Autocratic regimes routinely employ tools such as border walls, the selective issuance of passports, or even brute force to control their citizens’ freedom of international movement (Dowty Reference Dowty1987). To this day, emigration remains restricted in autocracies such as China, Iran, and North Korea (US Department of State, 2023). According to V-Dem data, 79% of contemporary autocracies—compared to only 4% of democracies—impose at least some restrictions on their citizens’ freedom of foreign movement.Footnote 1 What are the consequences of such emigration restrictions?
Past scholarship typically argues that emigration restrictions function as a “safety valve” (e.g., Barry et al. Reference Barry, Clay, Flynn and Robinson2014; Hirschman Reference Hirschman1970; Reference Hirschman1978). Autocratic governments allow the departure only of those individuals whose removal would most benefit their rule. This logic hinges on the central role of exit visa systems in closed regimes. Where exit is restricted, citizens must apply for a permit to leave, an exit visa. The application process itself acts as a screening mechanism (Michel, Miller, and Peters, Reference Michel, Miller and Peters2023): because applying is politically costly, those who persist through the process reveal themselves as potential troublemakers. Allowing their emigration, then, helps reduce internal pressure on the regime.
This study demonstrates that emigration from closed regimes can also have a secondary, countervailing effect that has received less attention in past work. I propose that emigration in one period can raise public demand for emigration among left-behind citizens. This creates new challenges for the government in later years precisely because exit visa systems restrict emigration to those who dare to challenge the regime. Restrictive emigration regimes can inadvertently reward antiregime behavior—a point that Michel, Miller, and Peters (Reference Michel, Miller and Peters2023) theorize about but do not test empirically.
Evidence for these arguments comes from an analysis of the emigration system of the former German Democratic Republic (GDR), an emblematic case of a closed autocracy with little freedom of international movement. My empirical analysis centers on an emigration reform in late 1983, when GDR leadership gave in to international pressure and liberalized the country’s emigration regime significantly. The reform led to the largest emigration wave since the construction of the Berlin Wall. The number of applications for exit visas skyrocketed, and an additional 65,200 citizens left. Although the regime retained control over who was allowed to leave, the large emigration wave that followed offers scholars a unique opportunity to study and quantify the consequences of emigration from an otherwise closed regime.
To that end, I leverage novel panel data on an outcome that played a key role in internal considerations about who was allowed to leave: criminal activity. Criminality featured prominently in the logic of the GDR’s emigration system, which criminalized emigration and attempted exits. Moreover, as I document, repeatedly breaking the law was a strategy commonly used by those seeking emigration to force their exit. In addition, the East German government explicitly sought to use emigration to get rid of criminals and other troublemakers. My data allow me to examine whether this strategy was successful.
A difference-in-differences approach demonstrates that emigration was indeed associated with criminal activity: I document a substantial drop in all forms of crime and in their severity in places with higher emigration immediately after the reform. This result aligns with the idea that the GDR’s exit visa system allowed the regime to identify and remove potential troublemakers. However, this safety valve effect was short-lived: there was a significant increase in politically motivated crimes several years after the emigration wave. This uptick was concentrated among border-related political crimes: citizens from high-emigration counties became more likely to attempt to cross the border illegally. Panel data on emigration-related petitions to the GDR government support the interpretation that this increase in border-related crimes stemmed from growing demand for exit among left-behind citizens.Footnote 2
This study makes key contributions to research on the consequences of emigration for authoritarian governance. Theoretically, I identify important limitations to the use of emigration as a safety valve (Barry et al. Reference Barry, Clay, Flynn and Robinson2014; Hirschman Reference Hirschman1978; Sellars Reference Sellars2019). In demonstrating that such a strategy can backfire in later years, exacerbating rather than alleviating pressure on the regime, I emphasize the need to distinguish between the short- and long-term consequences of emigration from closed regimes. The data I assemble for this project expand the scope of past work on emigration from authoritarian regimes in general by moving beyond the conventional focus on democratic quality, political contestation, or other high-profile systemic outcomes (e.g., Barsbai et al. Reference Barsbai, Rapoport, Steinmayr and Trebesch2017; Peters and Miller Reference Peters and Miller2022; Spilimbergo Reference Spilimbergo2009) and on emigration from the GDR in particular by studying an earlier period, rather than the time around the East German revolution (e.g., Hirschman Reference Hirschman1993; Pfaff Reference Pfaff2006).
In demonstrating how selective emigration can backfire for autocratic governments and create more demand for exit in the future, I build on a large literature on migrant networks and the cumulative causation of migration (e.g., Beine, Docquier and Özden Reference Beine, Docquier and Özden2011; De Haas Reference De Haas2010; Massey Reference Massey1990; Massey and Zenteno Reference Massey and Zenteno1999). This literature shows that international migration tends to be self-reinforcing. Migration creates and deepens migrant networks that then lower the costs and raise the benefits of subsequent migration. My contribution to this literature is to illustrate very similar patterns in a least likely case: despite its high state capacity, vast repressive apparatus, and tight control over the border, the GDR was ultimately unable to find an effective solution to growing migratory pressure that it had sought to eliminate through the initial emigration wave.
My research also contributes to nascent scholarship on how authoritarian sending countries formulate their migration policies (e.g., Fitzgerald Reference Fitzgerald2006; Iskander Reference Iskander2010; Norman Reference Norman2020). I concur with a key assumption in this line of work that autocrats are rational actors who employ emigration policy strategically (e.g., Miller and Peters Reference Miller and Peters2020). However, my research reveals that emigration had unanticipated consequences for the regime, and even though the government quickly realized these repercussions, it proved unable to counteract them. Moreover, the GDR regime did not act in an international vacuum: it confronted mounting international pressure to liberalize its emigration regime. In describing these international pressures, this study speaks to scholarship on “migration diplomacy” (Tsourapas Reference Tsourapas2021) and the foreign policy dimensions of autocratic emigration regimes (for an overview, see Adamson and Tsourapas Reference Adamson and Tsourapas2018). Much of this literature focuses on how states’ foreign policy interests inform their own migration policies (e.g., Oyen Reference Oyen2016; Tsourapas Reference Tsourapas2018b).
My contribution is to study an instance where more powerful actors influenced the authoritarian sending state. As such, I focus on the flipside of what Greenhill (Reference Greenhill2010) calls “coercive engineered migration.” Rather than using migration as strategic leverage to achieve other foreign policy goals (see Tsourapas Reference Tsourapas2018a for another example), West Germany and the Soviet Union used their own leverage to influence the GDR’s emigration policy and, perhaps unintentionally, the GDR’s eventual breakdown.
This study is most closely related to Pfaff’s (Reference Pfaff2006) work on the role that emigration played in the demise of the former GDR. I depart from his work by drawing attention to criminal activity in the years before the GDR’s collapse, rather than on political protest immediately before and during the East German revolution. In doing so, I introduce novel data that allow me to observe political and social developments in an otherwise highly secretive regime.
The Logic of Emigration Restrictions
I propose that emigration has countervailing effects in contexts where it is restricted. As noted in past scholarship, emigration restrictions serve as a screening device that can help identify and remove troublemakers—a mechanism often summarized in the idea of emigration as a safety valve. My contribution is to emphasize that emigration can also backfire, especially in later years. It can inadvertently fuel future demand for emigration and thus create new challenges for the regime. This section develops both arguments in greater detail and illustrates the key mechanism using archival material on the former GDR.
Emigration Restrictions Serve as a Screening Device
Emigration restrictions allow the autocratic regime to handpick who leaves (Peters and Miller Reference Peters and Miller2022). Autocrats seeking to inoculate their rule against threats from the population can use these powers to restrict emigration to those citizens who oppose their rule today or might challenge it tomorrow (henceforth, “troublemakers”). However, doing so requires the government to identify them.
This task is relatively simple for citizens with low “revolutionary thresholds” (Kuran Reference Kuran1991). Those citizens are willing to openly confront the government even when few others join them. The regime knows their identity because this “active opposition” routinely challenges the government out of ideological conviction. It is thus unsurprising that closed regimes routinely expel some of their most vocal critics, such as Aleksandr Solzhenitsyn (USSR, 1974) or Wolf Biermann (GDR, 1976).
Identifying citizens with medium revolutionary thresholds is more complicated. These members of the “latent opposition” (Miller and Peters Reference Miller and Peters2020) conceal their political attitudes and are only willing to reveal themselves in public if enough other citizens join them. In normal times, these individuals are indistinguishable from citizens whose revolutionary thresholds are so high that they are unlikely to threaten the dictator. But unlike such citizens, small changes in public approval or economic conditions might turn these members of the latent opposition against the regime. How, then, can the regime distinguish emigration seekers who are part of the latent opposition from those who are not?
Building on insights by Michel, Miller, and Peters (Reference Michel, Miller and Peters2023), I propose that exit visa systems can accomplish this task. The presence of emigration restrictions means that everyone who wants to leave the country legally has to apply for an exit visa. Such applications come with extremely high political costs. “Persecution and harassment [were] often employed to intimidate applicants” in the USSR (Pettiti Reference Pettiti, Vasak and Liskofsky1976, 181), for instance. Applicants there routinely challenged the authorities in public to prove their opposition to the regime (Polian Reference Polian2004).
The costs of an exit visa application were similarly high in the former GDR (Lueders Reference Lueders2021, 44–45).Footnote 3 Applicants had to prove themselves as hartnäckige Antragsteller (“persistent applicants”) during the application process, during which they faced routine intimidation and harrassment by secret police officials (Mayer Reference Mayer2002).Footnote 4 A first application was typically rejected by default to probe applicants’ resolve.Footnote 5 Efforts by government officials to “win them back” (Rückgewinnungsmaßnahmen) included showing applicants reports about dismal living conditions in West GermanyFootnote 6 and letters from East German refugees begging the authorities for the “privilege” to return to the GDR.Footnote 7 Expulsions from university, demotions, or layoffs, which targeted the applicant’s broader family members as well,Footnote 8 also occurred frequently.Footnote 9
In addition, secret service officials looked for evidence that applicants were fundamentally opposed to East Germany’s government and socialism. Because authorities considered how an exit would influence the country’s “inner stability,” many applicants sought to challenge the regime in public.Footnote 10 Prominent strategies involved quitting one’s job,Footnote 11 leaving socialist mass organizations (Mayer Reference Mayer2002), abstaining from elections,Footnote 12 or engaging in öffentlichkeitswirksame Aktionen (“visible acts of opposition”).Footnote 13 The last strategy was particularly widespread, involving activities such as the public display of political messages or white ribbons on car antennas.Footnote 14 Finally, and of particular relevance for the purpose of this article, many applicants repeatedly broke the law.Footnote 15 All these actions meant that applicants were in direct confrontation with the government, despite the high costs associated with these actions. Consequently, they can be interpreted as credible signals of applicants’ resolve.
In short, emigration restrictions and exit visa systems serve as “an ideal vetting system” (Corrales Reference Corrales, Tulchin, Bobea, Prieto and Hernández2005, 56). They help address what Wintrobe (Reference Wintrobe1998) coined the “dictator’s dilemma”: the dictator’s uncertainty about the population’s true level of support. Autocratic leaders are usually theorized to use surveillance (Dimitrov and Sassoon Reference Dimitrov and Sassoon2014) or citizen complaints (Dimitrov Reference Dimitrov2014; Lueders Reference Lueders2022) to reduce this uncertainty. Exit visa systems, in contrast, turn information gathering upside-down: they are screening devices that encourage troublemakers—especially those difficult to identify through other strategies—to reveal themselves.
This screening mechanism can have clear benefits for the regime (Michel, Miller, and Peters Reference Michel, Miller and Peters2023). Citizens who go through the application process credibly signal that they might challenge the regime in the future. As Hirschman (Reference Hirschman1970, 61) recognized in his work on the interplay between exit and voice, the emigration of these troublemakers constitutes a “conspiracy in restraint of voice.” It serves as a safety valve and makes mobilization against the regime or other acts of opposition less likely (Barry et al. Reference Barry, Clay, Flynn and Robinson2014; Hirschman Reference Hirschman1978).
Emigration Raises Public Demand for Further Emigration
Yet, this safety valve effect can also backfire for the regime and inadvertently raise demand for emigration among left-behind citizens in later years. As noted in an internal GDR government report, “every permitted exit generates, as a rule, one or two new applicants” (cited in Pfaff Reference Pfaff2006, 79).
This backfire effect is driven by two complementary mechanisms. One is rooted in what has been described as the cumulative causation of migration (Massey Reference Massey1990)—the idea that “once started, migration processes tend to become partly self-perpetuating” (De Haas Reference De Haas2010, 1587). An initial emigration wave creates new cross-border family ties and transnational associations. These migrant networks provide support and resources, which lower the costs and raise the benefits of subsequent migrations (Beine, Docquier, and Özden, Reference Beine, Docquier and Özden2011; Liu, Reference Liu2013). The result is the “self-feeding character of international migration” (Massey and Zenteno Reference Massey and Zenteno1999, 5330).
Migrant networks played a crucial role in sustaining demand for emigration in the former GDR. Each additional emigrant deepened what were already very strong family and social ties between East and West. These ties were important because Rückverbindungen (“return connections”) gave East Germans a reason to apply for exit (Davis, Stecklov, and Winters Reference Davis, Stecklov and Winters2002). At the same time, East German émigrés often joined organizations abroad—such as the West German International Society for Human Rights (Internationale Gesellschaft für Menschenrechte e.V.) or Calls for Help from Over There (Hilferufe von drüben e.V.)—to gain access to their resources and information.
The second mechanism is based on the idea that successful emigration in one period generates knowledge about the application process (Eisenfeld Reference Eisenfeld, Poppe, Eckert and Kowalczuk1995). As detailed earlier, the goal of the GDR’s exit visa system was to restrict emigration to those most likely to challenge the autocratic regime. Because committing crimes raised one’s chances of obtaining an exit visa, the exit visa system inadvertently rewarded such behavior. Information about the best strategies is likely to spread among the population through a process of social learning (Bandura Reference Bandura1977): citizens learn new behavior through observation and imitation of the behavior of others, especially when that behavior is seen as successful. In other words, witnessing their peers first challenge the autocratic government and then be granted an exit visa and leave the country generates new “social beliefs” (Bikhchandani et al. Reference Bikhchandani, Hirshleifer, Tamuz and Welch2024) among left-behind citizens about the most effective actions.
This social learning creates new challenges for regimes with closed borders precisely because they restrict emigration to individuals who can credibly signal their opposition. Indeed, once East Germans learned that citizens engaging in public acts of opposition were more likely to obtain an exit visa, applicants increasingly challenged the regime in public (Pfaff Reference Pfaff2006).
Importantly, we should expect the increase in demand for emigration to be locally concentrated. The East German government suppressed discussions of emigration and closely monitored any individual or group suspected of planning an exit. In this context, the relevant information could only spread through word of mouth in trusted social networks or through locally organized groups of emigration seekers. Moreover, up until the last year of the GDR, emigration occurred mostly in secret. Hirschman (Reference Hirschman1993, 194), for example, quotes a Protestant bishop who laments that people during the 1984 emigration wave left “heimlich, still und leise” (secretly, softly, and silently). This suggests that it was easier to learn about a realized exit through the sudden departure or disappearance of a neighbor, friend, or relative near by than ones elsewhere in the country.
That exit in one period can at times backfire and generate more public dissent and “voice” was described in Hirschman’s (Reference Hirschman1993) discussion of the interplay between exit and voice in the demise of the GDR in 1989. He observed that emigration throughout that year encouraged more East Germans to defy the regime and take to the streets to demand exit (cf., “Wir wollen raus” / We want out). At the same time, it mobilized the Bleiber, those who wanted to remain in the GDR and seek reform from within (cf., “Wir bleiben here” / We’ll stay here). This article contributes to this line of work by showing how the see-saw mechanism between exit and voice that Hirschman (Reference Hirschman1978; Reference Hirschman1993) described began to reverse before 1989.
Emigration Complements Surveillance and Repression
Past work on authoritarian durability describes how dictators use surveillance to identify regime opponents (Dimitrov and Sassoon Reference Dimitrov and Sassoon2014) and use repression to deal with popular discontent (Shen-Bayh Reference Shen-Bayh2018). Emigration and emigration restrictions can complement these two strategies.
Emigration restrictions complement surveillance by enabling the regime to identify latent opponents. Even the strongest surveillance apparatus is typically unable to monitor all citizens. The East German Stasi, for example, is widely seen as one of the largest and most effective secret police forces in history. Still, it was largely focused on the public sphere, social organizations, and known dissident groups. Because many East Germans had escaped into private life and rarely shared their political views openly (Popplewell Reference Popplewell1992), the Stasi was unable to identify precisely those individuals who are the target of the exit visa system: latent opponents who remain quiet in normal times.
Emigration also complements multiple forms of repression (Alemán and Woods Reference Alemán and Woods2014). First, emigration is more feasible than outright execution, which is reserved for the most active opponents. Executing opposition figures is risky because it can provoke international sanctions or backlash at home (Esberg Reference Esberg2021). Emigration is also more feasible than imprisonment, which again targets active opponents. The former GDR imprisoned few citizens for life. Imprisonment became even less feasible over time as the GDR sought to reduce its prison population and so improve its human rights record (Raschka Reference Raschka2001). Lastly, emigration can be more successful than state harassment. Although the target of both strategies may be the same—latent opponents—state harassment can backfire and raise discontent. The GDR’s Zersetzungsmaßnahmen (“degradation measures”), for instance, were intended to make life harder for opponents. Known opponents were often excluded from higher education, were assigned lower-quality housing, or had to work less attractive jobs. However, such poor economic prospects could increase the demand for emigration. After they became the target of these measures, citizens had little to lose by applying for an exit visa (Pingel-Schliemann Reference Pingel-Schliemann, Conze, Gajdukowa and Koch-Baumgarten2009).
Summary
In sum, the high political costs that come with exit visa applications allow the government to identify potential troublemakers. This “screening mechanism” has benefits for autocratic rulers, identifying and removing those emigration seekers who are most likely to challenge them. However, emigration in one period can also raise demand for emigration among left-behind citizens in the future. This “demand mechanism” can counteract the benefits of the screening mechanism by creating public opposition.
The GDR’s Emigration Regime
After construction of the Berlin Wall in 1961, the former German Democratic Republic became a paradigmatic case of a closed autocracy with little freedom of foreign movement. I focus on an emigration policy reform in late 1983, when the GDR gave in to international pressure and liberalized its emigration regime. It led to the emigration of an additional 65,200 citizens and offers researchers valuable insights into how the regime determined who was allowed to leave and the consequences of the resulting selective emigration.
Closing the Border
Concerns surrounding international migration played a key role throughout the history of postwar East Germany. The advance of the Red Army toward the end of World War II set in motion mass refugee movements westward from Germany’s former Eastern territories. The end of the war was accompanied by further refugee streams from East to West; many East Germans preferred the economic opportunities and political freedoms in West Germany to the emerging socialist order in the GDR. East German internal statistics report that the country lost about 250,000 individuals to West Germany every year in the first few years after its founding.Footnote 16
The government quickly grew worried about this mass emigration. Most of the émigrés were young and well educated, which raised concerns about the economic repercussions of this sustained loss of labor (Hirschman Reference Hirschman1993). Mass emigration was also at odds with the regime’s claim to widespread popularity and, in the context of the systemic conflict that emerged during the Cold War, the claimed superiority of socialism (Dowty Reference Dowty1987). To solve the emigration crisis, the regime declared emigration illegal in 1952 and began to erect physical barriers to seal off the land border between East and West Germany (Geier Reference Geier, Dietrich and Geier2019).Footnote 17 These actions remained largely unsuccessful, however. Although the closing of the inner border prevented East Germans from directly crossing into the West, the open border between East and West Berlin continued to offer East Germans a path to freedom. As emigration rates kept growing, the government ultimately decided to build a wall around West Berlin in August 1961. Emigration plummeted to between 25,000 and 40,000 per year as a result.
Emigration remained illegal for the next two decades. All citizens seeking to leave the GDR had to apply for an exit visa. Eligibility was mostly restricted to a small number of citizens who were seen as a burden for the regime—primarily retired, sick, or disabled individuals (Major Reference Major2009).Footnote 18
The 1983 Emigration Reform
Major change to the GDR’s emigration regime did not occur until 1983. That year, a confluence of factors forced GDR leadership to liberalize the emigration regime in a meaningful way. One factor was related to the credit boycotts that many Western countries had issued against the socialist world in reaction to the repressive Polish response to the Solidarność movement. The sudden inability to obtain foreign capital brought the GDR to the brink of bankruptcy (Graf Reference Graf2020). It desperately needed foreign capital to pay for the import of consumer goods and its expansive welfare programs, cornerstones of the regime’s strategy to maintain popular acquiescence (Dale Reference Dale2005).
West Germany knew how to exploit the GDR’s vulnerability. It agreed to provide the GDR with two loans of one billion deutschmarks each so it could avoid insolvency (Graf Reference Graf2020). The loans came with several conditions, among them a liberalization of the GDR’s emigration regime. West German negotiators successfully demanded that the GDR facilitate travel and communication between both Germanys and dismantle the shooting devices at the inner German border, among other measures (Kieninger Reference Kieninger, Blumenau, Hanhimäki and Zanchetta2018).
At the same time, the Conference on Security and Cooperation in Europe (CSCE) was completing its follow-up meeting in Madrid. Although the GDR had already recognized its citizens’ right of foreign travel by joining the United Nations in 1973 and signing the Helsinki Accords of the CSCE in 1975 (Eisenfeld Reference Eisenfeld, Poppe, Eckert and Kowalczuk1995), signing the Madrid document proved highly consequential for the country’s leadership. In doing so, the GDR recognized its citizens’ right to apply for family reunification and emigration (Hanisch Reference Hanisch2012, 329): the document reflected a commitment, on paper, that the government would rule on applications for family meetings “as expeditiously as possible” in emergency cases and, in cases of family reunification, “within six months.”Footnote 19
Historians again attribute the GDR’s agreement to these terms to external pressure, particularly from the USSR. At the time, the Soviet Union was struggling to keep pace with US military expansion, with costs becoming increasingly unsustainable. Seeking a disarmament agreement (outlined in Basket I of the accord), Soviet leaders were willing to push for humanitarian concessions (detailed in Basket III) from East Germany. In the face of Soviet pressure, the GDR leadership ultimately had no choice but to comply (Hanisch Reference Hanisch2012, 264).
To meet its obligations, the GDR implemented a far-reaching reform of its emigration regime. The “order to regulate questions of family reunification and marriage between citizens of the GDR and foreigners,” published in September 1983, was the first time that the government officially recognized the option of emigration and gave it a legal basis. All East Germans with family members abroad became eligible to apply for an exit visa.
This reform triggered a dramatic increase in exit visa applications. Applications jumped from a pre-reform annual average of 10,300 to 57,600 in 1984—an increase of about 460% (figure 1). Part of this jump can be attributed to the perception that the emigration reform had made leaving the country a real possibility. A more important reason for the massive increase in applications was East German citizens’ loose interpretation of the new regulations. Internal government documents indicate that East Germans seeking to emigrate used any family tie, even to very distant relatives, to justify their applications. Although precise numbers of East Germans with such family ties are unavailable, several statistics suggest that they were very high. First, at least three million Germans had moved from East to West Germany between the end of World War II in 1945 and the construction of the Berlin Wall in 1961, creating deep social ties between both parts of the country (Burchardi and Hassan Reference Burchardi and Hassan2013). A second indicator is the numerous care packages (Westpakete) that East Germans received from the West. Although not every package was sent by immediate family members, the total number of packages was impressive: approximately 6.6 million East German households (Staatliche Zentralverwaltung für Statistik 1989, 285) received a staggering 25 million packages every year (Härtel and Kabus Reference Härtel and Kabus2000). Finally, in a representative survey of East Germans conducted in 2018, about two-thirds of respondents reported that they had had relatives or friends in the West before reunification.Footnote 20 In short, the emphasis on family reunification in the 1983 emigration reform gave a surprisingly large number of East Germans across the country a reason to apply for emigration.
Consequences of the 1983 Emigration Reform
Note. This figure reports the number of first-time exit visa applications per year (gray line [Eisenfeld Reference Eisenfeld, Poppe, Eckert and Kowalczuk1995, 202]) and annual emigration from the former GDR (black line; data collected by the author). The period after the emigration reform is emphasized.
Emigration indeed rose sharply in response to the increase in applications. As shown in figure 1, it jumped from a pre-reform annual average of 32,800 per year to 74,000 in 1984 and 56,900 in 1985.Footnote 21 That is, an additional 65,200 East Germans left the country in the two years after the reform.Footnote 22 Importantly, there was significant subnational variation in emigration rates. Whereas the cities (city-counties or Stadtkreise) of Schwerin, Stralsund, and Wismar all lost no more than 0.2% of their population to emigration in 1984, the emigration rate topped 1% of the population in cities such as Dresden, Görlitz, and Rostock. There was similar variation in rural counties (Landkreise), where emigration rates ranged from less than 0.05% (e.g., Querfurt, Grimmen, or Anklam) to more than 1% (e.g., Cottbus-Land, Freiberg, or Saalfeld).
Emigration in 1984 was not only significantly higher in volume than in the previous year, but the demographic composition of emigrants also changed significantly (table 1; see also Lueders Reference Lueders2021, 44). In the first six months of 1983, 40.5% of the 4,177 emigrants were pensioners, and 10.5% had served prison sentences. During the same period in 1984, these groups accounted for less than 10% of all emigrants. Instead, those of working-age made up three-quarters of emigrants (up from 49% the year before). Nearly half were either semi-skilled or skilled workers (up from 25%). In short, emigration in 1984 was disproportionately concentrated among demographic groups most threatening to autocratic regimes: well-educated, young professionals, especially single young men (Nordås and Davenport Reference Nordås and Davenport2013; see also figure A3b in appendix C).
Comparison of Emigrants in 1983 and 1984 (January to June of each year)
Source. BStU, ZA, ZAIG, 4531.
The emigration reform offers scholars a unique opportunity to study emigration from an otherwise closed autocracy. Although the timing of the reform can be interpreted as exogenous, the East German government retained control over the emigration process and who was allowed to leave the country. As such, the reform allows scholars to study how the regime decided who to let out and, because of the large volume of emigration, to quantify the consequences of selective emigration. Naturally, the emigration rate in a particular county was not random: it stands to reason that the regime used the reform strategically to maximize its benefit. Indeed, archival evidence suggests that the regime prioritized criminals and other political opponents in an effort to maintain social order and “lance the boil of emigration once and for all” (Major Reference Major2009, 215). Given the regime’s efforts to strategically control emigration, it is surprising that emigration appears to have had unintended consequences, especially in later years.
Data and Observable Implications
To study the consequences of emigration in the context of the emigration reform, I assembled novel data on criminal activity in the former GDR covering the period from 1976 to 1989.Footnote 23 Criminal activity constitutes a theoretically and empirically relevant outcome in a study of emigration from socialist East Germany. First, unauthorized departures were criminalized under §213 of the GDR’s criminal code. Even attempted exits were classified as criminal offenses and are therefore recorded in the crime statistics that I analyze here. Second, as discussed earlier, criminality was integral to the logic of the regime’s emigration system. The system inadvertently encouraged persistent emigration seekers to engage in behavior deemed criminal by the state and so strengthen their case. Third, crime is a substantively meaningful outcome precisely because the 1984 emigration wave was explicitly seen by East German authorities as an opportunity to get rid of “enemies, criminal elements, and other incorrigibles.”Footnote 24 Focusing on criminal activity therefore enables an empirical assessment of whether the regime’s strategy achieved its intended effect.
Crime in the Former GDR
The GDR government understood criminal activity not as innate to human beings but as a product of societal forces. The belief was that capitalism created crime and that all incentives to engage in criminal activity would disappear during the transition to a communist society (Behling Reference Behling2017). Official propaganda often pointed to low crime levels as evidence of success of the socialist reform process (Hildebrand Reference Hildebrand1985). At the same time, the East German government spent vast resources on keeping crime rates low. The regime pursued a two-pronged crime-fighting strategy. A widespread police presence and government surveillance, coupled with severe punishments for convicted criminals, were intended to prevent crimes in the first place. In addition, the regime sought to eradicate what it saw as the root causes of crime. Among other measures, employment guarantees and a comparatively generous welfare state were intended to keep economic inequality low and satisfy citizens’ economic grievances (Buchholz Reference Buchholz2016; Sensch Reference Sensch2007).
My main data source is the country’s central, machine-readable database of all crimes, the Kriminalstatistik. Created in the 1970s, the database was maintained under the auspices of the attorney general of the GDR, who collected information on all crimes from the attorneys general of the 15 districts, who themselves assembled this information from local prosecutors (Rathje et al. Reference Rathje, Erben, Teichmann and May2023). I obtained a digital copy of the data from the German Federal Archives. Covering the years 1976, 1980, 1981, 1984, 1985, and 1987–1989,Footnote 25 the dataset records crime-level information on the type of crime (according to specific paragraphs of the GDR’s criminal code), number of persons involved, basic sociodemographic information on the main offender, damage caused (if any), and the county in which the crime was prosecuted.
There are good reasons to have faith in the usefulness of the data for the purposes of this study. Most importantly, the data likely capture the criminal activity that is of particular interest. We know that the GDR regime artificially deflated its criminal statistics in an effort to demonstrate the superiority of socialism. Many low-level crimes—such as petty theft, trespassing, or slander—were not counted as crimes and thus cannot be observed in the data (Freiburg Reference Freiburg1981); this was applied uniformly across the country regardless of a county’s post-reform emigration rate. In contrast, an accurate counting of politically motivated crimes was particularly important for the GDR government’s attempts to maintain social order (and for the purposes of this article). Moreover, data on politically motivated crimes were kept secret, which further limited incentives to undercount them. Although the regime did publish data on criminal activity, the information shared was typically limited to countrywide crime counts and always excluded the most sensitive crimes (Rathje et al. Reference Rathje, Erben, Teichmann and May2023).
Categories of Crime
I used the database to create counts of different categories of crimes, which largely map onto chapters of the GDR’s criminal code. Of particular interest is the distinction between “political crimes”—that is, crimes that involve an explicit challenge of socialism and the government—and “ordinary crimes,” which include all other crimes. This distinction helps me test various observable implications of the screening and demand mechanisms, as discussed later.
Figure 2 reports the distribution of the different types of crimes in the data (see also Behling Reference Behling2017, 405–11). Most crimes are classified as “ordinary” (81.1%). They comprise nine types: crimes against state property (e.g., theft or damage); private property (e.g., theft or damage); general security (e.g., arson, illegal weapons ownership); person, life, and health (e.g., murder, homicide); the administration of justice (e.g., perjury, false accusation); freedom and dignity (e.g., rape, human trafficking); youth and family (e.g., sexual abuse of minors, refusal to pay child support); the national economy (e.g., price speculation, violation of price regulations); and a small, miscellaneous category.
Crimes by Category (Share of All Crimes)
Notes. Examples (numbers refer to those in the y axis): (1) theft, fraud, damage to property; (2) theft, fraud, intentional damage; (3) arson, violation of occupational safety rules, DUI, illegal weapons ownership; (4) murder, homicide, bodily injury; (5) disturbance of public order through refusal to work or prostitution; (6) failure to report, perjury, false accusation; (7) Illegal border crossing; (8) rape, forced into prostitution, robbery, human trafficking, libel; (9) disturbance of public order through violence, harassment, or intentional damage; (10) sexual abuse of minors, refusal to pay child support, illegal abortion, double marriage; (11) insult or slander of government or individuals involved in government business; (12) threat or actual use of violence to prevent state official from carrying out official duties; (13) violence against citizens to prevent them from carrying out official duties; (14) crimes while serving in the military; (15) violation of tariff law or price regulations, speculation; and (16) planning of wars of aggression, war propaganda, treason.
The remaining 18.9% of crimes are classified as “political.” They fall into seven categories: four are what the GDR’s criminal code refers to as “asocial behavior” (e.g., disturbance of public order through a refusal to work), “rowdyism” (e.g., disturbance of public order through violence or harassment), “resistance to state measures” (e.g., use of force to prevent state officials from carrying out official duties), and “interference with state activities” (e.g., use of force to prevent other citizens from carrying out official duties). In addition, the GDR’s criminal code punished state slander (e.g., directed at government institutions or government officials) and crimes against state sovereignty (e.g., planning of wars of aggression, treason). Finally, the “political crimes” category includes attempts to cross the border illegally: such border-related crimes comprised 5.5% of all crimes, equaling 29% of all political crimes.Footnote 26 Given my interest in emigration, I separated illegal border crossings from the other political crimes.
On the basis of these classifications, I created five dependent variables that reflect the number of all crimes, ordinary crimes, all political crimes, political crimes other than border-related crimes, and border-related crimes in each county and year. During the aggregation process, I accounted for the fact that crimes may involve more than one offender by multiplying each crime by the number of persons involved.Footnote 27
The distribution of crime types changed significantly after the 1984 emigration wave (figure 3). There was a small (9%) decline in the total number of crimes—from an average of 160,000 to an average of 147,000. This modest decline masks significant heterogeneity across crime types. Although ordinary crimes decreased by about 14%, the number of political crimes rose by more than 25%. This increase is almost entirely driven by increases in the number of border-related crimes. Whereas political crimes displayed no clear trend, border-related crimes increased by about 164%: the data record an annual average of 5,000 attempted illegal border crossings before the emigration wave and more than 13,200 thereafter.
Crimes by Year
Note. This figure reports the overall annual number of persons involved in a crime, by type of crime. It emphasizes the period after the 1984 emigration wave.
Observable Implications
My argument makes divergent predictions about the development of these various crime types in response to emigration. The screening mechanism suggests that the removal of troublemakers should result in fewer ordinary crimes. The demand mechanism, in turn, implies an increase in border-related crimes: emigration created new cross-national ties and encouraged left-behind citizens to try to leave the country. Given the severe restrictions on legal emigration, some of this increased demand for emigration should have led to more attempts to cross the GDR border illegally.
The predictions are ambiguous regarding political crimes unrelated to the border. The screening mechanism suggests a decline in the number of these crimes because those who committed political crimes were let out. The demand mechanism, by contrast, implies an increase in these crimes, because the successful emigration of some East Germans taught others that open confrontation with the regime raised the chances of obtaining an exit visa. One way to adjudicate between both arguments is to distinguish between the short- and long-run consequences of emigration. The removal of troublemakers may lead to an initial decline in political crimes, but after an increased demand for emigration arises, the number of political crimes may increase in later years.
Results
An additional 65,200 citizens were allowed to leave the GDR after the 1983 emigration reform. Archival documents suggest that the GDR used the reform to remove criminals and reduce popular demands for emigration. How, then, did criminal activity change after the reform? And was the reform successful in keeping later demands for emigration low?
Evidence from Data on Criminal Activity
Difference-in-Differences Results. My empirical strategy to assess how criminal activity changed after the emigration reform used a standard difference-in-differences approach. I regressed each of the five measures of criminal activityFootnote 28 on the interaction between a county’s emigration rate in 1984Footnote 29 and an indicator of the period after the emigration wave. For ease of interpretation, I dichotomized the emigration rate using a median split, although this coding decision does not influence my results (see appendix D.1). All models include county- and year-fixed effects. In this setup, the coefficient of interest is that on the interaction term. It reflects the estimated change in the differences in crime between counties with high versus low emigration after the emigration wave.
The results, reported in table 2, indicate that different types of crimes developed differently following the emigration wave. We observe no clear change in the volume of criminal activity (Model 1). The relevant coefficient is statistically insignificant and trivially small: it implies a decline in all crimes by no more than 0.06% relative to the mean of this variable in the pretreatment period. This insignificant coefficient masks substantial heterogeneity between ordinary and political crimes, however. Model 2 indicates that emigration was associated with a statistically significant decline in ordinary crimes. I calculate that such crimes declined by about 9.2% in high-emigration counties after the emigration reform. By contrast, model 3 indicates that political crimes increased significantly—by a staggering 43%.
Change in Crime Incidence after the Emigration Wave
Notes. Change in the incidence of different types of crime between high- vs. low-emigration counties after the emigration wave. Robust standard errors, clustered by county, in parentheses. Lower-order terms are collinear with the fixed effects and thus not reported.
*p<0.1; **p<0.05; ***p<0.01.
When distinguishing between political crimes that do not (model 4) and do (model 5) involve the border, I find that the increase documented in model 3 is entirely driven by border-related crimes. Model 4 finds no statistically discernible change in political crimes when border-related crimes are not included. According to model 5, by contrast, we find a statistically significant increase in border-related crimes by more than 200% in high-emigration counties after the reform.
My interpretation of these coefficients rests on the assumption that crime would have developed in parallel across counties with varying emigration rates in the absence of emigration. As is standard practice, I used an event study approach to support this assumption. Specifically, I interacted the indicator for high-emigration counties with separate indicators for each year, using 1984 as a baseline. The results are summarized in figure 4. They support the parallel trends assumption: the coefficients of interest are small and insignificant for ordinary and border-related crimes. Although they are not always insignificant for nonborder political crimes, no clear trend is recognizable here—a perhaps unsurprising finding in light of the statistically insignificant result for this outcome (model 4 in table 2).
Event Study Plot for the Incidence of Different Types of Crime after the 1984 Emigration Wave
Notes. Each coefficient reports the estimated annual change (compared to 1984) in the difference in the crime incidence between counties below vs. above the median emigration rate in 1984. The period following the emigration wave in 1984 is emphasized.
I also explored whether omitted variables might simultaneously influence emigration and criminal activity by probing the robustness of my findings when including interactions between the post-1984 period and the following variablesFootnote 30: (1) the vote share of the East German ruling party (SED) in the 1946 regional elections, the only free elections before the establishment of one-party rule in East Germany (Falter Reference Falter1997) and indicators of counties (2) with at least one college or university; (3) that are urban (own coding); (4) that have access to West German TV (Crabtree, Darmofal, and Kern Reference Crabtree, Darmofal and Kern2015); and (5) that border West Germany (own coding). I also control for (6) the share of the population that is young (aged 18–24), which may be particularly likely to engage in antiregime behavior (Nordås and Davenport Reference Nordås and Davenport2013), and (7) the under-3 infant mortality rate, which is widely seen as a correlate of socioeconomic status, living conditions, and poverty in emerging (DaVanzo Reference DaVanzo1988) and industrialized economies alike (Udine et al., Reference Udine, Evans, Burns, Pearson and Kaltman2021; Zucconi and Carson, Reference Zucconi and Carson1994). Infant mortality was a key metric that the GDR regime used to measure societal progress and keep track of its commitment to provide healthcare and special protections to mothers and infants, as outlined in Article 38 of the GDR Constitution (Ockel Reference Ockel1995).Footnote 31 As shown in table A6 in appendix D.1, all results remain substantively unchanged when these variables are included.
A third concern is that the decline in ordinary crimes might be driven not by the exit of troublemakers but instead by an increase in government transfers facilitated by the West German loans secured in 1983; that is, the East German government may have increased social spending strategically to lower criminal activity in places with higher emigration rates. Available data on government spending in the former GDR are inconsistent with this idea, however. As I demonstrate in appendix E.1, patterns of government spending in general and on social services in particular looked similar before and after the reform. Overall government spending grew at an average annual rate of 7.1% between 1984 and 1988, but this was below the 8.2% annual average in the five years prior to the emigration wave. Similarly, social expenditures grew by 5.2% % annually after the reform, compared to 4.3% before (own calculations using data from Staatliche Zentralverwaltung für Statistik 1989).Footnote 32
Finally, the rise in border-related crimes may be related to an increase in enforcement and a more severe crackdown on related behavior, rather than an actual change in criminal activity. Although this is theoretically plausible, three data points are inconsistent with this idea. First, there was little change in border enforcement during this period. The internal structure of border troops, which had been reorganized in the early 1970s, was left untouched until the last year of the GDR (Lapp Reference Lapp1999). Although border fortifications were continuously upgraded throughout the existence of the GDR, it is unclear that any particular change had significant consequences for enforcement.Footnote 33 Meanwhile, the number of Stasi officials involved with issues related to the border and illegal border crossings remained stable during this period (Tantzscher Reference Tantzscher2009, 98). Second, and perhaps most importantly, the idea that changes in enforcement alone—rather than changes in the underlying demand—are responsible for the uptick in border-related crimes is inconsistent with the increase in demand for emigration I document later. Finally, the number of successful illegal border crossings increased sharply after the emigration reform. According to West German statistics, their number rose from an average of 3,450 between 1976 and 1983 to an average of 5,550 between 1984 and 1988 (Ritter and Niehuss Reference Ritter and Niehuss1991, 46). According to internal East German statistics, the numbers rose from 800 to 2,300 between the same periods (Eisenfeld Reference Eisenfeld1996, 49).Footnote 34 This sharp increase is again inconsistent with the idea that changes in enforcement alone are responsible for the patterns I describe. If they were, the number of successful illegal border crossings should have stayed the same or even declined.
Short- and Long-Term Consequences of Emigration. Emigration may have opposing short- and long-term consequences on criminal activity, in particular politically motivated crimes. The screening mechanism may be dominant immediately after the emigration wave, resulting in a decline in all crimes. In later years, more East Germans may be encouraged to demand emigration, resulting in an increase particularly in political crimes.
Table 3 offers evidence that the consequences of emigration for political crimes differed by year. It reestimates how ordinary crimes (model 1), border-related crimes (model 3), and all other political crimes (model 2) changed after the emigration wave. It uses a modified indicator of the time after the emigration reform, which takes on distinct values for each year after 1984. This modification allows me to examine whether the change in crimes after the emigration wave differed by year.Footnote 35
Short- vs. Long-Term Consequences for Different Types of Crimes
Notes. This table reestimates select models from table 2 when considering each year of the posttreatment period separately. Model 1 considers ordinary crimes, model 2 considers political crimes other than border-related crimes, and model 3 considers only border-related crimes. Robust standard errors, clustered by county, are in parentheses. Lower-order terms are collinear with the fixed effects and thus not reported.
*p<0.1; **p<0.05; ***p<0.01.
The results reveal that the emigration wave was associated with an immediate drop in ordinary crimes. Model 1 suggests that such crimes declined by about 5.7% (relative to the pretreatment mean) in places with high levels of emigration in the year immediately after the emigration wave. Although the relevant coefficient drops below standard levels of statistical significance in 1987, it becomes larger in later years; in 1989, ordinary crimes were 7.1% lower than before the reform. The insignificant coefficient in 1987 may be related to social learning: East Germans may have initially believed that engaging in ordinary crimes would be sufficient to support their desire to emigrate. If true, these changed incentives to engage in ordinary crimes would counteract the immediate declines in this outcome after the initial emigration wave. However, over time, East Germans may have come to realize that border-related crimes were a more effective strategy to demonstrate their resolve.
The results for political crimes support the idea that the screening and demand mechanisms follow a temporal sequence. All political crimes declined in the year immediately after the emigration wave. Border-related crimes dropped by about 15% relative to this variable’s pretreatment mean, whereas other political crimes declined by 12.7%. The pattern was reversed two years later, however. Starting in 1987, we observe a large and statistically significant increase in border-related crimes. Although the coefficient is also positive for all other political crimes, it is not statistically significant.
Taken together, the evidence supports both the screening and demand mechanisms. In the short run, the screening mechanism appears to have dominated, whereby the identification and removal of troublemakers were associated with declines in criminal activity. The demand mechanism, in turn, appears to have played a more important role in later years. More East Germans were encouraged to seek exit and engaged in political crimes to demonstrate their resolve.
Decline in Crime-Related Damage. Emigration was also associated with changes in the severity of crimes. Table 4 assesses the damage (reported in GDR marks) of individual crimes as recorded in the data. I considered two measures: the log-transformed total amount of damage (models 1 and 2) and differences between the 43% of crimes that caused any damage and the 57% that did not (models 3 and 4). To account for the significant skew in this variable’s distribution, I estimated additional models dropping the 5% of crimes with the largest amount of damage (models 2 and 4).
Change in Average Damage Caused after the Emigration Wave
Notes. Change in the difference in the average damage of a crime (models 1 and 2) or in whether a crime created any monetary damage (models 3 and 4) between high- vs. low-emigration counties following the emigration wave. Robust standard errors, clustered by county, are in parentheses. Lower-order terms are collinear with the fixed effects and thus not reported.
*p<0.1; **p<0.05; ***p<0.01.
All models regress the outcome on the interaction between the period after 1984 and counties with above-median emigration rates, as well as the main offender’s gender, party membership, age group, and the total number of persons involved in a particular crime. I include fixed effects for the crime type (which is based on individual paragraphs in the GDR’s criminal code) as well as county and year. Of interest is the coefficient on the interaction term, which reflects the estimated change in crime severity in high-emigration counties after the emigration wave.
The results reveal that emigration was associated with reductions in the severity of crimes: average damage declined by between 8.3% (model 2;
$ p<0.01 $
) and 9.1% (model 1;
$ p<0.01 $
) in high-emigration counties after the emigration wave. Likewise, the likelihood that a crime caused any damage dropped by 1.6 percentage points (models 3 and 4;
$ p<0.01 $
), equaling 3.7% relative to this variable’s mean. These results indicate that emigration may have brought about fiscal benefits for the perennially cash-strapped GDR government.
Robustness. I probed the robustness of my results in multiple ways (see appendix D.1 for details). I verified that all results remain unchanged when accounting for baseline differences in regime support, whether a county has a university, its urban status, access to West German TV, infant mortality, and the share of the population that is young (table A6). I next varied how I identified counties with high emigration rates, replicating my main results when replacing the dichotomous indicator of counties with above-median emigration rates with the log-transformed continuous emigration rate (table A7) and when considering the difference in emigration rates between 1984 and 1979 (table A9).Footnote 36
Two additional robustness checks are related to the measurement of my dependent variable. My results remained unchanged when considering the count of crimes as an outcome (table A10). I also replicated my main analyses when scaling the outcomes by county population (table A11). Although this coding decision leaves the results regarding border-related and other political crimes unchanged, it does not replicate the finding for ordinary crimes. One potential interpretation of this result is that the decline in ordinary crimes was driven by a decrease in population. However, a definitive answer is difficult to provide because the main explanatory variable influences both the numerator and the denominator of the outcome.
Another set of robustness checks modified the sample composition. I found that my results remain substantively unchanged when dropping county-year observations with unusually large values on the dependent variable (table A12) or restricting the sample to nonurban counties (table A13) or to counties without a college or university (table A14). Finally, all relevant coefficients remained stable when replicating the analyses in samples that sequentially drop one of the GDR’s 15 districts (figure A4).
Mechanism: Demand for Emigration
My findings align with my expectations. That emigration was associated with a decline in ordinary and—at least in the short run—political crimes is consistent with the screening mechanism, whereby emigration restrictions allowed the regime to identify and remove criminals and other potential troublemakers. That emigration was associated with an increase in border-related crimes, in turn, is consistent with the demand mechanism, whereby emigration in one period can raise future demand for emigration among left-behind citizens. I now offer additional evidence for this latter mechanism.
The goal of this analysis is to link emigration after the emigration reform to demand for emigration in later years. Because county-level data on exit visa applications are unavailable, I instead exploit the fact that the GDR encouraged citizens to write Eingaben (petitions) to communicate their demands directly to the government. Estimates suggest that East Germans wrote up to one million such petitions every year; the government used the information contained in these petitions to learn about public grievances and to signal its responsiveness (Lueders Reference Lueders2022).
Although most petitions were about everyday grievances, a few were more critical of the regime. Among the latter group were petitions related to emigration and pending exit visa applications. Applicants for exit routinely wrote such petitions to demonstrate their resolve to leave (Major Reference Major2009). My analyses thus take the volume of such petitions as a proxy for demand for emigration. I use two measures, which were both generated using detailed statistics on petition volume to the GDR’s Staatsrat (“State Council”), the main executive body. One is the number of petitions specifically about exit visas. Because the data are only available for two years after the 1984 emigration wave (1987 and 1989), I also consider petitions about domestic affairs more broadly (available for 1986–89). This category includes petitions about international travel and exit visas. Both variables are scaled by 1,000 capita.
The data underscore the importance of petitioning in the exit visa application process. In 1984, for instance, my data record 16,151 emigration-related petitions to the State Council—about 28% of the 57,600 initial exit visa applications submitted that year (Eisenfeld Reference Eisenfeld, Poppe, Eckert and Kowalczuk1995). Similarly, in 1987, there were 10,914 petitions, equaling 25% of that year’s 43,200 applications. The true number of exit visa applications was likely higher, because such petitions were routinely submitted to other national and local government entities as well, which is not captured in my data.
Table 5 uses the same difference-in-differences strategy I introduced earlier to examine how demand for emigration changed after the emigration wave.Footnote 37 The results are as expected: petitions about exit visas increased by about 51.4% (model 1) relative to this variable’s mean in high-emigration counties, whereas petitions about domestic affairs increased by about 26.0% (model 2).
Increase in Demand for Future Emigration after the Emigration Wave
Notes. Change in the difference in different types of petitioning between high- vs. low-emigration counties after the emigration wave. Robust standard errors, clustered by county, are in parentheses. Lower-order terms are collinear with the fixed effects and thus not reported.
*p<0.1; **p<0.05; ***p<0.01.
My interpretation of these coefficients rests on several key assumptions. I again assume that counties with varying emigration rates trended in parallel before the emigration reform. An event study plot substantiates this assumption (see figure A5 in appendix D): counties with varying emigration rates displayed similar patterns in demand for emigration in prior years but diverged sharply after the emigration reform. Importantly, this deviation does not start until 1987—a result that aligns with the previous finding that the repercussions of emigration were not immediately noticeable.
A second assumption is that my measure of the magnitude of emigration does not simply capture other factors that may covary with both variables of interest. It is possible that emigration was higher from places with varying economic conditions or levels of baseline political support and that these factors later resulted in heightened demand for emigration. To address this concern, I reestimated the regressions presented in table 5 when accounting for the same control variables I introduced previously. As shown in table A15 in appendix D.2, my results are unchanged when accounting for baseline support for the GDR’s ruling party, counties with colleges or universities, urban counties, access to West German television, whether a county borders West Germany, as well as the infant mortality rate and the share of the population that is young.Footnote 38
A related concern is that subsequent demand for emigration may be driven by economic prospects. It is possible that demand for emigration increased in the lead-up to the collapse of the GDR because East Germans were anticipating the economic hardships they were about to experience. Whether East Germans were able to do so is doubtful, however. For example, economic perceptions before the fall of the Berlin Wall were largely uncorrelated with the economic situation afterward (Lueders Reference Lueders2024, 157). Indeed, the correlation between both measures of demand for emigration (in 1989) used in this section and two key indicators of future economic prospects is very low. It is no more than 0.21 for the share of unskilled workers (measured in 1989),Footnote 39 who would be hit particularly hard by the economic transformation after German reunification. It is the same for the unemployment rate in the first full year after German reunification (1991).Footnote 40 Given these weak associations, it is unsurprising that the inclusion of these measures leaves my main results unchanged (see table A15 in appendix D.2).
Conclusion
This article revisits the consequences of emigration in the former German Democratic Republic—a canonical example of an autocratic regime that exerted tight control over its population’s freedom of international movement. Whereas past work largely focused on the relationship between emigration and protest immediately before the GDR’s demise (Hirschman Reference Hirschman1993; Pfaff Reference Pfaff2006), I expand the scope of inquiry to an earlier period and examine an outcome that played a key role in the government’s internal considerations: criminal activity.
Exploiting a large emigration wave in 1984, I show that the consequences of emigration are more complex than previous studies have shown. In line with past work, I propose that the GDR’s restrictive emigration system allowed the regime to identify troublemakers (Michel, Miller, and Peters Reference Michel, Miller and Peters2023). In that sense, the emigration system served as a safety valve, leading to the removal of those most likely to cause trouble in the future (Hirschman Reference Hirschman1993). Yet, I also identify a second, countervailing effect: letting some citizens out can fuel future public demand for emigration among left-behind citizens. Because the emigration system incentivized emigration seekers to express their demands in public acts of opposition, the result was an increase in crime.
This article presents empirical evidence from an in-depth investigation of the emigration system of the former GDR. Although my focus is on one regime in particular, its conclusions apply beyond this case. The key innovation described in the article is how the exit visa system helps address the “dictator’s dilemma”—that is, the dictator’s uncertainty about the population’s true level of support (Wintrobe Reference Wintrobe1998)—by encouraging potential troublemakers to reveal themselves. This discussion speaks to closed autocracies more broadly. These regimes often restrict the freedom of international movement of some if not all of their citizens. Emigration seekers routinely have to overcome numerous hurdles on their way out. Although many paradigmatic examples of such regimes—such as Nasser’s Egypt or the former Soviet Union—have since liberalized or broken down, emigration restrictions are not a phenomena of the past. Indeed, Michel, Miller, and Peters (Reference Michel, Miller and Peters2023) calculate that 30% to 45% of all contemporary autocracies require citizens to apply for an exit visa to leave.
An open question is whether the GDR regime had intended to create the demand mechanism I describe. Although the available archival evidence does not allow for a definitive answer, the materials consulted for this study suggest that the regime sought to minimize rather than encourage emigration: as discussed earlier, the goal of emigration policy was to permanently reduce demand for emigration. Moreover, the regime appeared increasingly concerned about its inability to keep public demand for emigration low, and it was acutely aware that the emigration question, if not solved, would pose a real danger to its stability.
In this context, it may seem surprising that the regime did not change course once the demand effect became apparent. Although East German officials were aware that each approved exit visa application would generate more public pressure, it did not change the system. One potential explanation is the regime’s institutional inflexibility. Hirschman (Reference Hirschman1993, 197) writes, “The managers of the GDR had inured themselves to it … and were in general an insensitive and inflexible bunch.”
This research points to the challenges of managing emigration flows over long periods of time. The example of the GDR illustrates how selective emigration may temporarily alleviate pressure on the regime, but it cannot suppress public opposition forever. The ultimate fate of socialist rule in East Germany is a powerful reminder that restrictions on the freedom of movement may in the short run sustain autocratic rule but may very well contribute to its demise in later years.
Supplementary material
To view supplementary material for this article, please visit http://doi.org/10.1017/S1537592725102144.
Acknowledgments
I am grateful to the editors, four anonymous reviewers, as well as to Aala Abdelgadir, Ala’ Alrababa’h, Lisa Blaydes, Brett L. Carter, Alberto Díaz-Cayeros, Vicky Fouka, Anna Grzymala-Busse, Carl Gustafson, Raymond Ha, Jens Hainmueller, Carl Levan, Elisabeth van Lieshout, Beatriz Magaloni, Will Marble, Rachel Myrick, Maggie Peters, Jonathan Rodden, Kathryn Stoner, Matt Tyler, and participants at the Berkeley European Politics Working Group in February 2020, APSA 2020, MPSA 2021, and the 45th annual Stanford-Berkeley conference on Russian, East European, and Eurasian Studies for helpful comments and suggestions. This research received generous financial support from the Europe Center, the Center for Philanthropy and Civil Society, the Graduate Research Opportunity Fund, and the Freeman Spogli Institute and the Shultz Fellowship for Economic Policy—all at Stanford University. I further thank Iris Haik for excellent research assistance and Patrick Munene and his team at Hewani Data for their help with data digitization.
Open access
