Migration aspirations in relation to border closures, employment opportunities and risk-taking attitudes: lessons from an online survey experiment

ABSTRACT

This article investigates the effect of structural and individual factors on migration aspirations in a secondary migration context. Through an online survey experiment conducted with Syrian migrants (N = 551) living in Turkey, we unpack factors explaining aspirations to stay and move onward from the current country of residence. The findings indicate that open borders alone do not compel migrants to move onward. Instead, employment opportunities in their current residence play a crucial role in shaping aspirations to stay put. Moreover, individuals inclined to take risks are more likely to migrate, even when strict border controls are in place. By highlighting the question of what motivates migrants to stay as well as to move onward, this research emphasizes individual differences in forming migration aspirations and contributes to migration aspirations literature in the secondary migration context.

KEYWORDS:

• Migration aspirations
• risk-taking
• border closures
• online survey experiment
• online survey experiment

Introduction

On 14 June 2023, eight years after so-called 2015 ‘refugee crisis’, a fishing trawler carrying 750 migrants from Italy to Greece sank in the Mediterranean Sea. Only 104 survived, making it one of the deadliest incidents of its kind in recent times. In the aftermath, Greek border guards were accused of intentional involvement in the trawler’s sinking, leading to significant criticism and controversy. The incident was much more than a tragic border spectacle: it has multiple layers, and offers insights into how migration policymaking affects migrant journeys that are full of risks.

Despite being one of the deadliest incidents, the June 14th incident is not unique or unusual in the Mediterranean. The EU’s migration and border policies have turned the Mediterranean route into a space of violence, death, drowning of refugee boats, and pushbacks (İşleyen and Qadim 2023; Tazzioli 2016). Meanwhile, funding schemes in the context of migration governance construct host countries – usually the ones neighboring conflict-torn regions – as spaces of opportunities to avoid migrants moving onward to developed Western countries (Müller-Funk, Üstübici, and Belloni 2023). Albeit limited, and insufficient, it is assumed that providing opportunities such as employment to migrants would make them integrate in the host countries and eventually increase their aspiration to stay put. In a way, migration policies oscillate between increasing border controls along migration routes, and providing migrants with opportunities in order to keep them where they are (Kipp and Koch 2018). Yet, some migrants who are more prone to take risks still embark on dangerous journeys, usually as a way of showing resistance towards border controls (Kiriscioglu and Ustubici 2023; Maâ 2023; Schapendonk, Bolay, and Dahinden 2021).

In this context, the following open question can be posed: To what extent are migration aspirations subject to exogenous shocks, such as border closures, or opportunities such as employment, in the context of secondary migration? While tackling this important empirical puzzle, our conceptual framework also aims to explore whether migrants’ risk-taking attitudes influence their aspiration to embark on dangerous journeys.

This research contributes to the migration literature by focusing on secondary migration aspirations, where individuals had to flee their home country for another and consider further migration from the first host country. To answer what motivates migration aspirations in such constrained mobility contexts, we test the effect of structural and individual level factors with a between-group online survey experiment conducted with Syrian migrants in Turkey.¹ We include border closures, employment opportunities, and subjective risk-taking attitudes as factors affecting migration aspirations. Survey experiments have so far been used to investigate migration decision-making outside the context of forced migration (Ewers and Shockley 2018; McKenzie and Yang 2012; Petzold 2017). Hence, using survey experiments to study migration aspirations in a forced displacement context is a novel approach. Recently, only a few experimental studies have investigated migration aspirations using the framework of push and pull factors between destination and origin countries (Bah and Batista 2018; Detlefsen, Heidland, and Schneiderheinze 2022; Hager 2021). We contribute to this research line by employing an experimental setting that enables us to manipulate the structural factors shaping onward migration aspirations in forced migration contexts. Plus, our approach allows us to gain insights into the variations in individual personality traits.

Existing research explores factors shaping migration aspirations mainly at the structural and individual levels, within the context of either the home or destination country. Structural-level-focused studies examine the impact of border externalization policies on migration management (Cobarrubias et al. 2023; Geiger and Pécoud 2010), while individual-focused research investigates perceptions of individuals, life cycles, and personal attributes (Alpes, 2014; Boccagni 2017; Kuschminder, de Bresser, and Siegel 2015; Syed Zwick 2022). A few studies examine migration aspirations at a combination of individual and structural levels (Helbling and Morgenstern 2023). Moreover, the majority of research on migration aspirations, with a few exceptions (Hager 2021; Kiriscioglu and Ustubici 2023; Müller Funk 2019), has concentrated on either the home or destination country, largely overlooking aspirations to stay in the current host country. This research aims to address this gap, generating insights into the onward migration decision-making patterns in forced displacement contexts.

Our findings indicate that implementing open border policies does not in itself compel migrants to move onwards. Instead, migrants consider employment opportunities in their current place of residence, even when borders are open for passage (i.e. to such welfare states as the EU countries). However, migrants’ aspirations are not only shaped by these policies. Migrants are neither victims, nor passive pawns without agency (De Haas 2021). Thus, our study also provides a room for human agency by analyzing differences in individual characteristics in forming migration aspirations. Our findings show that individuals with a higher inclination for risk-taking are more likely to move onward, even when confronted with strict border controls. Incorporating risk-taking attitudes as individual characteristics brings a novel approach to migration literature, showing that not every migrant can be deterred by strict border controls.

Turkey is a suitable case to explore secondary migration aspirations. Besides being a host country for around 3.1 million registered Syrian migrants, Turkey poses geographical significance as one of the European Union’s external borders.² After the EU-Turkey Statement in 2016, Turkey has been constructed as a space to deter migrants from entering the EU’s borders and equipped with, albeit inadequate, policy interventions which aim to foster migrants’ aspirations to stay put by providing opportunities for local integration (Üstübici, Kirişçioğlu, and Elçi 2021). Based on a convenient sample of Syrians living in Turkey (N = 551), we asked the participants to assess four experimental vignettes about a hypothetical person and a control group. We crossed two vignettes involving border closures and employment opportunities to evaluate the conditions that would shift migration aspirations.

The article is organized as follows. We first discuss the existing literature on migration aspirations, particularly explaining the relevance of border closures and externalization policies, employment opportunities, and risk-taking attitudes in migration journeys by presenting our hypotheses. Plus, we briefly present the social, political and legal contexts that Syrians are subjected to in Turkey that inspired the design of the survey experiment. Second, we explain our data and experiment design. The third section lays out the results using regression models. Finally, we discuss the policy implications of the interaction between border externalization policies and employment opportunities, and the role of risk-taking attitudes in migration aspirations literature.

Theoretical framework

We define migration aspirations as the perspectives of migrants on their future mobility. The aspirations – abilities framework combines structural and individual factors to understand how aspirations and abilities are formed (Carling 2002, 2014; Carling and Collins 2018; Carling and Schewel 2018). To unpack how aspirations are formed, the framework analyzes communal social, economic, and political contexts at the macro level, and individual characteristics at the micro level, in determining who wants to migrate and who wants to stay. On the individual level, the framework focuses on the characteristics that affect how people cope with restrictive migration policies. Against this complex theoretical model, this paper selects two aspects at the structural level (border externalization policies and employment opportunities), and focuses on risk-taking attitudes at the individual level.

Structural factors

Border closures and externalization policies

The externalization of migration management has become a significant concern for various countries, including the U.S., Australia, the member states of the European Union, and, more recently, countries in the Global South (Adamson and Tsourapas 2019). Externalization is a broader term and refers to the activities undertaken by states, directly or indirectly, outside their territorial boundaries to control borders (Cobarrubias et al. 2023).³ Notably, the states in the Global North have developed strategies to prevent the entry of undesirable migrants at their borders (Hyndman and Mountz 2008). Scholars have explored how this unwanted migration influences the processes of border externalization in receiving states (Stock, Üstübici, and Schultz 2019; Tsourapas 2019), focusing on the practices of state actors (Côté-Boucher et al. 2014; El Qadim 2017) at the discursive and practice levels (İşleyen, 2018a, 2018b; Karadağ, 2019). While the literature has debated the effectiveness of these border externalization practices, the impact of such policies on migrants’ perceptions has received less attention (Hager 2021). This study aims to contribute to this particular research trajectory.

Common sense would say that open borders would increase migration. While open borders may initially increase migration from less developed to developed regions of the world, evidence shows that they also bolster circulation and return (de Haas, Vezzoli, and Villares-Varela 2019; Flahaux and Vezzoli 2017; Vezzoli 2021). Second, closed borders are assumed to be stopping migrants without documents. Some scholars show that border enforcement policies decrease the number of migrants crossing the borders (Alden 2017; Sorensen and Carrion-Flores 2007). Despite the decreasing numbers, border closures fail to deter migrants from taking risks (Dávila, Pagán, and Soydemir 2002; Gathmann 2008; Massey, Pren, and Durand 2016). Rather than stopping migrants, closed borders direct them to pursue alternative, more dangerous channels (Vezzoli 2021). In a way, border externalization policies create a self-perpetuating cycle, forcing migrants to take risks at borders rather than deterring them (Massey et al. 2016).

Due to challenging living conditions, some Syrians residing in Turkey are inclined to consider onward migration, particularly towards Europe. Being deprived a full-fledged refugee status without a prospect of local integration (Baban, Ilcan, and Rygiel, 2017), experiencing economic vulnerability compared to natives (Elçi, Kirisçioglu, and Üstübici 2021), and facing limited access to protection (Karadağ and Üstübici 2021) pose significant challenges for Syrians living in Turkey. Recent strict visa policies and closed borders limit their options for legal migration channels (Welfens and Bekyol 2021), and direct them to take high risks in undocumented migration (Kiriscioglu and Ustubici 2023).

When Syrian migrants first arrived in Turkey, many moved on to seek asylum in EU countries. In 2015, around 1.3 million migrants fled Turkey. The EU and its member states responded to this influx by closing their borders under the discourse of ‘the refugee crisis’.⁴ In response, the EU-Turkey Joint Action Plan (November 2015) and the EU-Turkey Statement (March 2016) were implemented, representing the EU’s efforts to externalize its migration and border policies (İşleyen, 2018a; Üstübici 2019). Additionally, Turkey was promised financial support by the EU, under the Facility for Refugees in Turkey (FRIT), to be spent on the protection and integration of displaced communities, especially Syrians. The design of these policy interventions reflects the rationale of the current international asylum regime, perceiving the secondary mobility as politically undesirable. Although the legal basis and effectiveness of these migration and border policies are widely debated (Aydın Düzgit, Keyman, and Biehl Öztuzcu 2019; Danış 2021; Pries and Zülfikar Savci 2023; Schiefer, Düvell, and Sağıroğlu 2023; Tekin 2022; Ulusoy and Battjes 2017), to our knowledge, there is no study measuring the effectiveness of these policies and the validity of their assumptions in the context of Turkey using survey experiments.

Evidence suggests that aspirations to move onward from Turkey persist (Kuschminder et al. 2019; Üstübici and Elçi 2022), despite the strict border controls initiated with the EU-Turkey Statement of March 2016. According to the latest statistics from the Presidency of Migration Management (PMM) (2023), the number of migrants apprehended at the borders increased after the so-called refugee crisis in 2015. Despite the EU’s efforts to externalize its migration and border policies with the EU-Turkey Statement, border crossings continued. The number of apprehended individuals tripled between 2015 and 2019, from 146,485 to 454,662. Although the number of apprehended individuals declined during the pandemic to 122,302, in 2022 it reached a second peak of 285,027. The majority of apprehended migrants are Afghans and Syrians but also include Palestinians and Iraqis. Frontex (European Border and Coast Guard Agency) (2023) data also show a similar trend for the Eastern Mediterranean route for undocumented border crossings. The total number of people who have lost their lives in the Mediterranean Sea since 2014 has reached almost 25,000 (Sunderland 2022). In short, people who face dire living conditions may feel compelled to continue their journey despite the risks (Kiriscioglu and Ustubici 2023). Hence, we expect that:

Hypothesis 1: Decreasing (increasing) risks at the borders and welcoming (exclusionary) policies in wealthier countries motivate aspirations to move onward from (stay in) the current country of residence.

Employment opportunities

The second line of research focuses on the role of employment opportunities in migration aspirations. Although the effect of employment opportunities on migration aspirations is commonly measured from the origin-host country perspective, i.e. employment opportunities in host countries (Becerra et al. 2010; Chindarkar 2014; Groenewold, de Bruijn, and Bilsborrow 2012; Roth and Hartnett 2018), the role of employment opportunities in the secondary migration context is less studied (Sahin Mencutek and Nashwan 2021).

Previous research shows that economic concerns motivate onward migration towards countries with better employment opportunities (Collyer 2006; Collyer, Düvell, and de Haas 2012; Dimitriadis 2021; Esteves, Fonseca, and Malheiros 2018; Hager 2021; Kuschminder 2018; Van Hear 2006). In the secondary migration context, satisfaction with employment opportunities fosters aspiration to stay in the current place rather than move onwards. Some other research, however, shows that regular income and remittance capabilities can provide migrants with the opportunity to make savings for onward migration (Donini, Monsutti, and Scalettaris 2016), especially in countries where migrants do not have alternative channels to smuggling. Shortly, financial security is a strong determinant in forming migration aspirations. The direction of migration aspirations, based on employment opportunities, is still contextual: while for some, financial means may enhance capabilities to move onwards, for others, it may imply improving one’s livelihood, hence fostering aspirations to stay put (Mata-Codesal 2018).

Formal employment opportunities for Syrian migrants in Turkey are limited due to policies that restrict their integration into the labor market. For instance, the Regulation on Work Permits of Foreigners Under Temporary Protection launched in January 2016, based on the EU-Turkey Joint Action Plan, allowed Syrians to access work permit procedures through their employers, although implementation has been somewhat restricted. Of almost 2 million working-age Syrians, only 91,500 individuals were granted work permits in Turkey.⁵ As a result, most working-age Syrians have ended up in insecure or underemployed positions in the informal sector (Bélanger and Saracoglu 2020; Düvell, Schiefer, Sağıroğlu, and Mann 2021). The lack of or limited employment opportunities in Turkey poses a significant challenge for migrants who aspire to stay in Turkey but feel obliged to move onwards. In this regard, we hypothesize that:

Hypothesis 2: Legal (precarious) employment opportunities foster (hinder) aspirations to stay in the current country of residence.

Individual factors

Risk-taking attitudes

Even when border crossing is ‘a matter of life and death’ (Walters 2011, 138), some migrants aspire to migrate despite the risks involved (Kaytaz 2016; Strasser and Tibet 2019; Topak 2014). Research shows that incorporating risk-taking attitudes into migration aspirations research could explain why individuals aspire to move onwards, stay, or return (Kiriscioglu and Ustubici 2023; Williams and Baláž 2014). If individual differences in risk-taking attitudes significantly shape migration aspirations, then structural conditions, such as border externalization policies, would not impact all individuals equally. Border externalization policies are predominantly based on the assumption that being aware of risks on the route would deter migrants. But this assumption that migrants lack knowledge fails to account for individual differences in risk-taking propensity, which can vary widely among migrants. Therefore, the effectiveness of these policies in deterring migration cannot be assumed to be universal and must be understood within the context of differences in individuals’ characteristics.

Migrants who venture the risks of borders are neither incognizant of the risks involved in their journey nor ‘irrational’ decision-makers (Hernández-Carretero and Carling 2012; Pagogna and Sakdapolrak 2023). Conversely, migrants are often well-informed about the dangers en route (Belloni 2016). With this information, migrants anticipate the risks at each step of their migration journeys and form their aspirations accordingly (Kiriscioglu and Ustubici 2023). These findings challenge the notion that migrants are passive victims entirely reliant on external constraints (De Haas 2021). Instead, migrants exhibit agency, and their attitudes toward risk-taking are highly individualized. In this regard, we hypothesize that:

Hypothesis 3: High (Low) tendency to take risks decreases (increases) the effects of structural barriers on the aspiration to move onwards (stay put).

In light of this background, we developed hypothetical vignettes for a survey experiment to echo a real-life policy environment marked by border closures and the government’s selective provision of legal employment opportunities. Recent border and migration policy developments in Turkey, explained above, make the latter a suitable case to test these hypotheses on the impact of border controls and employment incentives on migration aspirations.

Methods and data collection

We conducted an online survey from 18 November to 31 December 2020, utilizing Facebook and Instagram advertisements to recruit participants through the social media accounts of MireKoc research center at Koc University (N = 551).⁶ These ads invited individuals to participate in a survey regarding their migration experiences in Turkey. To ensure privacy and anonymity, we collected data using the Qualtrics survey tool and randomly assigned respondents to different experimental groups. The survey and its advertisements were presented in Arabic and were geographically limited to Turkey. Furthermore, we targeted Syrian individuals through online advertisements, inviting them to participate. All participants were provided with an informed consent form before initiating the survey. Only those who granted their consent were given access to the survey questions. Respondents who only partially completed the survey, those who later indicate not having Syrian nationality, and those who failed attention checks were excluded from the analysis. On average, participants took 31 minutes (standard deviation = 42) to complete the survey.

We chose not to offer incentives to recruit participants for several reasons. Firstly, while incentives are commonly used to increase response rates in online surveys, our pilot surveys did not indicate any difficulties in obtaining responses that would necessitate the use of incentives. Secondly, research suggests that incentives can introduce measurement errors, as respondents may rush through the survey to obtain the incentive quickly (Elliott and Valliant 2017). Additionally, offering incentives may raise ethical concerns, particularly in the context of our study in Turkey, where Syrians often live in precarious conditions and rely on aid for their livelihoods. Providing a survey with a prize could be perceived as misleading and may create a sense of dependency on aid among respondents.

The convenience sampling technique employed in our study had certain constraints due to our reliance on Facebook and Instagram advertisements for participant recruitment. As a result, our sample may exhibit bias towards individuals who are more educated, literate, and financially affluent, potentially overlooking illiterate, elderly, and economically disadvantaged Syrian refugees. To avoid additional bias, we strictly relied on social media ads and did not share the link to the survey in other platforms. Furthermore, since the survey was conducted online, possession of a smartphone or computer and access to Facebook and Instagram accounts were prerequisites for participation. Consequently, our sample is limited to Syrian refugees who have access to these devices and online platforms.

Despite the inherent challenges associated with sampling, recent research suggests that online platforms can be valuable tools for surveying hard-to-reach populations, including immigrants and refugees (Elçi et al. 2021; Ersanilli and van der Gaag 2020; Pötzschke and Weiß 2021). Additionally, without comprehensive knowledge of the total population of Syrian refugees in Turkey, it is difficult to ascertain whether an in-person survey conducted in a specific neighborhood densely populated by migrants would be more accurate and representative than an online survey. It is important to note that during the COVID-19 pandemic, online surveys provided researchers with a safer means of collecting data. Furthermore, compared to face-to-face and phone surveys, gathering data through social media advertisements proves more cost-effective, time-efficient, user-friendly, and secure.

Experiment design

We employed a between-group experimental design to examine the impacts of border closures and employment conditions. Participants were randomly assigned to one of five different groups (see Table 1). Next, they were introduced to the following statement: ‘Now, you are going to read a story about Ahmad and his family. Please read the story and then answer the question’. Those in the treatment groups were exposed to one of the four distinct experimental vignettes (T1, T2, T3, or T4), each depicting a hypothetical individual, while participants in the control group were presented only with a baseline scenario. Table 1 summarizes the experimental design, illustrating the various combinations of border closures and employment opportunities manipulated to create favorable and unfavorable conditions for migration recommendations.

Table 1. Experimental vignettes.

No treatment	Control	Ahmad is a single man living in a shared flat in Istanbul and working in construction. His parents are in Syria, and their livelihood depends on the money he sends them.
Treatments for Open Borders	Treatment 1 (T1)	Unemployed	Control + Ahmad could not find work in the last two weeks because of the economic crisis. One day, he hears from a friend that the journey from Turkey to Europe is getting less risky because borders are now open, and countries such as Germany will accept more refugees.
	Treatment 2 (T2)	Employed	Control + Ahmad’s boss offers him a permanent job where he can work with a work permit and social security. One day, he hears from a friend that the journey from Turkey to Europe is getting less risky because borders are now open, and countries such as Germany will accept more refugees.
Treatments for Border closures	Treatment 3 (T3)	Unemployed	Control + Ahmad could not find work in the last two weeks because of the economic crisis. He hears from friends that the journey to Europe is now more dangerous than ever. Smugglers may demand extra money, or the boats may sink. Coastguards can also be cruel to migrants.
	Treatment 4 (T4)	Employed	Control + Ahmad’s boss offers him a permanent job where he can work with a work permit and social security. He hears from friends that the journey to Europe is now more dangerous than ever. Smugglers may demand extra money, or the boats may sink. Coastguards can also be cruel to migrants.

When designing the vignettes, we constructed a hypothetical person to uphold ethical considerations. This approach was chosen to avoid disseminating misinformation within refugee communities and networks, as direct questions and factual scenarios could potentially lead to speculation about the situation at the borders. Our primary concern was to prioritize the safety and well-being of the participants. Hence, opting for a hypothetical person was deemed the best option. In addition, we aimed to decrease variation in the stimulus, which would be high if, for example, we primed participants to think about their current situation.

Our objective was to examine the impact of border closures and employment opportunities on migration aspirations. To achieve this, we presented Ahmad’s situation to all groups. We sought to determine whether the border and employment conditions differed from the control group. We chose Ahmad as a familiar figure among the Syrian refugee population in Turkey, as they often work in precarious jobs. Additionally, we focused on a single male migrant rather than a migrant with a family, as the presence of a family can complicate migration decision-making (Dubow and Kuschminder 2021; Kiriscioglu and Ustubici 2023). In other words, by constructing a control group where no particular policy intervention is mentioned, we aimed to understand the respondents’ migration aspirations when the conditions of border and employment were not explicit. Then, we aimed to test whether the changing conditions of borders and employment opportunities make any difference in the direction of migration aspiration. Methodologically, we sought a baseline group to detect variations from the control group and significant differences across treatments, as experimental interventions affect aspirations (Gaines, Kuklinski, and Quirk 2007; McDermott 2002).

We checked the randomization and distribution of respondents in the experimental groups according to their demographic characteristics, whether they receive the Emergency Social Safety Net (ESSN) aid⁷, and their level of wealth. The Wald test after the multinomial logistic regression received an insignificant result (X²(20) = 25.64, p = 0.178), which means randomization is successful (see Table 2 for the distribution of demographic and ESSN variables). Hence, we do not need to use control variables.

Table 2. Distribution of respondents per experiment group.

	Control	Unemployment & Open borders	Employment & Open borders	Unemployment & Border Closures	Employment & Border Closures	All groups
Age	28.61 (178)	30.24 (97)	27.77 (92)	30.85 (87)	28.90 (87)	29.17 (541)
Male	0.60 (182)	0.61 (98)	0.50 (92)	0.47 (88)	0.54 (89)	0.55 (549)
Education	0.85 (181)	0.83 (98)	0.86 (92)	0.82 (88)	0.91 (89)	0.85 (548)
Wealth	0.46 (183)	0.45 (99)	0.44 (92)	0.42 (88)	0.42 (89)	0.44 (551)
ESSN	0.28 (183)	0.30 (99)	0.34 (92)	0.34 (88)	0.39 (89)	0.32 (551)

Note: The number of observations is in parentheses.

Variables

Following the presentation of the vignettes, we asked the following question as our dependent variable: ‘If you were Ahmad, would you be willing to stay in Turkey, or move on to another country, or return to Syria?’ Descriptive analysis of the responses reveals that 53.53% of respondents recommended staying in Turkey, 44.29% suggested moving to another country, and only 2.08% selected the option to return. While the return option was included in the question for theoretical reasons (Carling 2019), we did not specifically prime for it in the experimental design. As a result, we generated three dependent variables. First, our main dependent variable is binary, excluding return responses from the analysis. Second, as a robustness check, we generated another binary dependent variable in which we coded both move and return as 1 and 0 otherwise. Third, we generated a categorical dependent variable. We coded move responses as 1, stay responses as 2, and return responses as 3 to conduct a multinomial logistic regression. The results do not show a striking difference between dependent variables. To assess our assumptions, we conducted a logistic regression analysis using our main dependent variable and dropping the return responses (See Online Appendix for robustness checks and other information about variables in the randomization checks).

We considered subjective risk-taking attitude as a moderating factor for the effects of border closures and employment opportunities. Risk-taking attitudes are fundamental drivers of human decision-making processes, influencing individuals’ perceptions of risk, their willingness to take risks, and their overall behavioral responses to uncertain situations. In the context of our study, we argue that individuals’ advice in a hypothetical migration scenario reflects their risk-taking attitudes, as these attitudes shape how individuals assess and respond to constraints. By examining the relationship between risk-taking attitudes and advice-giving behavior, we aim to elucidate the underlying mechanisms forming individuals’ aspirations. This analysis contributes to a deeper understanding of the factors that influence migration aspirations in real-world scenarios.

We operationalized subjective risk-taking attitudes using a general question that captured individuals’ overall tendency towards risk-taking: ‘How do you see yourself in general? Are you generally a person who is fully prepared to take risks, or do you try to avoid taking risks in your life?’ Respondents provided their responses using a Likert scale ranging from ‘not willing to take risks’ (1) to ‘extremely willing to take risks’ (5). The average score was 2.63, with a standard deviation of 1.16, reflecting the participants’ self-perceived inclination towards risk-taking behavior. We opted for a self-reported assessment of risk for two primary reasons. Firstly, risk-taking attitudes are known to be influenced by individuals’ personal life experiences (Barnett and Breakwell 2001), making self-perceptions theoretically relevant for measuring risk-taking. Secondly, subjective measurement of risk attitudes is considered a valuable overall measure, as it is presumed to capture a stable risk preference capable of predicting risk-taking behavior across different domains (Charness, Gneezy, and Imas 2013).

Self-reported measurements of risk are often criticized for overlooking domain-specific differences in risk-taking attitudes. For example, an individual might be more inclined to take risks in financial matters but less so in health-related issues. To address this concern, Weber, Blais, and Betz (2002) developed the DOSPERT scale, which aims to capture variations in risk-taking preferences across different domains. However, due to the length of our survey and to avoid cognitive fatigue among participants, we opted for a simpler elicitation question. Therefore, we utilized a single, simplified self-reported measurement, which is also easier for participants to comprehend (Charness et al. 2013).

Results and discussion: the interplay between border closures and employment opportunities, and the mediating effect of risk-taking attitudes

In this section, we use evidence from the survey experiment to answer the following two questions: To what extent are migration aspirations subject to exogenous shocks, such as increasing or decreasing risks at the borders? And to what extent are they shaped by opportunities such as employment? We conducted two different logistic regressions to test the effects of treatments, including moderation with subjective risk-taking attitudes.

Table 3 illustrates that all treatments significantly differ from the control group except for the first treatment. All other three treatments significantly motivate aspiration to stay in the current country of residence. (see Figure 1).⁸ Moreover, the margins plot illustrates that respondents in the unemployed and open borders vignette are more likely to move on to a third country than stay in Turkey compared to other treatment groups. In particular, a significant difference between ‘Unemployment and open borders’ and ‘Employment and open borders’ indicates that employment is crucial to staying, even when borders are open.

Figure 1. Average marginal effects of treatments on move aspiration.

Table 3. Logistic regression results.

	Model 1	Model 2
	Aspiration to move	Aspiration to move
Reference: control
Unemployment & Open borders	0.895	0.625
	[0.539,1.487]	[0.169,2.320]
Employed & Open borders	0.349***	0.264*
	[0.206,0.588]	[0.079,0.889]
Unemployment & Border closures	0.117***	0.0129***
	[0.062,0.224]	[0.002,0.095]
Employed & Border closures	0.0785***	0.0363**
	[0.038,0.162]	[0.004,0.372]
Subjective Risk-taking Attitude		0.841
		[0.656,1.079]
Reference: Control * Subjective Risk-taking Attitude
Unemployment & Open borders * Subjective Risk-taking Attitude		1.151
		[0.730,1.816]
Employment & Open borders * Subjective Risk-taking Attitude		1.104
		[0.731,1.666]
Unemployment & Border closures * Subjective Risk-taking Attitude		2.218*
		[1.190,4.133]
Employment & Border closures * Subjective Risk-taking Attitude		1.322
		[0.614,2.847]
N	541	540
pseudo R^2	0.148	0.161

Odds Ratios; 95% confidence intervals in brackets * p < 0.05, ** p < 0.01, *** p < 0.001.

Policymakers assume that border externalization policies deter migrants from moving onward. This assumption on the effectiveness of deterrent migration policies overemphasizes the significance of border closures. Moreover, the emphasis on border closures overlooks the effect of opportunities provided for migrants. At first sight, our results indicate that migrants recommend staying in the country of residence when borders are closed, and countries do not implement welcoming policies for migrants (see Figure 1). However, what is surprising in our findings is that opening the borders does not increase the probability of recommending moving onward. Contrary to the expectations of policymakers, open border policies and welcoming policies of the destination countries are not the main drivers for migrants to recommend moving onward. In this regard, based on the interrelation between border closures and employment opportunities, we explain why the assumption that border enforcement policies deter migrants is wrong.

First, our experimental design confirms that border controls should not be considered separate from employment opportunities provided to migrants. As Figure 1 shows, employment opportunities significantly increase migrants’ probability of staying in the current country, regardless of the conditions of the borders. In other words, when migrants are provided with employment opportunities, staying in the current country is seen as a better option than moving onwards to Europe, even when the dangers on the route are minimal. In this regard, we reject Hypothesis 1, indicating that decreasing dangers at the borders and welcoming refugee policies in European countries motivate aspirations to move from the current country of residence. This finding has two implications regarding migration aspirations based on opportunities and obstacles. First, obstacles at borders per se do not determine migration aspirations. Plus, risks of border crossings per se cannot predict change in migration decision-making. Second, this finding provides counterevidence to policymakers’ assumption that all migrants will eventually aspire to move onwards when borders are open, and only stricter border policies can stop migration.

Second, even when borders are open, and countries are welcoming, opportunities can amplify the aspiration to stay in the current country. For instance, when comparing the first and second treatments, we observe that employment opportunities strongly decrease the probability of recommending moving onward (see Figure 1). While the probability of moving onward is more than 60% in the first treatment [unemployment, open border], it decreases to 37% with an employment opportunity. If we imagine a single male working in odd jobs in Turkey, an employment opportunity including a secured work permit would significantly reinforce aspiration to stay. Thus, the findings support Hypothesis 2, indicating that employment opportunities motivate aspirations to stay in the current country of residence. We argue that relatively open border policies for migrants do not necessarily feed the aspiration to move. The employment opportunities generate a significant difference between aspirations to stay and move, and are more determinant in forming migration aspirations than border externalization policies.

We conducted the second logistic regression by taking risk-taking attitudes as a moderator (see Model 2). Figure 2 shows the predicted probability of the move aspiration when there are no employment opportunities in the current place and when borders are closed, moderated by subjective risk-taking. We assume that these are the conditions under which only risk-taking migrants would recommend moving onward. Hence, the results show that recommending moving onward significantly increases depending on to what extent participants are risk-avoidant or risk-taking. As risk-taking attitudes increases, the probability of moving to another country increases from 8% to 40% when borders are closed, and our hypothetical person is unemployed. Thus, we can confirm Hypothesis 3, namely that as the tendency to take risks increases, the impact of border closures on the aspiration to move onward decreases.

Figure 2. Predicted probability of the ‘Move Aspiration’ of ‘Unemployment and Border Closures’ vignette moderated by subjective risk-taking.

The lack of employment opportunities prime plays a significant role in understanding the aspiration to move onward when the borders are risky to cross. Respondents provided with the unemployment and border closures prime are significantly more likely to recommend moving onward when they are more willing to take risks in general (see Figure 2). This finding implies that strict border policies may discourage some migrants from onward migration but do not deter every migrant on the move in the same manner. We can explain this variation by personal attitudes towards risk-taking. If we imagine an unemployed risk-taking migrant whose family is dependent on the money they send, border control policies would not be deterring. The lack of employment opportunities in the country of residence could be highly unacceptable for risk-taking migrants who aspire to move onward even when border crossing is dangerous. Hence, structural factors, opportunities, and/or border closures cannot be separated from individual characteristics in general, and risk-taking attitudes in particular, in migration decision-making.

Conclusion

This study examined factors influencing secondary migration aspirations in forced displacement contexts. Through an online survey experiment, our findings indicate that simply opening borders and implementing welcoming policies do not necessarily incentivize all migrants to consider moving onward. Instead, legal employment opportunities are perceived as preferable to migrating to EU countries, even when border conditions are relatively safe. Additionally, our results highlight the role of individual differences in risk-taking attitudes in determining whether migrants recommend moving or staying. Specifically, migrants with a higher inclination for risk-taking are less deterred by border closures and are more likely to consider moving onward.

Our findings carry significant implications for policy-making, ethical considerations, and analytical approaches. First, there is a growing demand for evidence-based migration policies, as many responses to undocumented migration tend to be reactive and driven by public discontent (Massey et al. 2016). Especially in the current European context, policymakers are increasingly inclined to implement externalization policies to deter people on the move (Adamson and Tsourapas 2019; Cobarrubias et al. 2023). De Haas (2023) makes a similar critique, drawing on historical examples to demonstrate that border restrictions are counter-productive. He argues that such restrictions increase the costs associated with migration, thereby leading to a rise in irregular entry and stay. Additionally, he suggests that these restrictions disrupt circulation and diminish returns. In line with existing studies, our study emphasizes the need for more research that collects evidence on the effectiveness of different migration and development policy options. We aim to stimulate a discussion on how survey experiments, which capture the perspectives of migrants, can contribute to building this evidence base.

Secondly, our findings have political and ethical implications concerning the allocation of limited public resources for migration control. The relative effectiveness of employment opportunities in increasing aspirations to stay informs policymakers about where they can strategically invest their resources. We argue that investing in employment opportunities is a more ethical, humane, and sustainable approach than solely focusing on border controls or externalizing such controls to third countries. However, it is essential to exercise caution in implementing this policy recommendation, as further evidence is necessary to evaluate the effectiveness of various policies. Overall, our study highlights the importance of evidence-based policymaking, ethical considerations in resource allocation, and the potential of survey experiments to inform migration policy discussions.

Third, from an analytical perspective, our analysis raises the need for further research to explore why it is crucial to incorporate risk-taking attitudes into migration aspirations literature. As demonstrated in our research, when migrants’ risk-taking attitudes are taken into consideration, the generalizability of the effectiveness of border externalization policies for all migrants living under similar conditions becomes questionable. By revealing that the onward migration aspirations of Syrian migrants are significantly influenced by their subjective risk-taking attitudes, we challenge the assumption that border closures effectively deter every migrant. Hence, we suggest incorporating risk-taking attitudes into the analysis of micro-level factors that impact migration aspirations (Carling 2019). Further research in this area would contribute to a more comprehensive understanding of the complexities involved in migration aspirations and abilities taking into account variations at the individual level.

Our research has some limitations. First, the subjective risk attitude is a self-reported measurement, which can be influenced by individual biases or social desirability effects. Individuals may have difficulty accurately assessing their risk attitudes, especially if they lack experience in the specific domain being examined. Despite this shortcoming, research shows that subjective measurement of risk attitudes is still a valuable tool in predicting risk-taking behavior across different domains of research (Charness et al. 2013).

Another limitation is around the use of online surveys. Online surveys effectively reach difficult-to-reach populations (Ersanilli and van der Gaag, 2020), yet some population segments may not be adequately represented. For example, individuals without Internet or social media access may be less likely to participate in an online survey, which could lead to a biased sample. However, we lack publicly available demographic data on Syrian migrants living in Turkey. Hence, it is difficult to assess the potential impact of limitations on data representativeness (Elçi et al. 2021). Moreover, our experimental design presented the single-dimensional profile of a hypothetical migrant – a single male migrant working in odd jobs. We acknowledge that these scenarios do not reflect the heterogeneous profiles of Syrians in Turkey including dimensions such as being single versus with one’s family, gender, age, education or socio-economic background. However, the predominant image of border crossers is single male migrants. In this regard, we can justify our choice of the hypothetical profile of a migrant based on our research aims. We acknowledge these potential limitations and encourage further research with alternative methods of data collection that may account for different profiles among the population of interest.

Our study focused primarily on the individual-level factors influencing migration aspirations of Syrian migrants in Turkey. One important limitation is that we did not explore the role of meso-level factors such as social networks, which prior research suggests significantly impacts migration aspirations (Della Puppa and King 2019; McIlwaine 2012; Müller Funk 2019). Social networks can influence destination choices and adaptation to a new country (Müller Funk 2019). Future research can address this gap by examining how social networks, or more specifically, transnational social ties, shape the destination choices, migration aspirations, or, their perceptions about onward migration in forced migration contexts. By analyzing these connections, we can gain a more nuanced understanding of the decision-making process for migrants at different levels of analysis.

Finally, our data was collected during the COVID-19 pandemic, which also affected border controls. Thus, the pandemic might have added another layer impeding migration aspirations to move on. More research should be conducted in the post-COVID-19 period on this topic. We hope that the findings of this survey experiment, which was staged in Turkey, will have implications for similar contexts, where opportunities for migrants are scarce, out-migration is risky, and international cooperation on the ‘containment’ of refugee flows is high.

The data used in this research is part of a larger dataset collected for the ADMIGOV project, which received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No. 822625. Quantitative data is available upon request from the authors.

Acknowledgements

We are immensely grateful to all the participants who took the time to complete the questionnaire and share their experiences. The authors would like to thank Duygu Merve Uysal, participants of the Midwest Political Science Association 78th Annual Conference, participants of the IMISCOE 2023 ‘New Data and Measurements on Migration’ session, participants of the Politicologenetmaal 2023 methods workshop, our collaborators in the ADMIGOV project, and the anonymous reviewers for their constructive comments and suggestions on earlier versions of this paper.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Horizon 2020 Framework Programme [grant number: 822625].

Notes

1 Turkey does not offer any long-term solutions for Syrian migrants, and the overwhelming majority of them have a temporary legal status. While official numbers indicate that over 200,000 Syrians received Turkish citizenship (see: https://www.diken.com.tr/soylu-vatandaslik-alan-suriyeli-sayisi-200-bin-950/), an overwhelming majority of Syrians have a dubious legal status. The Regulation on Temporary Protection grants access to public services such as healthcare and education, but it does not grant the right to work in the formal labor market. Despite the provision of basic services, many Syrians face obstacles on a daily basis due to informal employment, scarce livelihood opportunities, rising living costs, and widespread anti-immigration attitudes in society (Baban, Ilcan, and Rygiel 2017). To avoid confusion with the legal status of refugees, this paper will refer to Syrians with temporary protection as ‘migrants’.

2 The PMM website https://en.goc.gov.tr/temporary-protection27 Last accessed: 20 May 2023.

3 In our study, we employ the term ‘externalization’ as a comprehensive concept encompassing various practices, such as border closures, controlling smuggling networks, or border guards’ attitudes towards refugees. Although our research does not specifically delve into distinct forms of border externalization policies, it aligns with and contributes to the existing literature in this field. Acknowledging the differences between the concepts of border closures and externalization, we use both terms interchangeably.

4 See Pallister-Wilkins (2019) for a critical view on the representation of migration journeys: https://www.societyandspace.org/articles/walking-not-flowing-the-migrant-caravan-and-the-geoinfrastructuring-of-unequal-mobility.

5 See Work Permits of Foreigners 2021, published by the Turkish Ministry of Labor https://www.csgb.gov.tr/media/90062/yabanciizin2021.pdf.

6 The survey experiment was part of a broader study, ADMIGOV, on migration aspirations among displaced communities in Turkey.

7 Initiated by the EU-Turkey Statement of March 2016, the Emergency Social Safety Net (ESSN) is a major cash transfer program funded by the EU to improve the living conditions of refugees in Turkey.

8 The first treatment priming for the lack of employment opportunities in the current country of residence and the lack of obstacles on borders does not significantly differ from the control group. In Turkey, being employed often means working in odd jobs, which places many migrants in unbearable and uncertain situations, blurring the line between unemployment and insecure employment. Therefore, the migrant experience described in the control group and in the first treatment may have been perceived similarly by the participants.