Does cultural proximity contain terrorism diffusion?

Abstract

What are the security consequences of population movements? This article seeks to provide a better understanding of when, how, and under what conditions terrorism diffuses across countries via migration flows as a vehicle. We contribute to this debate by studying the influence of migrants’ cultural proximity to the native population of their host country. It is argued that cultural closeness can contain such terrorism diffusion. Similarities in societal norms, customs, or beliefs seem likely to induce trust in the social interactions between migrants and locals. This, in turn, makes it more difficult for terrorist organizations to exploit transnational population movements for radicalization and as a recruitment pool – one of the core mechanisms linking population flows with terrorism. Conversely, migrants from culturally distant societies may find it more challenging to integrate into their new homes. A fertile ground for terrorist organizations for the recruitment of new followers is thereby more likely. Our analyses present consistent evidence that the effect of terrorism diffusing across countries weakens when accounting for cultural closeness between migrants and host societies. This key finding of our research has crucial implications for policy’s and scholars’ understanding of terrorism, the diffusion of terrorism across countries, and the security consequences of population movements.

Keywords

cultural proximity diffusion migration terrorism violence

Introduction

The world’s migrant population has grown by more than 40% during the past 15 years and, indeed, most states now have significantly more immigrants than in the 1990s (UN DESA, 2016). Not surprising, international migration has become one of the most salient contemporary policy issues and its consequences have attracted particular attention from policymakers and scholars alike. For instance, the literature generally agrees that host countries benefit from migration economically (Dustmann & Frattini, 2014). However, we must not ignore some of the more difficult political consequences of population movements, including security concerns linked to migration. There is evidence that migrants can have an important role in global security politics (Greenhill, 2010) and act as a conduit for transnational action, such as third-party military intervention in civil wars (Bove & Böhmelt, 2019). Transnational migration may also challenge the stability of receiving countries by making it harder for states to control their territory (e.g. Adamson, 2006; Helbling & Leblang, 2019). Migration inflows may further affect the ethnic composition of host nations (e.g. Dowty & Loescher, 1996) or could facilitate the traveling of weapons, combatants, and ideologies across borders (Lischer, 2015; Salehyan & Gleditsch, 2006). Additionally, population movements are frequently targeted by combatants and terror organizations (Choi & Salehyan, 2013) and they might provoke retaliatory cross-border incidents between neighbors (Salehyan, 2008).

The traditional focus in this literature focusing on the security implications of population movements has recently shifted toward the types of migrants and the surrounding contextual parameters, which drive the impact of migration in diverse ways (see also Skeldon, 2008). Bove & Böhmelt (2016) find that a large number of immigrants increases the odds of terrorism in the host country only when migrants stem from terrorist-prone countries of origin. The underlying theoretical mechanism for this effect is that migrant communities tend to have common views, loyalties, and a strong sense of community, which form pre-existing networks that can – more easily than in the case of other groups – be exploited by terrorist organizations for radicalization and recruitment (Sageman, 2004, 2011). Following this research, several studies explore the conditions under which migrants can actually give rise to security threats in recipient states, in particular terrorism. Dreher, Gassebner & Schaud (2017) ask whether immigration affects the risk of terrorism and Böhmelt & Bove (forthcoming) analyze to what extent national migration policies moderate the diffusion of terrorism.

The following article contributes to a better understanding of when, how, and under what conditions terrorism diffuses across countries via migration flows as a vehicle. Specifically, we investigate whether and how migrants’ cultural proximity to the host society mitigates the diffusion of terrorism via population movements. Culture can be defined as a ‘system of meaning and value shared by a community, informing its way of life, and enabling it to make sense of the world’ (Cohen, 1996: 109). Cultural proximity potentially has the power to shape people’s societal norms, customs, and beliefs. In light of this, theoretically, we offer a thorough account of the conditions under which migration can be a vehicle for terrorism to diffuse from one country to another by taking into account migrants’ cultural closeness to the local population of the host state. Previous empirical research on immigration and security often treats immigrants as homogeneous populations, without fully considering the at times significant variation within migrant streams (see e.g. Kubrin, Hipp & Kim, 2018). Merely pointing to the dichotomy of ‘all immigrants’ vs. ‘all natives’ does not comprehensively incorporate the cultural diversity that may exist across and within such population groups (Ousey & Kubrin, 2018). We overcome these limitations by directly modeling the rich diversity within migration flows, and between immigrants and native populations, as we incorporate the cultural proximity of foreign-born individuals to native populations into the relationship of migration flows and terrorism diffusion.

We develop the argument that diffusion of terrorism across countries via migration depends on the cultural closeness between immigrants’ countries of origin and their destination state. A shared understanding and a cultural bond between migrants and natives, that is, cultural proximity, contribute to a common identity and increase the level of trust among individuals (see Guiso, Sapienza & Zingales, 2006; Spolaore & Wacziarg, 2016), which facilitates the integration of migrants into the host society. In turn, this should make it more difficult for terrorist organizations to exploit migrant communities for radicalization and as a recruitment pool – one of the key mechanisms that associate migration with terrorism (Sageman, 2004, 2011). Conversely, ‘people of different cultures will have greater difficulty in interaction, in understanding, and valuation’ (Carnevale & Choi, 2000: 16), and larger cultural distances erode trust and social cohesion within societies, making it more difficult for migrants to integrate (Allport, 1954; Velasco González et al., 2008; Ghosn, Braithwaite & Chu, 2019). These conditions of marginalization can, as we contend, fuel migrants’ radicalization and make it easier for terrorist organizations to recruit individuals (Lyons-Padilla et al., 2015).

Despite the lack of extensive data on when immigrants are the perpetrators – or rather the victims of terrorist violence by, for example, anti-migration groups in the destination country – terrorist attacks are often carried out by native-born citizens.¹ In addition to taking an active part in terrorist violence, there is a variety of indirect means of support that terrorist groups can obtain from immigrants. Immigration flows from terrorist-prone countries could amplify the interaction between and interdependence of terrorist organizations across countries. Social ties based on migration flows may also contribute to terrorism through the diffusion of ideologies and the exchange of information within migrant networks. Migration further increases the exposure of domestic groups to prospects for mobilization and can allow for the exchange of ideas, resources, and knowledge (Zimmermann & Rosenau, 2009; Bove & Böhmelt, 2016). We argue that regardless of whether the support is direct or more indirect, this is facilitated by the marginalization of migrant groups. In the words of Zimmerman & Rosenau (2009: 9), ‘the failure of integration has an obvious, but pernicious consequence – unabated marginalization leaves diasporic communities vulnerable to exploitation by radicals’. In an interview, former jihadist David Vallat offers examples of how marginalization and lack of integration, particularly of French Muslims, imparted a sense of not being full citizens. The lack of economic and social assimilation of immigrants or those living in immigrant communities often promotes a vicious cycle of victimization: ‘the recruitment process across the world capitalizes on the sense of marginalization young people feel and that creates a powerful rhetorical argument that the enemy is the state’.²

As a result, we claim that cultural proximity between migrants and host countries can moderate the cross-national diffusion of terrorism via migrants. Our empirical results are based on quantitative, spatial-econometric analyses. We show that cultural closeness can indeed dampen the commonly perceived strong link between migration movements and the diffusion of terrorism: in fact, when directly considering cultural proximity for our estimations, there is little evidence that migration is linked to terrorism diffusion in significant ways. This main result carries important implications for policy’s and scholars’ understanding of terrorism.

Terrorism diffusion via migration: The role of culture

The economic, political, and social determinants of terrorism are widely studied (e.g. Enders, Sandler & Gaibulloev, 2011; Young & Findley, 2011; Wilson & Piazza, 2013; Gaibulloev, Piazza & Sandler, 2017; Gaibulloev & Sandler, 2018). In this literature, the diffusion of terrorism refers to a particular aspect, namely spatial links connecting countries to each other, which allow for the possibility that terrorism in one national context is influenced by terrorism in other states. Braithwaite & Li (2007) shed light on the clustering of terrorist incidents in space as they identify ‘hot spots’ of terrorist attacks, that is, areas in which terrorism occurs much more frequently than in others.³ In addition, Neumayer & Plümper (2010) report spatial dependence of international terrorism along civilizational lines, while Findley, Piazza & Young (2012) examine the influence of state rivalries on terrorism traveling across borders. Blomberg & Hess (2008) and Li & Schaub (2004) focus on the impact of globalization on the diffusion of terrorism between state pairs, whereas Braithwaite & Chu (2018) show that militants from civil wars abroad affect terrorism in other states.

Population movements are one form of spatial links that connect countries to each other, making it possible that one national context and its level of terrorism are affected by other countries and their degree of terrorist violence (e.g. Buhaug & Gleditsch, 2008; Bove & Böhmelt, 2016; Braithwaite & Chu, 2018). However, we still lack systematic research on the diffusion of terrorism via migration that incorporates contextual factors and moderating conditions that could enhance or hamper terrorism travelling across borders. In the following, we add to this debate as we investigate one crucial aspect of population movements, which also allows us to move beyond the rather simplistic view of migrants as homogeneous populations: the cultural closeness of immigrants to their host countries. While migration can be a cross-national diffusion path and immigrants can be a vehicle (directly or indirectly) for terrorism to travel from one state to another, we contend that the cultural proximity between migrants’ home and host countries matters and, in fact, can work against such diffusion of terrorism.

The economic literature has long argued that cultural diversity, the range of citizens with different origins, religions, and traditions living and interacting together, plays a pivotal (and mostly positive) role in shaping economic growth (Alesina & Ferrara, 2005). Yet, depending on the circumstances, diversity can be associated with both normatively adversarial and beneficial effects. That is, although diversity can lead to positive organizational synergies, a larger cultural distance is often characterized by different norms practiced, different perceptions held, and more misunderstandings between, in our case, migrants and local populations (Bakaki, Böhmelt & Bove, 2016; Alesina & Ferrara, 2005). Particularly relevant to our research, by potentially being culturally distant to the local population and eventually further affecting cultural diversity in host societies, immigration influences individual interactions in decisive ways: in fact, cultural barriers are usually seen as customary impediments to interpersonal trust, solidarity, and social capital (see e.g. Guiso, Sapienza & Zingales, 2006).

In line with this mechanism, several studies show that immigrants from countries that are culturally more similar to the host state are better able to integrate (e.g. De Wit & Koopmans, 2005; Maxwell, 2010; Isphording, 2014). Conversely, cultural distance can be a major hurdle to social and economic interactions across groups (Gokmen, 2017). Immigrants with culturally more distant backgrounds who are more concerned over the preservation of their own identity are particularly less likely to assimilate (Dowty & Loescher, 1996), fueling the fear of host societies that migration may threaten their existing cultural identity (Velasco González et al., 2008). And indeed, the cultural distance between social groups is likely to be a core driver of locals’ negative attitudes and prejudice towards immigrants (e.g. Allport, 1954; Weiner, 1992; Mahfud et al., 2018). Clearly, cultural proximity does not by default lead to trust-building among immigrants and natives, but it serves as an important facilitating factor. When subscribing to the link between cultural proximity and trust, this bond might ease migrants’ integration through, for example, the social inclusion of migrants in receiving societies and improved labor market participation. Cultural closeness between migrants and the local population of the destination country is therefore likely to be a significant factor for the integration of immigrants (see also Angelini, Casi & Corazzini, 2015).

A larger cultural distance of migrants to the population of their new home could marginalize the former and impede their integration – which, in turn, leads to one of the main mechanisms associating migration flows with terrorism diffusion: radicalization and the recruitment of individuals by terrorist organizations (Sageman, 2004, 2011). That is, cultural identity plays an important role in the radicalization of individuals (for an overview, see Lyons-Padilla et al., 2015). Immigrants who are culturally homeless lack a clear feeling of belonging and are more likely to become marginalized, making them likely to be more attracted to groups that offer a sense of inclusion, purpose, and identity. And this frequently constitutes the first step toward developing a radical belief system (Hogg, 2011; Doosje, Loseman & Bos, 2013; Lyons-Padilla et al., 2015). Not surprisingly, then, higher levels of prejudice toward immigrants have been associated with less cultural closeness between immigrants and the majority group (Mahfud, Badea & Ngbala, 2015). This may make it easier for marginalized communities to be attracted to groups offering a feeling of identity – and immigrants with a more culturally distant background are not only more likely to feel a loss of significance, but are also then more susceptible to radicalization (Zimmerman & Rosenau, 2009; Lyons-Padilla et al., 2015). All this is facilitated by strong social bonds that usually connect individuals within migrant populations. Sageman (2004, 2011) contends here that these ties predating recruitment into terrorist organizations are the crucial element of this process (see also Bove & Böhmelt, 2016): the existence of social bonds comes first and ideology follows (Sageman, 2004: 133). The process of joining terrorist groups is then ‘through mutual emotional and social support, development of a common identity, and encouragement to adopt a new faith. All these factors are internal to the group’ (Sageman, 2004: 135). Terrorist organizations can exploit these social linkages within migrant groups, particularly if migrants feel marginalized and excluded, for radicalization and to recruit members (Sageman, 2011).

Particularly when immigrants are not fully integrated in host societies, resentment and anger can facilitate their recruitment – or the recruitment of their children – by terrorism organizations (Schmid, 2016). For example, in both France and Belgium, many immigrants and their offspring live in deprived city neighborhoods, ‘separated culturally but linked physically to the surrounding urban landscape’, sparking a heated debate on the link between terrorism and lack of assimilation.⁴ The Molenbeek district of Brussels, a melting pot of different nationalities, illustrates this. Many recent terrorism attacks across Europe have been linked to people residing in this area, and a main reason that has often been put forward for this is the lack of access to important services, particularly active integration programs for migrants.⁵ Consider also Cherif Kouachi, involved with his brother in the 2015 Charlie Hebdo shooting. Raised in a northern suburb of Paris, he was driven by a sense of social estrangement. According to Mohammed Benali, president of the local mosque, he was of a ‘generation that felt excluded, discriminated against and, most of all, humiliated. They spoke and felt French, but were regarded as Arabic; they were culturally confused.’ Similarly, Mohammad Sidique Khan, the leader of the 7/7 bombings in London (2005), was caught ‘between no cultures’ and expressed his rage through terrorist violence.⁶ In contrast, the promotion of immigration and assimilation is regarded as the best deterrent against radicalizing people to join terrorist organizations.⁷

Marginalized migrants from terrorist-prone countries do not necessarily have a direct effect on terrorist activities by, for example, taking an active part in terrorist attacks, but may support domestic terrorist group in the adopted country in an indirect way, which eventually heightens the level of terrorism. Migration can increase the exposure of domestic groups to prospects for mobilization, thus making emulation more likely to emerge (Adamson, 2006). Similarly, members of immigrant communities may allow for the exchange of ideas, as terrorist groups in the host country often lack the relevant experience to organize terrorist activities (Salehyan & Gleditsch, 2006; Choi & Salehyan, 2013). As such, the connection between foreign and domestic terrorist organizations mainly focuses on information exchange. Finally, migrant inflows could facilitate the establishment of transnational links between terrorist groups, which induce cooperation, the pooling of resources, and the access to knowledge that would be unavailable otherwise.⁸

Cultural proximity between immigrants and destination countries could offset this – if migrants are not culturally distant from, but closer to the host society and, thereby, potentially more integrated and less marginalized. Cultural proximity between immigrants and natives is a crucial intervening factor that moderates the impact of migration on terrorism diffusion. On one hand, assimilation and integration are likely stronger when immigrants are culturally closer to the native population of the destination country. Integration promotes social, economic, and civic well-being (Vigdor, 2015). By reducing the odds of failed assimilation, cultural proximity also lowers feelings of personal uncertainty, injustice, and perceived intergroup threats that are among the key determinants of radicalization (see also Rahimi & Graumans, 2015). On the other hand, cultural closeness can prevent the development of negative out-group attitudes (Zimmerman & Rosenau, 2009; Lyons-Padilla et al., 2015). To the extent that social integration matters for predicting a group’s vulnerability to terrorist recruitment (Zimmermann & Rosenau, 2009; Sageman, 2004, 2011), and subscribing to the claim that cultural proximity facilitates assimilation and integration, terrorist groups should find it more challenging to radicalize and recruit followers under those circumstances. This should weaken the effect of migration as a vehicle for the diffusion of terrorism. Hence, cultural proximity of migrants to their host country contains, and acts as a barrier to, the diffusion of terrorism via population movements. This leads to the following hypothesis:

Hypothesis: Cultural proximity weakens the diffusion of terrorism across countries.

Research design

Data

We created a monadic (country-year as the unit of analysis) panel dataset comprising OECD countries from 1980 to 2010. We concentrate on OECD states as possible destinations for migrants, but allow for migration flows from anywhere in the world (see below). In other words, non-OECD states are not considered as destinations of migration movements in our study, but all countries worldwide are possible origins of migration flows. Focusing on OECD host countries has the benefit of examining a set of states that is rather homogeneous in a number of socio-economic and political characteristics, for example, levels of democracy, development, and membership in international organizations. This helps to isolate the effects stemming from the spatial diffusion of terrorism. Furthermore, OECD countries are among the top destinations of international migration. As of 2015, more than half of the international-migrant population was hosted in the OECD region (see UN DESA, 2016).

For our dependent variable, we measure the level of terrorism at the country level each year using the Global Terrorism Database (GTD) that defines terrorism as ‘the premeditated use or threat to use violence by individuals or sub-national groups against noncombatants in order to obtain a political or social objective through the intimidation of a large audience beyond that of the immediate victims’ (Enders, Sandler & Gaibulloev, 2011: 321). We do not distinguish between national and transnational attacks as our argument applies to both cases (Bove & Böhmelt, 2016; Sageman, 2004, 2011).⁹ Given the skewed distribution of the number of terrorist incidents, our final outcome variable is the log-transformed number of terrorist attacks in a given year after adding 1.

Data on migrant populations are taken from the World Bank (Özden et al., 2011: 17), which relies on two main criteria to define migration: being born in or being a citizen of a foreign country. That said, the place-of-birth definition is considered superior by the World Bank as ‘while nationality can change, place of birth cannot’. And when data are available for both criteria, the World Bank gives priority to the birthplace (Özden et al., 2011: 17). Furthermore, Özden et al. (2011) subtract the number of refugees from total migrant numbers as the focus is mostly on economic migrants.¹⁰ The data are therefore consistent with our theoretical argument, which pertains to diaspora communities or people who are permanently settled in a country. Refugee flows are a temporary movement of people that flee violence to seek protection, but we focus on the longer time horizon of migrants as opposed to refugees. Moreover, case-specific narratives highlight that there is usually a longer period of radicalization and, hence, using the stock of immigrants rather than recent entrants is a more suitable approach (see also Dreher, Gassebner & Schaud, 2017: 5). In addition, while migration is a phenomenon of global scope with a much wider reach, refugee flows are more localized, usually constrained to neighboring countries.¹¹ The estimates are derived from national census and population register records for 232 destinations (countries and country-like territories), including our OECD states.¹² As each census round was conducted during a ten-year window, we linearly interpolate all missing data between two consecutive rounds. After accounting for missing values and temporally lagging all our explanatory items, our sample comprises 32 potential host states from the OECD that we combine with more than 200 countries of origin to create bilateral matrices of migration populations between sender and receiver nations each year between 1980 and 2010.¹³ We use the data on migration for the construction of the spatial lags, which we describe in the methodology section.

A central challenge for our empirical study is that the estimates may be biased by endogeneity stemming from omitted variables. This becomes even more important for the spatial-econometric approach that we describe below as we must distinguish a genuine diffusion effect from mere country-level influences and spatial clustering (Franzese & Hays, 2007, 2008; Plümper & Neumayer, 2010; Buhaug & Gleditsch, 2008). To mitigate this concern, we control for several relevant country attributes that are both spatially clustered and potentially related to terrorism (e.g. Enders, Sandler & Gaibulloev, 2011; Krieger & Meierrieks, 2011; Young & Findley, 2011; Wilson & Piazza, 2013; Gaibulloev, Piazza & Sandler, 2017; Gaibulloev & Sandler, 2018), which allows us to accurately identify a real spatial diffusion effect – and it is our aim then to see how cultural proximity between migrants and host societies alters this. First, Gaibulloev, Piazza & Sandler (2017: 15) recommend controlling for a state’s involvement in foreign policy. To this end, we consider a variable on alliance ties with the USA, which is binary and based on the Correlates of War Formal Alliance dataset (Gibler, 2008).

Second, we include information on per-capita GDP and population as wealthier and less populous states are likely to experience significantly less terrorism (Li, 2005; Piazza, 2006). Both items are log-transformed and, as all other variables, lagged by one year. We also control for the level of democracy using the revised and combined polity score from the Polity IV database (Marshall & Jaggers, 2015). The score potentially assumes values between –10 and +10, with higher values denoting more democratic forms of government.

Third, immigrants might select themselves into countries based on potential economic opportunities and the generosity of local migration policies, and this has long been a concern in the cross-country migration literature

Table I.

Descriptive statistics: Dependent variable and controls

Variable	Mean	Std. dev.	Min.	Max.	Obs.
Terrorism (ln)	1.364	1.442	0	6.267	911
Alliance with US	0.682	0.466	0	1	911
Democracy	8.904	3.151	–8	10	911
GDP per capita (ln)	9.943	0.696	7.972	11.382	911
Population (ln)	9.648	1.371	5.898	12.634	911
Total migration population	9.445	8.715	0.424	38.537	911
Internal immigration policy restrictions	0.512	0.089	0.314	0.812	911

as it introduces selection bias. Income and democracy, therefore, not only control for alternative mechanisms of terrorism, but also should correlate with migration flows (e.g. Breunig, Cao & Luedtke, 2012; Alarian & Goodman, 2017; Helbling et al., 2017). We further control for the number of immigrants (as a percentage of the population), and the level of immigration restrictions for the same reason. The number of immigrants as a share of the total population in a given country-year disentangles the effect of the spatial variables we present in the next section from the impact of immigration alone, which is theoretically and empirically different. Migrants also select themselves into those destinations that already have larger diaspora communities.

Finally, as immigrants are more likely to move to countries with less stringent immigration policies in terms of, for example, regulations or control mechanisms (e.g. Breunig, Cao & Luedtke, 2012; Alarian & Goodman, 2017; Helbling & Leblang, 2019), we include a control using the Immigration Policies in Comparison (IMPIC) dataset (see Helbling et al., 2017).¹⁴ These data offer information on the total level of restrictiveness of immigration policies across four dimensions for all OECD countries between 1980 and 2010. The data are presented on a scale between 0 and 1, where 1 is the maximum relative level of restrictiveness. We use the data’s aggregated variable that combines internal and external regulations with control mechanisms. Table I summarizes the descriptive statistics of the variables discussed so far.

Methodology

We estimate spatial temporal autoregressive models based on ordinary least squares (spatial-OLS) and specify a weighting matrix on migrant-population movements and cultural proximity. The core part of this is $W y_{t - 1}$ , which is the spatial variable comprising the product of a row-standardized connectivity matrix (W) with the temporally lagged dependent variable ( $y_{t - 1}$ ).¹⁵ The elements ( $w_{i, j}$ ) in the weighting matrix measure the relative connectivity of country j to country i (with $w_{i, i}$ = 0). A positive and statistically significant effect constitutes a positive diffusion effect in that, in our case, terrorism in the sending state increases terrorism in the recipient country. A negative and statistically significant effect would capture a negative diffusion effect in that, in our case, more terrorism in the sending state decreases terrorism in the recipient country. Following Franzese & Hays (2008, 2007: 142), we control for a number of relevant ‘exogenous-external conditions or common shocks and spatially correlated unit level factors’. Next to the control variables we outlined above and the temporally lagged dependent variable, country effects are added to account for unobserved heterogeneities that are specific to each country and to leverage changes occurring within units. Time effects are entered to control for temporal shocks or systemic effects that might affect terrorism, such as the 9/11 attacks.

The core of this methodological approach and, hence, our empirical analysis are three spatial lags. All spatial items are based on a matrix that links countries via migrant populations.¹⁶ That is, the elements of each variable’s matrix measure for each pair of countries in a given year the size of the migrant population of a foreign state (country of origin) in the country under study (host state). While one (the first) spatial lag’s underlying weights matrix is only based on migrant populations, the other two spatial lags rely on weighting matrices that further include information about cultural proximity and distance, respectively, between countries. Ultimately, for the first spatial lag that is only based on migration populations, each element $w_{i, j}$ of the connectivity matrix measures the migrant population in country i that has country j as the state of origin in t–1. In the absence of any migration population from j in i, $w_{i, j}$ takes the value of 0. This row-standardized matrix is multiplied with the temporally lagged dependent variable ( $y_{t - 1}$ ) to create the first spatial lag, which then measures the average degree of terrorism in other countries weighted by migrant populations. As indicated above, we take the data on migrant populations from Özden et al. (2011) and we expect a positive and significant effect for this first spatial lag.

The second spatial lag is similar to the first one except for an additional component added to the weighting matrix. That is, each element $w_{i, j}$ of the second connectivity matrix captures the migrant population in country i that has country j as the state of origin in t–1 – but only in culturally close countries. We measure cultural proximity between two countries $i, j$ in the matrix via the cultural-differences information in Kandogan (2012), which is a revised variable of the standardized measure of cultural distance as introduced in Kogut & Singh (1988).¹⁷ For countries’ cultural links in the connectivity matrix, we first paired each state with all other countries and then merged the Kandogan (2012) variable into this dyadic dataset. We reversed the scale of that variable so that higher values pertain to more cultural similarities between states. Afterwards, we divided the item into ten equally sized groups (quantiles), while the five most similar groups eventually receive an entry of 1 in the elements $w_{i, j}$ of the connectivity matrix and the five least similar quantiles are coded as 0 for $w_{i, j}$ . Finally, these entries $w_{i, j}$ are multiplied by the migrant population in country i that has country j as the state of origin in t–1 in order to capture migrant flows between i and j, which are then two states that are by construction culturally more similar to each other.

Our final spatial lag combines information about migrant stocks and cultural dissimilarity in its underlying matrix. That is, using the reversed scale on cultural proximity (Kandogan, 2012) and the ten equally sized groups, the five least similar quantiles are now coded as 1 in the elements

w_{i, j}

of the connectivity matrix (all other entries as 0). As above, we multiply these entries with the migrant population in country i that has country j as the state of origin in t–1, which then measures migrant

Table II.

Descriptive statistics: Spatial lags and lagged dependent variable

Variable	Mean	Std. dev.	Min.	Max.	Obs.
Lagged dependent variable	1.401	1.459	0	6.267	911
W_y: Migrant inflow	1.271	0.425	0.539	2.536	911
W_y: Migrant inflow cultural proximity	1.479	0.753	0	3.365	911
W_y: Migrant inflow cultural distance	1.473	0.575	0	3.06	911

populations across two culturally dissimilar states. In light of our theory, we expect an insignificant estimate for the culturally similar spatial lag, but a positive and significant impact stemming from our last spatial variable that focuses on culturally distant states. Table II summarizes the descriptive statistics of the spatial variables and the temporally lagged dependent variable. The weighted level of terrorism in all other countries is around 1.27, when the weights are based on migrant populations only. When focusing on migrant flows from culturally proximate countries, the value increases to 1.479, while it is around 1.473 for migration from culturally distant states. In the following, we assess whether these items shape terrorism ‘at home’ in systematic ways.

Findings

All models in Table III are identical in terms of controls and fixed effects, but they differ in the spatial variable(s) considered. Model 1 focuses on W _y: Migrant inflow. Model 2 replaces that item by the spatial lag that is weighted by migration inflows and cultural proximity. Model 3 concentrates on terrorism weighted by migration in culturally distant states. Finally, Model 4 comprises both W _y: Migrant inflow cultural proximity and W _y: Migrant inflow cultural distance. Due to the row standardization, the spatial lags can be interpreted directly. However, as we include a temporally lagged dependent variable, our coefficient estimates of the spatial lags only reflect the short-term effect, that is, the impact in a current year. The levels of asymptotic long-term influence of these items are calculated

Table III.

Terrorism diffusion: The role of cultural proximity

	Model 1	Model 2	Model 3	Model 4
Lagged dependent variable	0.475**	0.474**	0.477**	0.475**
	(0.030)	(0.030)	(0.030)	(0.030)
W _y : Migrant inflow	0.529^†
	(0.271)
W _y : Migrant inflow cultural proximity		0.074		0.073
		(0.093)		(0.093)
W _y : Migrant inflow cultural distance			0.229^†	0.228^†
			(0.139)	(0.139)
Alliance with US	0.185	0.136	0.152	0.162
	(0.152)	(0.150)	(0.150)	(0.151)
Democracy	0.002	–0.003	–0.003	–0.002
	(0.011)	(0.011)	(0.011)	(0.011)
GDP per capita (ln)	–0.439*	–0.372^†	–0.389^†	–0.437*
	(0.214)	(0.215)	(0.210)	(0.218)
Population (ln)	1.213 *	0.953^†	1.000*	1.047*
	(0.514)	(0.494)	(0.493)	(0.496)
Total migration population	0.012	0.006	0.003	0.004
	(0.018)	(0.018)	(0.018)	(0.018)
Internal immigration policy restrictions	0.293	0.222	0.223	0.303
	(0.502)	(0.508)	(0.499)	(0.509)
Constant	–11.346^†	–7.753	–8.516	–8.838
	(6.660)	(6.376)	(6.389)	(6.403)
Observations	911	911	911	911
Country fixed effects	Yes	Yes	Yes	Yes
Year fixed effects	Yes	Yes	Yes	Yes
Prob. > F	0.000	0.000	0.000	0.000

Standard errors in parentheses. ^† p < 0.10, *p < 0.05, **p < 0.01.

according to Plümper, Troeger & Manow (2005: 336) and discussed in the text where relevant.

First, W _y: Migrant inflow is positively signed and significant at the 10% level in Model 1. This finding mirrors and replicates Bove & Böhmelt (2016) in that it underlines that migration flows can be a vehicle for terrorism to diffuse from one state to another. We can rule out that this is driven by geographical proximity, since our analysis combines OECD-only host states with all other countries in the world. In substantive terms, the marginal effect in Model 1 suggests that a one-unit increase in W _y: Migrant inflow leads to a rise in terrorist attacks of about 1.67 events. As indicated above, this is merely the short-term effect, though. The asymptotic long-term marginal effect of this first spatial lag is about 0.968 (coefficient estimate), which translates into 2.6 attacks. While these results mirror existing findings on terrorism diffusion via population movements, therefore highlighting that the degree of terrorism ‘at home’ increases with migrants from countries with a high level of terrorism, the first model does not take into account that states might be linked via cultural ties to each other. In other words, our first estimation in Table III treats immigrants as homogeneous populations, merely pointing to the dichotomy of ‘immigrants’ vs. ‘natives’ without fully incorporating the cultural diversity that may exist across and within such populations (Ousey & Kubrin, 2018; Kubrin, Hipp & Kim, 2018). To this end, Models 2–4 concentrate on W _y: Migrant inflow cultural proximity and W _y: Migrant inflow cultural distance.

In light of our hypothesis, we do not suggest that the cultural-proximity spatial lag should be negatively signed and/or significant, but that it should weaken and, thus, render insignificant the original spatial variable. As Model 2 shows, the spatial lag incorporating cultural closeness is now statistically insignificant, providing little evidence for a strong – or any – influence stemming from migration populations on terrorism diffusion when directly modelling that migrants come from culturally close countries. Changing model specifications by adding control variables, altering the way we calculate the standard errors, or considering refugee populations does not affect this finding qualitatively: as the Online

Figure 1.

Terrorism diffusion: The role of cultural proximity

appendix demonstrates, W _y: Migrant inflow cultural proximity remains statistically insignificant even then. The conclusion holds when examining the simulated marginal effects following King, Tomz & Wittenberg (2000) in Figure 1: while the marginal-effect estimates of W _y: Migrant inflow center around 0.5 with only a small portion of the 1,000 simulated coefficients being 0 or negative (less than 2.3 percent of the simulated marginal effects), the simulated estimates of W _y: Migrant inflow cultural proximity converge toward 0. Hence, we are unable to identify a statistically significant effect once we incorporate migrants’ cultural proximity to their host society: cultural closeness does indeed dampen the diffusion of terrorism via migration flows. Interestingly, the substantive finding of migration being a vehicle for terrorism to diffuse across countries is unaltered for W _y: Migrant inflow cultural distance. Substantively, Models 3 and 4 emphasize that a one-unit increase in W _y: Migrant inflow cultural distance is associated with an increase in terrorist events of about 1.26. These results, in sum, emphasize in line with our theoretical expectations that more cultural proximity can address, contain, and indeed be a barrier to the diffusion of terrorism via migration. In other words, cultural proximity can dampen the diffusion effect of migration, while cultural distance does not.

This conclusion is based on statistical significance, but statistical significance may be a poor criterion for policy prescriptions (Ward, Greenhill & Bakke, 2010). In their words, ‘a variable that might at first be thought to represent an important conceptual breakthrough in our understanding of conflicts, owing to its statistical significance, often only leads to a very modest improvement in our ability to predict’ outcomes (Ward, Greenhill & Bakke, 2010: 365). Put differently, next to statistical significance, predictive power is an important factor to consider and out-of-sample heuristics must not be ignored for deciding whether the inclusion of a variable contributes to our understanding of, in our case, terrorism diffusion or not. This implies that we should be able to demonstrate not only that W _y: Migrant inflow cultural proximity exerts an effect that is indistinguishable from 0, but also that the inclusion of this variable leads to predictive gains that help us to generalize the effect of cultural proximity to out-of-sample contexts. Hence, not only should W _y: Migrant inflow cultural proximity render insignificant the original spatial variable, but also the underlying model should perform better from a prediction point of view than the original setup when adding the information on cultural proximity. We thus conducted a four-fold cross-validation quasi-experimental exercise, which we repeated ten times for Models 1, 2, and 4 above and an additional model that is similar to the estimations in Table III, but we omit any of the spatial lags.

To this end, as an initial step, we randomly divided our sample into four segments of about the same size. We then used three random segments to estimate the parameters, while the fourth segment was retained for assessing the predictive power of either of the four models on the pooled subsets. We provide two goodness-of-fit measures in this out-of-sample setup. First, Theil’s U is the square root of the ratio between the sum of squared prediction errors of a model and the sum of squared prediction errors of a naive model, that is, a ‘no-change prediction’ where the level of terrorism in t–1 fully corresponds to the level in t. If Theil’s U is larger than 1, the model performs worse than the naive model; values of Theil’s U smaller than 1 indicate that the ‘theoretically informed model’ performs better than the naive specification. Second, the mean squared prediction error (MSPE) pertains to the expected value of the squared difference between the observed values of the outcome variable and the predicted ones.

We calculated both measures for Models 1, 2, and 4 as well as a constrained model that does not include any of the spatial lags. As indicated above, we repeated the cross-validation ten times and, thus, obtain for each model ten different values for Theil’s U and the MSPE,

Table IV.

Four-fold cross-validation

	Constrained model	Model 1	Model 2	Model 4
	0.823	0.801	0.814	0.803
	(0.462)	(0.438)	(0.451)	(0.439)
	0.820	0.823	0.806	0.818
	(0.458)	(0.461)	(0.444)	(0.456)
	0.828	0.828	0.815	0.817
	(0.468)	(0.467)	(0.453)	(0.455)
	0.821	0.813	0.819	0.817
	(0.459)	(0.450)	(0.457)	(0.455)
	0.824	0.817	0.816	0.818
	(0.463)	(0.465)	(0.454)	(0.456)
	0.816	0.806	0.813	0.807
	(0.454)	(0.443)	(0.451)	(0.445)
	0.821	0.828	0.802	0.816
	(0.459)	(0.468)	(0.438)	(0.455)
	0.802	0.812	0.820	0.815
	(0.438)	(0.450)	(0.459)	(0.452)
	0.815	0.828	0.818	0.813
	(0.453)	(0.467)	(0.456)	(0.451)
	0.830	0.826	0.818	0.817
	(0.470)	(0.465)	(0.456)	(0.455)
Mean value	0.820	0.818	0.814	0.814
	(0.458)	(0.457)	(0.452)	(0.452)

Table entries are Theil’s U values with MSPEs in parentheses.

respectively. We calculated the average values for both model-fit statistics to arrive at global values. The results are summarized in Table IV. For the constrained model, the average Theil’s U across all ten iterations of the cross-validation is 0.820, while the corresponding MSPE stands at 0.458. Interestingly, both the corresponding average Theil’s U and the MSPE values are lower for Models 1, 2, and 4. As a result, the predictive power of our core variables of interest is established as the prediction error tends to increase when omitting W _y: Migrant inflow, W _y: Migrant inflow cultural proximity, or W _y: Migrant inflow cultural distance.

In terms of control variables, our results are consistent with recent studies on the economic, political, and social causes of terrorism (e.g. Wilson & Piazza, 2013; Gaibulloev & Sandler, 2018). We find that whereas larger countries attract more terrorist attacks, and thus Population (ln) is positively and significantly correlated with the incidence of terrorism, higher income is associated to a lower degree of terrorism, as GDP per capita (ln) is negatively signed (e.g. Young & Findley, 2011). Similarly, and consistent with earlier studies, we find that the lagged dependent variable is positive and significant; in other words, terrorism displays temporal dependencies and a higher number of terrorist attacks in the previous year is correlated with more terrorism in the current period. The Democracy variable is statistically insignificant as is the alliance item, which is expected given our focus on OECD host countries. Finally, migration policies or the (‘raw’) migration population in the host country do not seem to crucially shape the level of terrorism. While omitting these variables from the models does not alter our results for W _y: Migrant inflow cultural proximity and W _y: Migrant inflow cultural distance, the insignificance of their coefficient estimates not only highlights yet again that migrants are not to blame for an increased risk of terrorism, but also that more restrictive national migration policies per se are unlikely to be an all-encompassing instrument against terrorism.

In the Online appendix, we assess the robustness of our results along various avenues. First, we consider genetic distance in lieu of cultural distance. Second, we include additional variables to our estimations. Third, we run a series of models with bootstrapped standard errors. Fourth, we check whether the effect of democracy is non-linear. Fifth, although our theoretical focus is on the longer time horizon of migrants as opposed to refugees, we control for refugee inflows. Sixth, we drop all country-years based on interpolated migration data and distinguish between domestic and transnational terrorism.¹⁸ Finally, we examine short-term migration influx, employ a different estimator, and compare explanatory variables across culturally close and distant dyads.

Conclusion

A recent strand of academic research finds that terrorism at home can be influenced by terrorism abroad, and that population movements may facilitate this transnational diffusion of terrorism (Bove & Böhmelt, 2016). One central argument of these works suggests that strong social bonds facilitate the establishment of ‘terror networks’ and a pre-established social framework is a key requirement for terror organizations radicalizing and recruiting individuals (Sageman, 2004, 2011). As migration flows comprise social ties and linkages, terrorist groups may exploit those networks of migrant communities for radicalization, as a recruitment pool, and to obtain indirect support through the exchange of ideas, resources, or knowledge.

Yet, cultural proximity between countries may shape individual interactions in diverse ways – and, as we contend, this has implications for terrorist organizations’ recruitment possibilities as well as individuals’ motives to directly or indirectly engage in terrorism. Migrants from culturally close societies may have fewer difficulties integrating into host societies. In turn, social cohesion within societies evolves more easily with migrant communities from culturally closer places, which makes it more difficult for terrorist organization to radicalize and recruit followers. Combining this mechanism with the arguments about terrorism diffusion via migration flows led to the theoretical expectation that cultural proximity can contain, and act as a barrier to, such terrorism diffusion. In fact, we find little evidence for migrants being an instrument for the transnational diffusion when accounting for and directly incorporating information on cultural closeness.

We build on existing arguments on how terrorism may diffuse via migration inflows, but we do not isolate the role of each mechanism that might drive this effect. There is a range of means of direct and indirect support that terrorist groups can obtain from marginalized immigrants, but the main obstacle for investigating them separately is the lack of comprehensive fine-grained data at the level of individuals or organizations. Future work should further disaggregate individuals and organizations as spatial units, and include information about their activities and positions in terrorist organizations, organizations’ links across countries, and, most importantly, the possible role played by immigrants within these groups. In addition, our data do not allow determining whether immigrants are the perpetrators or victims of terrorist violence. There is ample evidence suggesting that foreign-born individuals are vulnerable to the violation of their rights and are too often the targets of violent persecution from hostile local populations (Savun & Gineste, 2019; Böhmelt, Bove & Gleditsch, 2019). Within the broader micro-turn in the study of immigration, recent scholarship has begun to compile new data on refugees’ involvement in acts of physical violence in their host state, either as the victims or perpetrators of violence (see e.g. Gineste & Savun, 2019). Similar coding efforts with detailed information on the nationality and identity of victims and perpetrators of terrorism will allow scholars to assess under which conditions immigrants are targeted by terrorist violence, when and how they offer direct or indirect material support, and when they are exploited as a vehicle for such episodes of violence.

Footnotes

Replication data

The dataset and do-file for the empirical analysis in this article, along with the Online appendix, can be found at .

Acknowledgements

We thank the Associate Editor, Halvard Buhaug, the anonymous reviewers, and Todd Sandler, as well as the participants of the 8th Political Violence and Policy Conference at the University of Texas at Dallas for helpful comments and suggestions. All remaining errors are our own.

ORCID iD

Tobias Böhmelt

Notes

References

Adamson

Fiona

(2006) Crossing borders: International migration and national security. International Security 31(1): 165–199.

Alarian

Hannah

Goodman

Sara

(2017) Dual citizenship allowance and migration flow: An origin story. Comparative Political Studies 50(1): 133–167.

Alesina

Alberto

Ferrara

Eliana

(2005) Ethnic diversity and economic performance. Journal of Economic Literature 43(3): 762–800.

Allport

Gordon W

(1954) The Nature of Prejudice. Boston, MA: Addison-Wesley.

Angelini

Viola

Casi

Laura

Corazzini

Luca

(2015) Life satisfaction of immigrants: Does cultural assimilation matter? Journal of Population Economics 28(3): 817–844.

Bakaki

Zorzeta

Böhmelt

Tobias

Bove

Vincenzo

(2016) Barriers to coordination? Examining the impact of culture on international mediation occurrence and effectiveness. Political Studies 64(3): 492–512.

Blomberg

Brock

Hess

Gregory

(2008) The Lexus and the olive branch: Globalization, democratization, and terrorism. In: Keefer

Philip

Loayza

Normal

(eds) Terrorism, Economic Development, and Political Openness. Cambridge: Cambridge University Press, 116–147.

Böhmelt

Tobias

Bove

Vincenzo

(forthcoming) How migration policies moderate the diffusion of terrorism. European Journal of Political Research. https://doi.org/10.1111/1475-6765.12339.

Böhmelt

Tobias

Bove

Vincenzo

Gleditsch

Kristian

(2019) Blame the victims? Refugees, state capacity, and non-state actor violence. Journal of Peace Research 56(1): 73–87.

10.

Bove

Vincenzo

Böhmelt

Tobias

(2016) Does immigration induce terrorism? Journal of Politics 78(2): 572–588.

11.

Bove

Vincenzo

Böhmelt

Tobias

(2019) International migration and military intervention in civil war. Political Science Research and Methods 7(2): 271–287.

12.

Braithwaite

Alex

Chu

Tiffany

(2018) Civil conflicts abroad, foreign fighters, and terrorism at home. Journal of Conflict Resolution 62(8): 1636–1660.

13.

Braithwaite

Alex

Quan

(2007) Transnational terrorism hot spots: Identification and impact evaluation. Conflict Management and Peace Science 24(4): 281–296.

14.

Breunig

Christian

Cao

Xun

Luedtke

Adam

(2012) Global migration and political regime type: A democratic disadvantage. British Journal of Political Science 42(4): 825–854.

15.

Buhaug

Halvard

Gleditsch

Kristian Skrede

(2008) Contagion or confusion? Why conflicts cluster in space. International Studies Quarterly 52(2): 215–233.

16.

Carnevale

Peter

Choi

Dong-Won

(2000) Culture in the mediation of international disputes. International Journal of Psychology 35(2): 105–110.

17.

Choi

Seung-Whan

Salehyan

Idean

(2013) No good deed goes unpunished: Refugees, humanitarian aid, and terrorism. Conflict Management and Peace Science 30(1): 53–75.

18.

Cohen

Raymond

(1996) Cultural aspects of international mediation. In: Bercovitch

Jacob

(ed.) Resolving International Conflicts: The Theory and Practice of International Mediation. Boulder, CO: Lynne Rienner, 107–128.

19.

DeWit

Thom

Koopmans

Ruud

(2005) The integration of ethnic minorities into political culture: The Netherlands, Germany, and Great Britain compared. Acta Politica 40(1): 50–73.

20.

Doosje

Bertjan

Loseman

Annemarie

Bos

Kees

(2013) Determinants of radicalization of Islamic youth in the Netherlands: Personal uncertainty, perceived injustice, and perceived group threat. Journal of Social Issues 69(3): 586–604.

21.

Dowty

Alan

Loescher

Gil

(1996) Refugee flows as grounds for international action. International Security 21(1): 43–71.

22.

Dreher

Axel

Gassebner

Martin

Schaudt

Paul

(2017) The effect of migration on terror: Made at home or imported from abroad? CESifo Working Paper (6441).

23.

Dustmann

Christian

Frattini

Tommaso

(2014) The fiscal effects of immigration to the UK. Economic Journal 124(580): F593–F643.

24.

Enders

Walter

Sandler

Todd

Gaibulloev

Khusrav

(2011) Domestic versus transnational terrorism: Data, decomposition, and dynamics. Journal of Peace Research 48(3): 319–337.

25.

Findley

Michael

Piazza

James

Young

Joseph

(2012) Games rivals play: Terrorism in international rivalries. Journal of Politics 74(1): 235–248.

26.

Franzese

Robert

Hays

Jude

(2007) Spatial econometric models of cross-sectional interdependence in political science panel and time-series–cross-section data. Political Analysis 15(2): 140–164.

27.

Franzese

Robert

Hays

Jude

(2008) Interdependence in comparative politics: Substance, theory, empirics, substance. Comparative Political Studies 41(4/5): 742–780.

28.

Gaibulloev

Khusrav

Sandler

Todd

(2018) What we have learned about terrorism since 9/11. Journal of Economic Literature 57(2): 275–328.

29.

Gaibulloev

Khusrav

Piazza

James

Sandler

Todd

(2017) Regime types and terrorism. International Organization 71(3): 491–522.

30.

Ghosn

Faten

Braithwaite

Alex

Chu

Tiffany S

(2019) Violence, displacement, contact, and attitudes toward hosting refugees. Journal of Peace Research 56(1): 118–133.

31.

Gibler

Douglas

(2008) International Military Alliances, 1648–2008. Washington, DC: CQ.

32.

Gineste

Christian

Savun

Burcu

(2019) Introducing POSVAR: A dataset on refugee-related violence. Journal of Peace Research 56(1): 134–145.

33.

Gokmen

Gunes

(2017) Clash of civilizations and the impact of cultural differences on trade. Journal of Development Economics 127(1): 449–458.

34.

Greenhill

Kelly

(2010) Weapons of Mass Migration: Forced Displacement, Coercion, and Foreign Policy. Ithaca, NY: Cornell University Press.

35.

Guiso

Luigi

Sapienza

Paola

Zingales

Luigi

(2006) Does culture affect economic outcomes? Journal of Economic Perspectives 20(2): 23–48.

36.

Helbling

Marc

Leblang

David

(2019) Controlling immigration? How regulations affect migration flows. European Journal of Political Research 58(1): 248–269.

37.

Helbling

Marc

Bjerre

Liv

Römer

Friederike

Zobel

Malisa

(2017) Measuring immigration policies: The IMPIC database. European Political Science 16(1): 79–98.

38.

Hogg

Michael

(2011) Uncertainty-identity theory. In: Lange

Paul AM Van

Kruglanski

Anrie W

Tory Higgins

(eds) Handbook of Theories of Social Psychology. Thousand Oaks, CA: Sage, 62–80.

39.

Isphording

Ingo

(2014) Disadvantages of linguistic origin? Evidence from immigrant literacy scores. Economics Letters 123(2): 236–239.

40.

Kandogan

Yener

(2012) An improvement to Kogut and Singh measure of cultural distance considering the relationship among different dimensions of culture. Research in International Business and Finance 26(2): 196–203.

41.

King

Gary

Tomz

Michael

Wittenberg

Jason

(2000) Making the most of statistical analyses: Improving interpretation and presentation. American Journal of Political Science 44(2): 347–361.

42.

Kogut

Bruce

Singh

Harbir

(1988) The effect of national culture on the choice of entry mode. Journal of International Business Studies 19(3): 411–432.

43.

Krieger

Tim

Meierrieks

Daniel

(2011) What causes terrorism? Public Choice 147(1): 3–27.

44.

Kubrin

Charis

Hipp

John

Kim

Young-An

(2018) Different than the sum of its parts: Examining the unique impacts of immigrant groups on neighborhood crime rates. Journal of Quantitative Criminology 34(1): 1–36.

45.

Quan

(2005) Does democracy promote or reduce transnational terrorist incidents? Journal of Conflict Resolution 49(2): 278–297.

46.

Quan

Schaub

Drew

(2004) Economic globalization and transnational terrorism: A pooled time-series analysis. Journal of Conflict Resolution 48(2): 230–258.

47.

Lischer

Sarah

(2015) Dangerous Sanctuaries: Refugee Camps, Civil War, and the Dilemmas of Humanitarian Aid. Ithaca, NY: Cornell University Press.

48.

Lyons-Padilla

Sarah

Gelfand

Michele J

Mirahmadi

Hedieh

Farooq

Mehreen

Egmond

Marieke van

(2015) Belonging nowhere: Marginalization & radicalization risk among Muslim immigrants. Behavioral Science & Policy 1(2): 1–12.

49.

Mahfud

Yara

Badea

Constantina

Ngbala

Ahogni

(2015) Cultural distance and prejudice against immigrants in France: The role of integration models. Revue Internationale de Psychologie Sociale 28(2): 53–79.

50.

Mahfud

Yara

Badea

Constantina

Verkuyten

Maykel

Reynolds

Kate

(2018) Multiculturalism and attitudes toward immigrants: The impact of perceived cultural distance. Journal of Cross-Cultural Psychology 49(6): 945–958.

51.

Marshall

Monty

Jaggers

Keith

(2015) Polity IV Project: Political regime characteristics and transitions, 1800–2014. Users’ Manual. College Park, MD: University of Maryland.

52.

Maxwell

Rahsaan

(2010) Evaluating migrant integration: Political attitudes across generations in Europe. International Migration Review 44(1): 25–52.

53.

Nemeth

Stephen

Mauslein

Jacob

Stapley

Craig

(2014) The primacy of the local: Identifying terrorist hot spots using geographic information systems. Journal of Politics 76(2): 304–317.

54.

Neumayer

Eric

Plümper

Thomas

(2010) Galton’s problem and contagion in international terrorism along civilizational lines. Conflict Management and Peace Science 27(4): 308–325.

55.

Ousey

Graham

Kubrin

Charis

(2018) Immigration and crime: Assessing a contentious issue. Annual Review of Criminology 1(1): 63–84.

56.

Özden

Çağlar

Parsons

Christopher

Schiff

Maurice

Walmsley

Terrie

(2011) Where on earth is everybody? The evolution of global bilateral migration 1960–2000. World Bank Economic Review 25(1): 12–56.

57.

Piazza

James

(2006) Rooted in poverty? Terrorism, poor economic development, and social cleavages. Terrorism and Political Violence 18(1): 159–177.

58.

Plümper

Thomas

Neumayer

Eric

(2010) Model specification in the analysis of spatial dependence. European Journal of Political Research 49(3): 418–442.

59.

Plümper

Thomas

Troeger

Vera

Manow

Philip

(2005) Panel data analysis in comparative politics. European Journal of Political Research 44(2): 327–354.

60.

Rahimi

Sadeq

Graumans

Raissa

(2015) Reconsidering the relationship between integration and radicalization. Journal for Deradicalization 5(1): 28–62.

61.

Sageman

Marc

(2004) Understanding Terror Networks. Philadelphia, PA: University of Pennsylvania Press.

62.

Sageman

Marc

(2011) Leaderless Jihad: Terror Networks in the Twenty-First Century. Philadelphia, PA: University of Pennsylvania Press.

63.

Salehyan

Idean

(2008) The externalities of civil strife: Refugees as a source of international conflict. American Journal of Political Science 52(4): 787–801.

64.

Salehyan

Idean

Gleditsch

Kristian Skrede

(2006) Refugees and the spread of civil war. International Organization 60(2): 335–366.

65.

Savun

Burcu

Gineste

Christian

(2019) From protection to persecution: Threat environment and refugee scapegoating. Journal of Peace Research 56(1): 88–102.

66.

Schmid

Alex

(2016) Links between terrorism and migration. International Centre For Counter Terrorism (ICCT) ICCT Research Paper.

67.

Skeldon

Ronald

(2008) International migration as a tool in development policy: A passing phase? Population and Development Review 34(1): 1–18.

68.

Spolaore

Enrico

Wacziarg

Romain

(2016) War and relatedness. Review of Economics and Statistics 98(5): 925–939.

69.

UN DESA (2016) International Migration Report. New York: UN Department of Economic and Social Affairs, Population Division.

70.

Velasco González

Karina

Verkuyten

Maykel

Weesie

Jeroen

Poppe

Edwin

(2008) Prejudice towards Muslims in the Netherlands: Testing integrated threat theory. British Journal of Social Psychology 47(4): 667–685.

71.

Vidino

Lorenzo

Marone

Francesco

Entenmann

Eva

(2017) Fear Thy Neighbor: Radicalization and Jihadist Attacks in the West. Milano: Ledizioni Ledi.

72.

Vigdor

Jacob

(2015) The civic and cultural assimilation of immigrants to the United States. In: Powell

Benjamin

(ed.) The Economics of Immigration: Market-Based Approaches, Social Science, and Public Policy. Oxford: Oxford Scholarship Online.

73.

Ward

Michael D

Greenhill

Brian

Bakke

Kristin

(2010) The perils of policy by p-value: Predicting civil conflicts. Journal of Peace Research 47(4): 363–375.

74.

Weiner

Myron

(1992) Security, stability, and international migration. International Security 17(3): 91–126.

75.

Wilson

Matthew

Piazza

James

(2013) Autocracies and terrorism: Conditioning effects of authoritarian regime type on terrorist attacks. American Journal of Political Science 57(4): 941–955.

76.

Young

Joseph

Findley

Michael

(2011) Promise and pitfalls of terrorism research. International Studies Review 13(3): 411–431.

77.

Zimmermann

Doron

Rosenau

William

(2009) Introduction. In: Zimmermann

Doron

Rosenau

William

(eds) The Radicalization of Diasporas and Terrorism, Volume 80. Zurich: ETH.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB