over 0.20 on a 0 to 1 scale. That is, even after accounting for well-known correlates of voting, national identification with the state can be the difference between an individual who feels no duty to vote versus one who does. This result is more robust for subsets of defectors for whom the potential for confounders is least likely. Model 3 looks at recent defectors only—those who have lived in South Korea less than two years—for whom bias from reverse causality is least likely. The effect of national identification is 10 percentage points larger than in the full sample. Another subgroup is defectors who left during or shortly after the great famine, where hunger and immediate survival—not any desire for political freedom—was likely the main reason for escape (Good Friends 1999, 14). For this famine subset, Model 4 shows that the main effect is again 16 percentage points greater than in the full sample.

Both variables rescaled 0–1. Prediction based Model 3 in Table 3 using actual values of covariates. Error bars mark 95% confidence intervals.

Figure 1 Predicted duty to vote of defectors by strength of national identification

Alternative Explanations

Several alternative explanations could account for the patterns, which I address in turn. The most obvious one is that the results could be explained by a “joiner” personality instead. Joiners may actively seek a stronger connection with the national community by pursuing opportunities to be part of the collective, of which voting in elections happens to be one. If this were the case, then national identification should have a positive effect on all kinds of communal participation, not just those that contribute to the state. The survey included a measure of the duty to pay taxes, ¹⁰ another citizen activity that supports the democratic state, as well as a battery on involvement in cultural, religious, or social organizations, which do not. Table 3 shows that, while stronger national identification leads to 15 percent greater duty to pay taxes, it has no meaningful effect on forms of participation that are unrelated to state welfare. These patterns suggest that the main results are indeed driven by the theorized mechanism.

Table 3 Effect of National Identification on State-related versus State-unrelated Participation

***p < .01, **p < .05, *p < .10. OLS regressions with all variables rescaled 0–1.

Second, defectors with strong versus weak national identification might differ in other ways that affect the duty to vote. Strength of national identification is not randomly assigned, and we cannot observe all of the factors that influenced it in the North. To address this issue, I use propensity score matching to control for the observed differences between weak and strong national identifiers (Rosenbaum and Rubin 1984; Dehejia and Wahba 2002; Austin 2011). For every strong national identifier, I matched her with the most comparable weak identifier counterpart based on age, gender, party membership, and class in the North. ¹¹ I then re-estimated the average treatment effect of strength of national identification on the duty to vote among matched pairs only.

Even after matching, strong national identification still produces a 17 percent higher duty to vote, as shown in Table 4. The test demonstrates that it is unlikely that the positive effect of national identification on the sense of duty to vote is reducible to systematic differences between strong versus weak national identifiers from the North. A formal sensitivity analysis (Imbens 2003) in Appendix 5 further suggests that the existence of an unobserved third factor that would render the effect of national identification to be spurious is highly unrealistic.

Table 4 Matched Re-estimate of the Effect of National Identification

Weak versus strong national identifiers matched on age, gender, party membership, and social class in the North. N is the number of units used for nearest neighbor matches. Bootstrapped standard error with 500 replications.

A third concern comes from the stable, but not static, nature of national identification. As defectors stay longer in South Korea, experiences with the South Korean state or society can gradually begin to influence the baseline national identification they came with from the North. Then we might spuriously observe an association between national identification and pro-state attitudes, such as seeing voting as a duty, not because of an independent effect of their national identification as I hypothesize, but because of a series of positive experiences with the South Korean state or society. While there is no way to control for such time-dependent contamination of the independent variable with observational data, it is possible to estimate the extent of confounding.

Bounds analysis is appropriate for this kind of problem, because it allows the researcher to simulate, in a very transparent manner, the degree of contamination over time and how it changes the overall findings (Keele and Titiunik 2014). A model of “interference” is specified and the main effect of interest is re-estimated to produce bounds around the original estimate. The relevant unit of bounds is number of years lived in South Korea. Let I denote the interference set—the set of subjects for whom we suspect contamination from state or society feedback. To form I, I identify all subjects with strong national identification within year bound j. For those subjects, I assume that an observed positive outcome Y _i  = 1 (voting as duty) is fully a function of state feedback and re-compute the interference-free outcome $\hat Y_t = 0$ (voting as choice). After replacing Y _i with $\hat Y_t $ for all subjects with strong national identification in I, I recalculate the main effect of national identification. Bounded effects are generated by varying the scope of j, which ranges from 0 to 14 years of residence in the sample.

Table 5 shows that the main effect of national identification is not reducible to such contamination. The effect of national identification on the sense of duty to vote changes little up until the point of a five-year bound, which assumes that national identification is entirely contaminated by experiences with the South Korea state or society for all defectors who have stayed five years or more. It is only under the widest and most unrealistic bounds that the main effect disappears, where strong national identification is assumed to be the result of state or society feedback for every non-recent defector. In other words, feelings of national unity play an important role in shaping defectors’ sense of democratic duty up until at least five years after entry—arguably the most formative phase of their political development.

Table 5 Bounded Effect of National Identification on the Duty to Vote

Estimates are the coefficient on national identification using the same specification as Model 3 in Table 3, while varying the “interference set” by years of residence in South Korea (j).

Does a sense of duty to vote actually matter for the electoral participation of defectors? Numerous studies based on native citizens of mature Western democracies suggest that it should (Campbell 2006; Blais and Achen 2010; Gerber, Green, and Larimer 2008). In the absence of validated turnout for defectors, Table 6 shows how the duty to vote predicts reported turnout in the most recent federal election. A strong sense of duty to vote predicts 11 times higher odds of turnout, meaning that each incremental increase in its strength is likely to have real consequences for how regularly these new democratic citizens participate.

Table 6 Duty to Vote and Turnout among North Korean Defectors

***p < .01, **p < .05, *p < .10. Logistic regression with variables rescaled 0–1.

The data paint a consistent, yet counterintuitive picture. That stronger national identification might be associated with greater duty to vote or turnout is perhaps not very surprising in most advanced democracies. However, the fact that we find this relationship to hold even in the case of defectors, where nationalist socialization occurred in the authoritarian North, is quite surprising. For many defectors, it appears that their national identification solidified in the North is a significant part of what seeds their nascent sense of democratic duty as a new citizen of the South.

CONCLUSION

Coming from one of the most closed off, stringently controlled authoritarian regimes, North Korea defectors represent a hard case of adapting to democracy. What factors aid their political development into responsible democrats once in South Korea? What can be learned from this case about the democratic adaptability of citizens coming from post-authoritarian contexts in general?

Conventional approaches to the integration of refugees have focused predominantly on building positive citizen–state relationships. Receiving host states often spend a great deal of resources to offer satisfactory deliverables, implicitly in expectation of political loyalty from their newest members down the road. South Korea's integration policies toward defectors are no exception in this regard, focusing on providing adequate material aid, assisting with socioeconomic integration, and trying to ensure fair legal treatment.

This article finds that identity—particularly a sense of belonging to the national community of the new state—also matters. In fact, in the case of North Korean defectors, a sense of duty to the new state appears to be grounded more in shared national identity than in contractual factors such as satisfaction with state aid or perceived fairness of state treatment. Strength of national identification consistently carries more explanatory weight than either contractual variable across all specifications of the duty to vote.

In most other observational contexts, this kind of analysis is plagued by the simple problem of simultaneity: new citizens are usually exposed to a new national community and new state at the same time, making it nearly impossible to discern the direction of the relationship. However, the unique identity politics of North Korean defectors in South Korea offered a setup in which national identification credibly precedes exposure to a new, democratic state. The findings therefore offer a rare and clear picture of the role of national identity in how post-authoritarian individuals adapt to democracy.

The policy implications that follow are quite distinct from the current approach toward defectors in South Korea. This article shows that a greatly underemphasized aspect of adapting to democracy is the psychological shift in national identity that often accompanies political transitions. Currently, the task of sustaining or strengthening feelings of national belonging for defectors is primarily left to informal channels of churches, volunteer groups, or other non-profits. More formal programs, such state-sponsored inclusive campaigns, recurring naturalization rituals, pro-diversity curricula in schools, and more stringent regulations against discriminatory practices could produce real gains in defector commitment to the democratic state.

Of course, North Korean defectors in South Korea are a unique sample, and any efforts to extend the findings here to other political refugee populations warrant careful comparisons. In particular, the way that Korean national identity is primarily understood in ethnic terms probably lends itself to easier co-national identification for North Koreans. But the political traumas that serve as barriers to adapting to democracy for defectors are not unlike those experienced by political refugees from similarly exploitative and authoritarian regimes. At the very least, this study offers a diversified template for researchers studying the democratic adaptability of political refugees elsewhere.

A limitation of the present study is that it only examines a small part of what it means to be a democrat. What constitutes a good democratic citizen is a multi-dimensional and ever shifting concept (Gutmann 1999), of which having a sense of duty to vote or pay taxes is an important, but limited part. Democratic citizenship includes not only state-supportive actions, but also the responsibility to voice dissent against the state when needed. How national identification affects such forms of counter-state or protest participation in democracies is not examined in this study, but points to an important avenue for future research.