Foreign Fighters in Syria and Iraq and the Socio-Economic Environment They Faced at Home: A Comparison of European Countries

by Philip Verwimp


This article looks at the gap in labour market and school outcomes between non-EU immigrants and natives in European countries. It then correlates the related socio-economic data with the number of foreign fighters per million inhabitants as well as per million non-EU immigrants. It emerges that larger gaps correlate with higher numbers of foreign fighters, with Belgium emerging as the worst performer in the sample. While not offering a full, causal and micro-level model to understand the link completely, the present analysis nevertheless finds a clear and robust pattern across much of Europe.

Keywords: foreign fighters; exclusion; labour markets; PISA; Belgium; Europe; terrorism


There is a vibrant debate in the literature on the educational background and the employment status of militants who leave European countries to join the Islamic State (IS). The question to be answered is whether and to what degree we are dealing with educated, employed individuals, or with rather uneducated and often unemployed ones. The background to this question has far-reaching consequences: in case we are dealing with a group of individuals that can be situated at the lower end of the socio-economic spectrum, the policy conclusion would be that something can be done about it, to wit, offer them a descent education and a job. That will allow them to get a positive self-image, cater for their families and increase the opportunity cost to leave the country. If the opposite turns out to be the case, then policies aimed at improving the socio-economic situation of potential militants will not help much. When we are dealing with above-average educated and well-employed people, it is clear that they have much to lose by joining IS, hence the reason(s) for joining should be situated primarily in the ideological, religious or political domain.

While it is not necessary to look for one profile, or for the average profile of a foreign fighter, indeed, such average profile may miss the point completely when, for instance, a group of highly educated, radical Islamists are leading a second group of poorly educated, rank-and-file. In that situation, the average outcome ‘somewhat educated’ is next to meaningless. However, it is by detailing the individual profiles of foreign fighters that one can arrive at a better understanding of the role of (lack of) education and employment. There is not one road that leads to IS. Hence, the study of individual profiles is very important. However, this requires access to information currently only available to the intelligent services, and not to this author. Regretfully, this article will therefore have to do without such micro-level data.

In order to address the question of joining IS at the aggregate, cross-country level, two domains – the labour market and the school system – will be discussed. In essence these represent two of the most important institutions that young adults face in their encounters with “society”. The outcomes in these two domains will be linked to the number and percent of young adults who join IS from European countries. The use of aggregates does not mean that a portrait of the average foreign fighter is sketched. Rather, an attempt is made to compare labour market and educational outcomes of non-EU immigrants with those of natives (non-migrants). Since non-EU immigrants are considered as one group, however diverse it may be, this does not present a profile of foreign fighters. For this comparison, natives are also treated as one group. There will be highly educated individuals in both groups, as well as poorly educated ones. Here the focus is on the gap between both groups. Because of the lack of micro-level data it is not possible to analyze the profile of immigrants joining IS with that of those not joining IS.

For the estimates of per capita numbers of fighters joining IS, only those European countries will be considered for which comparable data on labour market and schooling outcomes between natives and non-EU immigrants could be found. For many countries in the Middle East or in Asia that have also sent fighters to Syria/Iraq, such group-level data are not available. Due to the small sample size, the interpretation of data is limited to (Western) Europe. The focus is on the potential variation between European countries. For an account of the high absolute number of foreign fighters going to Syria, the work of Hegghammer should be consulted[1] while for a history of Jihadi transnational networks, the work of Nesser is most relevant.[2]

The analysis leads to the conclusion that a larger gap between non-EU immigrant and native populations groups in the labour market and the school system in a European country is correlated with a higher per capita number of foreign fighters sent from that country. Again, this does not mean that all foreign fighters have experienced this gap themselves. It could be that those young adults who join IS are not the most disadvantaged within their group.

Review of Related Literature

As Lee has argued,[3] among the politically informed (which requires some degree of education), it is the volunteers with lower social status (drop-outs from high school for example) who may be most inclined to participate in violent activities, as compared to the high status militants who are politically active in non-violent groups, mainly because they have more to lose. Or, as Krueger and Laitin suggested,[4] a well-educated migrant may find it all the more frustrating when (s)he realizes that his group is disadvantaged and may therefore wish to act as a type of “Robin Hood”. Hence, such arguments are compatible with the gap shown below. It suffices here to document the existence of a gap and the potential for recruiters to exploit such a gap. In that sense this represents a qualified version of the grievance hypothesis–‘qualified’ in the sense that, in a cross-country approach, more marginalization and exclusion vis-à-vis the native group are correlated with a higher level of participation in IS–without implying that the poor among the immigrants constitute the bulk of the participants.

Long before the major terrorist attacks in Paris (2015) and Brussels (2016) and before the ascent of IS, social scientists and representatives from civic society in Europe pointed out the dangers arising from the backward position of immigrants in the labour market and in the school system.[5] In fact, the disadvantaged position of immigrants in these two key areas has been part of common knowledge in European policy circles for some time as well.[6] While this position has improved in some countries in recent years, a new report by the Belgium Ministry of Labour [7], still paints a dark picture. Immigrants and their offspring continue to face daunting challenges, in some countries more than in others.

Apart from the publications of the authors already mentioned, two other publications bear some similarity to the approach taken here. Since both come to different results and interpretations, a discussion of them is warranted here. The first publication is authored by McCants and Meserole [8] who apply machine learning to find out which one of a series of 40 variables correlates most with a foreign fighters score. They computed the latter by multiplying the radicalization rate (the number of foreign fighters per Muslim population) with the foreign fighters share (the share of each countries’ contingent in the total number of foreign fighters in Iraq and Syria). They argue that a policy maker should be interested in this score rather than in the two indicators that make up the score. That maybe the case for the global policy making community who want to stop IS, a national policy maker, however, will be interested first of all by the radicalization rate, as this gives him/her a measure of the extend of the problem in his/her own country. Two more comments on their very interesting study are in place here. The first one is this: among the 40 variables there is none that allows us to identify disparities or inequalities between Muslims and non-Muslims in the countries under study. In other studies, such inter-group inequality has been found to correlate with within-country violent conflict.[9] The country-level Gini coefficient, the literacy rate, or the youth unemployment rate do not inform us about that. Second comment: McCants and Meserole have 50 countries in their sample. On the one hand, this is an advantage (larger samples have more statistical power), but these countries also have very idiosyncratic histories (including, for example, Russia, Australia, Saudi-Arabia, Israel, China). Here, we run the risk that the Gini coefficient, youth unemployment or any other variable may mean very different things in these countries. Can we really assume that youth unemployment captures the same condition in each of these countries?

The second study, authored by Benmelech and Klor [10], argues that the flow of fighters to IS is not determined by poor economic conditions, but rather by ideology and by the difficulty of assimilation into homogenous Western countries. As in the case of McCants and Meserole [11], these authors use data aggregated at the national level (e.g. GDP per capita and Gini Coefficient) to conclude that poverty and inequality are unlikely to be root causes for young Muslims to join IS as Belgium, Austria, Sweden and Norway are highly developed welfare states. In the present contribution it is argued that we should not focus on national aggregates but on the differences of outcomes between Muslims and non-Muslims, or, as done here, between non-EU immigrants and natives.

What both publications have in common with the approach used in this article is the difficulty to attribute causality to one or more variables to explain the size of the foreign fighter phenomenon. All three studies also point out that they do not use/have micro-level data. Hence, it is prudent to talk about correlations, rather than causality. In that sense there is not really a contradiction in findings. All three have an element in common: the difficulties of non-EU immigrants to integrate into rich, secularized societies. Setback in the labour market and the school system, as well as the mainstream political discourse around Islam, or the homogeneity of the country may all be manifestations in different domains (economic, political, cultural) of the same underlying problem.

The sample size (N=14 countries) informing this article is too small to make causal inferences. It limits itself therefore in its interpretation to correlations between variables, realizing fully well that other, omitted variables, such as the presence of Salafist networks, preachers or recruiters, ought to be taken into account to arrive at a fuller understanding of the phenomenon. The purpose here is to observe variation across European countries, but refraining from causal and micro-level interpretations. The take here on the ‘structural’ factors can only serve as an incomplete analysis of recruitment.

Data and Methods

In search for comparative data on European countries, Eurostat, the European Union Labour Force Surveys (EU-LFS) and OECD are the first sources to turn to. The groups of persons that are of particular interest for this analysis are first and second generation non-EU immigrants, in other words, those immigrants born outside the EU as well as native-born offspring of foreign-born parents. As IS recruits from both groups, the present analysis utilizes two definitions of ‘immigrant’. Some immigrants of the second generation have obtained the nationality of the host country before leaving to Syria or Iraq, or acquired the nationality of another European country. As most recruits are in their 20s or early 30s, the focus here is on the labour market situation and educational achievements of this age group. A comparison between the situation of this group in the educational system and in the labour market with that of the native-born offspring of native-born parents is made.

The data found in the sources just cited need to be teased out. Eurostat, to start with, distinguishes between persons with and without the nationality of the country of residence, and between countries of birth. Based on that source alone, one cannot distinguish second generation immigrants from native-born offspring of native-born parents since both were born in the country of residence and (most) possess the nationality of one or another European country. Fortunately, the recent OECD report on immigrants and their integration presents data on the second generation (native-born offspring of foreign-born parents).[12] For some countries however, key variables are missing in the OECD data, thereby reducing the sample’s size from 14 to 12 countries.

Due to this, two definitions of migration will be utilized to compare data between natives and immigrants. The first is from Eurostat and EU-LFS; it defines an immigrant as someone having the nationality of a country other than a EU member state. This definition allows one to present comparative data across 14 European countries on the employment rate of nationals and non-EU nationals. This should provide an indication about the (variation in) challenges faced by immigrants on the labour markets of European countries. These challenges certainly do not disappear when an immigrant acquires the nationality of the country of residence, as discrimination is often linked to a non-European name, especially on the labour market.[13] The second definition of immigrants used is: native-born offspring of foreign-born parents. Here, data for 12 countries are available which allow a comparison of the unemployment rate for 25 to 34 year-old immigrants with native-born offspring of native-born parents. This is as close a definition one can get to locate comparable data on the target groups. It should be pointed out that these are not samples of convenience. In each case the largest group of European countries for which data were available for comparison was taken, utilizing definitions by either Eurostat or OECD.

Given the focus on the educational achievements of immigrants and their offspring, PISA test score data were also consulted. The Programme for International Student Assessment (PISA) is the largest, comparative school testing program in the world. It tests the mathematical and reading skills of a cross-section of 15-year old students in OECD countries every few years. With the help of data assembled by PISA one can compare indicators for second generation students with those of native-born students from native born parents.

To sum up: four data sources are utilized: (i) from Eurostat and from the European Union Labour Force Surveys (EU-LFS) employment data for the year 2010 are used.[14] These data sketch an approximate picture of the situation potential fighters faced on the labour market before their departure. During 2005-2008, the situation of immigrants on the labour market in Europe had improved somewhat, compared to the 2001-2004 period [15]; however, it worsened again with the 2009-2010 economic crisis; (ii) data on the employment situation of second generation immigrants, as defined by OECD, were consulted.[16] These data are available for the year 2013, which is later than one would wish, but the assumption is warranted that the gap between second generation immigrants and native-born young men and women from native-born parents did not change much between 2010 and 2013; (iii) pooled PISA test scores in mathematics for the PISA waves of 2003, 2006, 2009 and 2012 were utilized. The reasons to use these educational achievement data are the same as for the labour market: they are readily available and reflect the situation of immigrants in the school system at the moment that most Syria fighters were in school, being around 15 years old at the time. As students take the PISA test at age 15, the pooled sample should capture the situation of Syria fighters when they were school-going; and (iv) finally the number of fighters that left for Syria and Iraq per million inhabitants was calculated. A robustness test will be done by looking at the number of foreign fighters per million non-EU immigrants. It is not clear from the outset which of the two is the better denominator; arguments can be found for each of the two. The data on foreign fighters are retrieved from those collected by the International Centre for the Study of Radicalisation and Political Violence in London.[17]

Main Results

Graphs 1(a) and 1(b) show a positive correlation between the gap in employment of immigrants and natives on the one hand and the number of Syria-going fighters per million inhabitants on the other hand in a sample of European countries. Belgium is performing worst on both indicators. The Pearson correlation coefficient between both variables is 0.74*** in Graph 1(a) and -0.55* in Graph 1(b) - which is high. The first is statistically significant at the 1% level, the second at the 7% level. Since correlation is not the same as causation one should refrain from attributing the high number of Syria fighters in Belgium and elsewhere solely to their situation on the labour market. To search for one single causal factor would be imprudent for the complex foreign fighter phenomenon. Having said that, the graphs show a clear pattern across Europe.

Graph (1a): First definition of migrant (N=14)

(1b) Alternative definition ( N=12)

On average, across the 14 European Countries, the youth unemployment rate among native-born offspring of foreign born parents is 15% of the labour force whereas it is 10% for the native-born offspring of native-born parents. This let the authors of the 2015 OECD report to conclude: “In the European Union, the youth unemployment rate among native-born offspring of immigrant parents is almost 50% higher than among the young with native-born parents.” [18] Graph (1b), based on the OECD data, shows remarkable variations between European countries: from 6.3% versus 21.9% for Belgium making a difference of -15.6 and from 4.9% to 14.6% for the Netherlands, making a difference of -9.7% to Germany, with the UK and Norway showing smaller differences.

Is There a Selection Effect?

It could be that the profile of immigrants in Belgium and other under-performing countries differs from that of other countries. Immigrants in badly performing countries could, for instance, be more violent compared to immigrants in other countries, and/or, they could be less well-educated and therefore more vulnerable to experience unemployment. Social scientists and economists refer to a situation as ‘selection effect’, meaning that the bad performers would be home to a group of immigrants with particular characteristics that ill-prepare them for school and for the labour market.

(i) Selection on Violence?

The scarce literature on the attitudes towards violence mostly deals with differences between migrant communities within one country, for example those of Moroccan vs. those of Turkish descent. However, it seems very unlikely that immigrants in badly performing countries have other innate attitudes than their colleagues in other European countries, at the time of migration or at the outset. However, it could be that several other factors influence attitudes towards violence among immigrants after migration. These factors could indeed differ across Europe. One of those factors might be the recruitment of young adults by Salafist preachers. Another might be the presence of mosques spreading an ultra-conservative interpretation of the Koran and of Islam (Wahhabism). While these factors exist they are not considered here. In any event, these factor would constitute an external intervention and as such cannot be attributed to any innate characteristics of migrant populations. It also begs the question why Salafist preachers are more present in certain countries/cites than others.

(ii) Selection on the Level of Education?

The second part of a potential ‘selection effect’ lies in the level of education and its corollary, the resulting greater vulnerability to unemployment of disadvantaged immigrants. Here, there is more evidence. The OECD classifies Austria, Belgium, France, the Netherlands and Germany as longstanding destinations of lower-educated immigrants.[19] That might therefore play a role. It should, however, be noted that four of these five group members appear to belong to a ‘middle-performing’ cluster of countries in Graph 1a–not the best, but not the worst either.

As any school director knows, it is much easier to achieve superior test scores if your students were high-performers already before they entered their school. Here some evidence from PISA studies is illustrative. Graph 2 shows the correlation between the difference in test scores between immigrants and natives on the mathematics test on the one hand and the gap between the percentage of natives in employment compared to immigrants on the other hand (as also used in Graph 1). Graph 2a shows a positive correlation, larger differences in test scores correlate with larger gaps in employment. The Pearson correlation coefficient is 0.87*** with test scores of 2003 (this is when the average recruit for IS was 15 years old). The correlation coeff. is 0.67*** when we pool all math scores for the years 2003 to 2012 as they are reported in PISA in FOCUS.[20] (It should be noted that these coefficients are not reproduced in Table 1).

The result using the alternative definition of immigration is shown in Graph 2b. The Pearson corr. coeff. of -0.72**, stat. significant at the 4% level. The sample size, however, is further reduced from 12 to 9 countries in Europe as PISA 2003 data for three countries are not available. As before, Belgium performs the worst of the entire class, having the largest difference in immigrant test scores compared to natives of any PISA country. Backed by several research reports by Glorieux et al (2006, 2009), a 2012 report of the regional labour bureau of Flanders (VDAB) on the situation of immigrants on the labour market already noted: the dire situation on the labour market is a consequence of the equally dire situation in the school system”.[21]

When one performs a similar analysis at the school level, comparing PISA test scores for mathematics in schools with a high concentration of immigrants and without immigrants, the difference in Belgium is +70 points–the second highest of all PISA countries.[22] Importantly, the OECD also reports a difference after accounting for type of school and student socio-economic status. This captures the above mentioned selection effect to a great extent. In this case, the difference in test scores for Belgian schools with a high concentration of immigrants and without immigrants is reduced to +30 points. Belgium remains, however, in terms of difference the third largest in all of the PISA countries. Two countries from the group with former “longstanding migration of lower-educated” (see above), The Netherlands and Germany, manage to eliminate the difference completely after accounting for the socio-economic situation of the students and the schools.

Hence, the level of education at entry point can only partly explain the dismal performance of immigrants in Belgian schools. The other part is to be found in the school system itself, in particular for second and third generation students, who were born in Belgium and went to school there. There is much evidence that the school system in itself is a major contributor to the dismal performance of (the children of) immigrants. It caters very well to the children from the middle-class, but its curriculum, teacher training, way of testing and school culture fail to connect with the lives and aspirations of migrant children.

Graph 2a First Definition of Migrant

(2b) Second Definition of Migrant

Table 1: Pearson Corr. Coeff. between a Set of Socio-economic Indicators and the per capita Number of Foreign Fighters who left 14 European Countries for Syria/Iraq



Gap in (Un-) Employment Rate (x)

Gap in Residence in Urban Areas (b)

Gap in Rel. Poverty rate (c)

Gap in (Un-) Employment

Rate (y)

Definition 1 of migrant, N=14




0.69*** (a)

Definition 2 of migrant, N=12

- 0.55* (a’)



- 0.67** (a’)

Notes: The difference between immigrants and non-migrants is used, as the non-differentiated value of the indicator risks to capture the state of the economy in general, which is to be avoided. Here the focus is on the difference between the two groups within each economy. [*** stat.sign at 1% level, ** at 5% level, * at 10%/].
(a) The employment rate refers to the percentage of Non-EU immigrants
in employment. Source is the difference in the employment rate of individuals bearing the nationality of each of the countries and Non-EU immigrants. Source: European Union Labour Force Surveys (EU-LFS) 2010; retrieved from Kansengroepen in Kaart, VDAB, 2012, p.11. This definition as well as an alternative definition “comparing native-born children from foreign-born parents with native-born children from native-born parents (arguably better for the present purpose, as this is the group of second generation immigrants) is utilized. For this alternative data for 12 of the 14 countries (excluding Ireland and Italy) could be found [data retrieved for 2013 from OECD, 2015, p.263], which reduces an already small sample. Nevertheless, when calculation of the Pearson Corr.Coeff.[see (a’) in Table 1 for these 12 countries with the alternative definition applied to the unemployment rate of the group of 25-34 years old,] a corr. coefficient of -0.55, stat. sign. at the 6% level can be found. See the row for N=12. This is measured as difference in the two percentages.
(b) Refers to the difference in the percentage of non-migrant and migrant populations residing in cities.
(c) Refers to the difference between the relative poverty rate for natives and immigrants. Source: OECD, 2015,
(x) Here the denominator for the percentage of foreign fighters is the total population.
(y) Here the denominator for the percentage of foreign fighters is the Non-EU immigrant population.

Robustness Analysis: Using Another Denominator

One can argue that one should not use the total population as the denominator in the calculation of the number of foreign fighters per capita, but rather the number of Muslims in a country, or the number of immigrants. I do not use the number of Muslims as I have not found an OECD or Eurostat definition (meaning the same one used for each country) to identify the number of Muslims. In addition, such a denominator would also miss non-Muslims or recent converts, estimated to be around 20% of all foreign fighters. Hence, for sake of comparison, I use the total number of non-EU immigrants residing in the countries in my sample. This denominator is used to perform a robustness check for graphs (1a and b). It has also been used and I also use it in the last column of Table 1. In case one would use yet another denominator (“the number of immigrants born outside of the EU”) the correlations are confirmed with Pearson Corr.Coeff of 0.72*** for (a) in table 1 and -0.52* for (a’). Another type of robustness check is the removal of extreme values from the sample–in this case Italy and Belgium for Graph 1(a) and Spain and Belgium for Graph 1(b). This procedure, which reduces further an already small sample and is thus not a preferred procedure, results in the confirmation of the finding in Graph 1(a), with a statistically significant and positive correlation, but in doing so one loses the statistical significance for Graph 1(b). The latter is thus driven by the extremes. When, however, when performing the same check with the new denominator (the total number of non-EU immigrants) as in Graphs 3(a) and 3(b), both correlations are statistically significant.

Graphs 3a-b: Labour Market Gap and Number of Foreign Fighters, with #Non-EU Immigrants as Denominator

(3a) Robustness for Graph 1(a)

(3b): Robustness for Graph (1b)

Concluding Remarks

In its edition of November 28, 2015, The Economist argued for stopping the bashing of Belgium because we are dealing with a European-wide phenomenon.[23] That plea must have pleased the Belgian government. The Economist is correct to state that other European countries also have to deal with young adults joining IS in Syria and Iraq. As in the case of Belgium they are also drawn from immigrant populations. But Belgium is at the extreme end based on the indicators used in present analysis, namely number of fighters per capita, gap between the employment status of natives and immigrants as well as in the PISA test scores. Therefore, it is perfectly legitimate to single out Belgium: the magnitude of the problems in all but two or three other countries appears to be significantly smaller.

Like other West-European countries, Belgium has a very elaborate state-financed welfare system. All citizens have health insurance coverage, schools and universities charge no or low fees while child- and unemployment benefits as well as pension schemes are all in place. While there does not appear to be a strong correlation with the gap in relative poverty in the sample of 14 countries examined here (see Table 1), there is one strong correlation with the gap in residence in cities with concentrations of disadvantaged neighbourhoods in some European countries. Almost all foreign fighters from Belgium originate from Brussels, Vilvoorde or Antwerp. As stressed in the introduction, this does not mean that each individual recruit who headed for Syria was or needs to be unemployed for the correlation to hold. Here is what Blattman and Ralston [24] said about this is a recent extensive overview of employment and violence:

“If terrorist groups want a small cadre of highly motivated, high-performing recruits, it may make sense to recruit from the better educated and even employed. Indeed, giving up gainful employment could be a signal of ideological commitment or other non-material incentives for committed participation”.[25]

There is some evidence, however, that poor economic conditions help terror organizations recruit more able, better-educated people to participate in more complex, higher-impact terror missions. For instance, high levels of unemployment seem to have enabled Palestinian terror organizations to recruit better educated, more mature, and more experienced suicide terrorists, who in turn attacked more important Israeli targets”.[26]

The correlations do suggest that the larger the gap of immigrants with the native population (the larger the disadvantage experienced at the group level) is in a given country, the more jihadi recruits are likely to emerge from that country.

The welfare state system comes at a cost of a closed labour market, meaning a labour system that heavily protects those who are in, but makes entry for newcomers very difficult – contrary to the situation in many Anglo-Saxon countries. In Belgium, just as in the Scandinavian Countries, inequality between citizens–measured for example by the Gini coefficient–is very low. What the Gini does not show, however, is the inter-group inequality between ‘natives’ on the one hand and ‘immigrants’ or ‘non-EU nationals’ on the other hand. From the present analysis, a pattern emerges across much of Europe that links the number of Syria fighters to low labour market participation and lack of educational achievements among immigrant populations.

About the author: Philip Verwimp teaches Development Economics at Université Libre de Bruxelles where he is a Fellow of the European Centre for Advanced Research in Economics and Statistics (ECARES) as well as the Centre Emile Bernheim. Valuable comments on the first draft were received from Chris Blattman, Micael Castanheira and Yvan Guichaoua. Participants at the Foreign Fighters Conference in Oslo in September 2016 also provided insightful comments. I am particularly grateful to Petter Nesser, Thomas Hegghammer and two anonymous reviewers for comments and suggestions. All errors are the responsibility of the author only.


[1] T. Hegghammer (2013). Syria’s Foreign Fighters, Foreign Policy, December 9, .

[2] P. Nesser (2015). Islamist Terrorism in Europe, a History. London: Hurst Publishers, 2015.

[3] A. Lee (2011). Who Becomes a Terrorist?:Poverty, Education and the Origins of Political Violence, World Politics, Vol 63, No 2, April, pp.203-245.

[4] A. Krueger and D. Laitin (2008). Kto Kogo?: A Cross-Country Study of the Origins and Targets of Terrorism, In Philip Keefer and Norman Loayza (Eds.)Terrorism and Economic Development. Cambridge: Cambridge University Press, pp. 148-73.

[5] For Belgium, see N. Duquet, I. Glorieux, I. Laurijssen & Y. Van Dorsselaer (2006). Wit krijt schrijft beter. Schoolloopbanen van allochtone jongeren in beeld, Garant Uitgevers Antwerpen/Apeldoorn,106p.; I. Glorieux, I. Laurijssen & Y. Van Dorsselaer (2009). Zwart op wit. De intrede van allochtonen op de arbeidsmarkt. Garant Uitgevers, Antwerpen/Apeldoorn, 117p. For France, see C. Adida, D.Laitin, M.A.Valfort (2010), Identifying barriers to Muslim integration in France, 2010, Proceedings of the National Academy of Sciences of the United States of America 107(52): pp. 384-390; C. Adida., D.Laitin, M.A.Valfort (2014). Muslims in France: identifying a discriminatory equilibrium, Journal of Population Economics 27(4), pp. 1039-1086; C. Adida, D.Laitin, M.A.Valfort (2015), One Muslim is Enough!, Evidence from a Field Experiment in France; forthcoming in Annals of Economics and Statistics.

[6] OECD ( 2015). Indicators of Immigrant Integration, Paris.

[7] Federal Ministry of Labour of Belgium (2015). Socio-Economische Monitoring, Arbeidsmarkt en Origine, Brussels, p. 262.

[8] W. McCants and C.Messerole (2016). The French Connection: Explaining Sunni Militancy Around the World, Foreign Affairs, March 24; see also the accompanying blog post, Radicalization, Laïcité, and the Islamic Veil, April 25.

[9] F. Stewart (2008). Horizontal inequalities and conflict: understanding group violence in multiethnic societies.New York: Palgrave Macmillan

[10] E. Benmelech, and E.F. Klor (2016). What Explains the Flow of Foreign Fighters to ISIS?, Unpublished Manuscript.

[11] W. McCants and C. Messerole,op. cit, 2016, .

[12] OECD, (2015). Indicators of Immigrant Integration, Paris.

[13] See the studies of C. Adida, D. Laitin and M.A. Valfort (note 5).

[14] Eurostat (2010).[ADD DETAILS] and from the European Union Labour Force Surveys (EU-LFS).

[15] VDAB Studiedienst (2012). Kansengroepen in Kaart, 2e editie, Brussels: VDAB.

[16] OECD (2015). Indicators of Immigrant Integration, Paris.

[17] Number of Foreign Fighters in Syria and Iraq per million inhabitants (2015), International Centre for the Study of Radicalisation and Political Violence, London: Kings College. Downloaded from Radio Free Europe – Radio Liberty, retrieved on November 28, 2015; URL: .

[18] OECD (2015). Indicators of Immigrant Integration, Paris.p. 232.

[19] Ibid., p. 28.

[20] Ibid., p. 4.

[21] VDAB Studiedienst (2012). Can Schools help to, integrate immigrants? Regional Labour Bureau of Flanders, Brussels p. 2.

[23] The Economist (2015) November 28, Charlemagne: A Continent like Belgium, p. 28.

[24] C. Blattman and L.Ralston (2015). Generating employment in poor and fragile states: Evidence from labor market and entrepreneurship programs. Unpublished manuscript.

[25] Referring to A.B. Krueger and J. Maleckova (2003). Education, Poverty and Terrorism: Is There a Causal Connection?, The Journal of Economic Perspectives 17, number 4, pp.119-144..

[26] Referring to Benmelech, E., C. Berrebi, and E. F. Klor (2012). Economic conditions and the quality of suicide terrorism. The Journal of Politics, 74 (1), pp.113-128.

Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.


Perspectives on Terrorism is  a journal of the Terrorism Research Initiative and the Center for Terrorism and Security Studies

ISSN  2334-3745 (Online)

Disclaimer, Terms and Conditions