Are Home Grown Islamic Terrorists Different? Some UK Evidence P R I F Y S G O L B A N G O R U N I V E R S I T Y Are Home Grown Islamic Terrorists Different? Some UK Evidence* By Yener Altunbas and John Thornton *Forthcoming in the Southern Economic Journal, 2011 P R I F Y S G O L B A N G O R U N I V E R S I T Y Are Home Grown Islamic Terrorists Different? Some UK Evidence • Introduction • UK homegrown terrorist acts • What makes a terrorist?• What makes a terrorist? o Conventional wisdom o Empirical evidence • Data on UK Islamic terrorists • Data on UK Muslims • Empirical results • Conclusions P R I F Y S G O L B A N G O R U N I V E R S I T Y Introduction • Shift in emphasis in terrorism literature from focus on poverty and lack of education as drivers of terrorist supply • Reflects results from survey-based and econometric studies of terrorism participationparticipation • This paper: o Focuses on UK homegrown Islamic terrorists o Uses a new dataset comparing UK Islamic terrorists with UK Muslims o Includes many demographic predictors of engaging in terrorism P R I F Y S G O L B A N G O R U N I V E R S I T Y UK Islamic Terrorist Acts • ‘Homegrown terrorist’ concept broadly defined--wide range of convictions o 2004 fertilizer plot o 2006 transatlantic aircraft terrorist plot o July 7, 2005 London Terrorist attacks o July 21 2005 London bombing o Planning Multiple bomb attacks o Conspiring to commit murder and launch radioactive or chemical attack o Conspiracy to cause explosions o Withholding necessary information P R I F Y S G O L B A N G O R U N I V E R S I T Y UK Islamic Terrorist Acts (continued) o Attempting to blow up a restaurant o Attempting to blow up a plane o Recruiting for terrorism o Inciting terrorismo Inciting terrorism o Possessing bomb making equipment o Fundraising for terrorism o Supplying military equipment to al-Qaeda training camps o Incitement to murder for terrorist purposes using the internet o Possessing details of how to fire mortar bombs and secret codes to facilitate terror attacks P R I F Y S G O L B A N G O R U N I V E R S I T Y Conventional Views on the Drivers of Terrorism • Poverty and lack of education are main drivers o Increased aid and educational assistance to end terrorism: former VP Gore (2002), President Bush (2002), Tyson (2001), Sokolsky and McMillan (2002)(2002), President Bush (2002), Tyson (2001), Sokolsky and McMillan (2002) o Economics of crime (e.g. Becker, 1968); low wage/education may lead to property crime (Freeman, 1996; Piehl, 1998) but not violent crime (Piehl, 1998; Ruhm, 2000) P R I F Y S G O L B A N G O R U N I V E R S I T Y Empirical Evidence on the Drivers of Terrorism • Evidence from 2 sources: o Survey-based profiles of terrorists: Russell and Miller (1983); Hudson and Majeska (1999); Sageman (2004); Bakker (2008); Gartenstein-Ross and Grossman (1999); Sageman (2004); Bakker (2008); Gartenstein-Ross and Grossman (2009) o Econometric-based studies: Krueger and Maleckova (2003); Berrebi (2007); Benmelech and Berrebi (2007); Krueger (2008) o Common finding: Terrorists with higher educational attainment and higher living standards are more likely to participate in terrorist activity � Exceptions: IRA-related studies by Hudson (1999); Paxson (2002); and leaked MI5 (Travis, 2008) P R I F Y S G O L B A N G O R U N I V E R S I T Y The Drivers of Terrorism Revised—Recent Theory • Seeks to explain terrorists coming from relatively privileged backgrounds in terms of a market for terror attacks (e.g. Bueno de Mesquita, 2005; Benmelech, Berrebi, and Klor, 2009; Iannaccone, 2006). o Supply function is the terrorists; demand function is organizations that select terrorists to commit attacks o Terrorist organizations want to succeed so they select the most educated terrorists o Poverty can play role—supply increases when economic conditions are poor • Summary: much survey-based and all econometric-based literature suggests terrorists not typically characterized by lack of education and poverty P R I F Y S G O L B A N G O R U N I V E R S I T Y Data on UK Homegrown Islamic Terrorists • Need predictors of individual Muslims participating in terrorism o Compare key characteristics of homegrown UK Islamic terrorists to those of UK Muslims • Terrorist sample: Muslims convicted in the UK for terrorist acts, or who died while participating in them during 2001-2009. • Sources: Gartenstein-Ross and Grossman (2009); Simcox, Stuart & Ahmed, (2010); publicly available UK Home Office documents, Wikipedia links, Google searches, press reports o We find 77 convicted Muslim terrorists with partial information on education, recent employment status, age, citizenship, sex, ethnicity, and residence o Unable to track down all the information for all the terrorists P R I F Y S G O L B A N G O R U N I V E R S I T Y Data on UK Muslims • Sample of UK Muslims (non-terrorists): Annual British Crime Survey (BCS) o Aims at measuring amount of crime in UK but contains information on religion, age, education level, citizenship, birthplace, employment, and marital status of those interviewedthose interviewed o Survey is designed to be representative of private households, and of adults aged 16 and over living in private households • Sample period: BCS surveys for 2001/02 and 2006/07, providing responses of 80,207 of which 1,363 stated that they were Muslims. • Pooled sample comprises the 1,363 Muslims that participated in the two surveys and the 77 terrorists (i.e homegrown terrorists are 5.3% of the pool) P R I F Y S G O L B A N G O R U N I V E R S I T Y Table 1 Sample averages for UK Muslims UK Muslims UK Islamic All Male terrorists p -value Education (years) 12.36 12.75 13.85 0.00 (3.46) (3.60) (2.33) Employed (percent) 48.26 65.6 37.66 0.07 Age (years) 32.84 33.12 26.35 0.00Age (years) 32.84 33.12 26.35 0.00 (9.71) (10.36) (6.26) Male (percent) 49.65 100.0 96.1 0.00 Married (percent) 57.08 55.12 37.66 0.00 UK citizen (percent) 56.48 55.2 79.22 0.00 Ethnic origin: Africa 8.64 8.19 29.9 0.00 Mixed South Asia-White 2.31 2.78 23.4 0.00 India 11.50 11.26 3.9 0.00 Pakistan 56.48 55.85 27.3 0.00 Other 10.80 11.70 3.9 0.00 Notes: Sample size is 1,440 UK Muslims, of which 77 were convicted homegrown Islamic terrorists. p -value is for test of the independence of the characteristics of homegrown Islamic terrorists versus those of all UK Muslims. Figures in parenthesis are the standard deviations of the respective sample averages. P R I F Y S G O L B A N G O R U N I V E R S I T Y 25 30 35 40 45 P er ce n t Terrorists UK Muslims 0 5 10 15 20 Graduate Undergraduate "A" Levels "O" Levels Other/none P er ce n t Figure 1. Educational attainment of homegrown Islamic terrorists versus UK Muslim population Chi-square test of independence: 12.53; p-value=0.014 P R I F Y S G O L B A N G O R U N I V E R S I T Y Table 2 Probit estimates for the likelihood of being a homegrown Islamic UK terrorist (1) (2) (3) (4) (5) (6) Mean values assigned to Weighted missing education probit Males only observations estimation Constant -1.4475*** -1.1888*** -0.1707 1.4261** -0.8866* -0.3139 (0.2641) (0.2837) (0.6082) (0.5916) (0.4840) (0.5522) Education 0.1270*** 0.1476*** 0.2009*** 0.2015*** 0.1265*** 0.2036*** (0.0281) (0.0296) (0.0434) (0.0449) (0.0330) (0.0404) Employment -0.4071** -0.5794*** -0.6501*** -0.6722*** -0.5172*** (0.1506) (0.1816) (0.1879) (0.1597) (0.1673) Age -0.0630*** -0.0567*** -0.0514*** -0.0560*** (0.0144) (0.0144) (0.0106) (0.0131) Male (1=yes) 1.6872*** 1.7589*** 1.5999*** (0.3317) (0.2777) (0.3063) Married (1=yes) 0.1706 0.1398 0.3372* 0.1100 (0.2102) (0.2194) (0.1797) (0.1928) UK citizen (1=yes) 0.5030*** 0.5413*** 0.5744** 0.5903*** 0.7192*** 0.5229** (0.1664) (0.1701) (0.2008) (0.2077) (0.1789) (0.1834) Ethnic origin Africa 0.6014** 0.6161** 0.4179 0.2436 0.6823** 0.3157 (0.2556) (0.2586) (0.3096) (0.3290) (0.2637) (0.2818) Mixed South Asia-White 0.9893*** 0.9528*** 0.5795* 0.6228* 0.7613** 0.5203* (0.2841) (0.2878) (0.3503) (0.3671) (0.3025) (0.3202) India -0.4332 -0.4037 -0.8711** -0.8884** -0.8771** -0.9096** (0.3106) (0.3164) (0.3897) (0.3998) (0.3561) (0.3602) Pakistan -0.3762* -0.4016* -0.8795** -0.8957** -0.8215*** -0.9119*** (0.2251) (0.2303) (0.2954) (0.3066) (0.2524) (0.2729) Other -0.4280 -0.5021 -0.6964 -0.7564* -0.6098* -0.7489 (0.3645) (0.3762) (0.4486) (0.4602) (0.3767) (0.4126) Pseudo R-square 0.199 0.215 0.421 0.338 0.418 0.390 Percent false negatives/positives 0.0/96.1 50.0/96.3 64.7/96.9 61.1/94.2 55.8/95.8 64.7/96.9 Chi-square 91.35 98.89 193.13 127.28 251.19 202.30 Sample size 1,416 1,416 1,416 734 1,439 1,416 Notes: Dependant variable equals 1 for convicted terrorists and 0 for others. Robust standard errors are in parenthesis. ***, **, and * indicate statistical statistical significance at the 1, 5, and 10 percent levels. The estimates reported in column 4 are for the pooled subsamble of male Muslims. The estimataes reported in column 5 assign the sample mean number of years of education to the terrorists for whom there are missing values. The estimates reported in column 6 scale the number of non-terrorists in the pool by assigning a weight to them that reflects the number of Muslims in England and Wales, as indicated in the British Crime Survey. P R I F Y S G O L B A N G O R U N I V E R S I T Y P R I F Y S G O L B A N G O R U N I V E R S I T Y Conclusions • Dealing with terrorism requires understanding the determinants of the supply of terrorists. • Results consistent with survey-based & econometric based results of drivers of terrorist participation:terrorist participation: o Muslims with higher educational attainment are somewhat more likely to participate in terrorist activities • Reinforces the wisdom of shift in emphasis in the literature from poverty and lack of education to e.g. political objectives. • Little reason to believe that additional financial/foreign aid or improving educational circumstances would help reduce the desire to participate in terrorism P R I F Y S G O L B A N G O R U N I V E R S I T Y Future research • Same (updated) UK sample, more BCS waves, more explanatory variables (local area income, local politics, local hate crimes)(local area income, local politics, local hate crimes) • Evidence from European countries o Pooling US & European data • The quality of Islamic terror: relating terrorist characteristics to terrorist acts • Economic impact of terrorist acts