Wagner Analiese Wagner Undergraduate Thesis Spring 2018 Economic Political, and Social Determinants of Terrorist Activity I. INTRODUCTION Security economics, as defined by Schneider et al "is understood as those activities affected by, preventing, dealing with and mitigating insecurity (including terrorism) in the economy and it "further refers to the application of economic tools to analyze the or i gins and dyn a mics of (in ) security 1 P olicy makers and counterterrorism experts around the world seek to understand the root causes of terrorist activity in order to develop strategies for comb atting and preventing terrorism at its origin This raises the important question s within security e conomics of which and to what extent economic factors contribute to terrorist activity. By analyzing the economic conditions that lead to increased terrorist activity within a nation, while also accounting for the political and social factors involved, the re is potential to better understand and combat the rise and spread of terrorism. II. SAMPLE This paper will use data from the 122 unique nations with incidents recorded in the Global Terro rism Database (GTD) and for which data on all other variables are available Note that there are over 200 nations listed in the GTD and nations were removed from consideration for one of two reasons only: 1) they no longer exist (e.g. East and West Germany) or 2) data were not available for that nation for one or more v ariables in the time range of interest. III. DEPENDENT VARIABLE Terrorist Attacks per 100,000 Population (mean of 2014 2016 ) (TERR_PER_100000 ) #$%&'()'*+",*(')*($%+"'-"./0"12%'"3$4&45($6"47"2'**4*(65".&)"849&-'* : 2'**4*(65;"<"#9*='>0?"@AAB0
Wagner @ The Global Terrorism Database (GTD) defines a terrorist attack as "the threatened or actual use of illegal force and violence by a non state actor to attain a political, economic, religious, or social goal through fear, coercion, or intimidation 2 In order to make this definition into a specific set of guideline s for classifying incidents as terrorist attacks, the GTD requires that for it to be considered a terrorist attack, the incident be "intentional," "entail some level of violence or immediate threat of violence," and "the perpetrators of the incident s must be sub national actors." 2 In addition, the incident m ust meet two of the follow criteria: "the act must be aimed at attaining a political, economic, religious, or social goal," "there must be evidence of an intention to coerce, intimidate, or convey some other message to a larger audience (or audiences) than the immediate victims," or "the action must be outside the context of legitimate warfare activities 2 T he n umber of terrorist attacks in a year per 100,000 population in a given nation will be used as a measure of the terrorist activity in that nation. Attempted attacks that are unsuccessful are counted in the GTD, but not plots or conspiracies; as a rule, attackers have to have been in the process of carrying out the attack before it was prevented in order for it to be counted For the purposes of this analysis, the value for number of terrorist attacks in a year in a nation will be taken as the mean of the number of terrorist attacks in that nation from 2014 to 2016 This will account for any large fluctuations in data caused by unusual incidents in a single year and will help align with other variables for which the data are available across this time frame. The data for population to compute the value of attacks per 100,000 population were obtained from the World Bank and the value of the population in 2016 was used. In the following analysis it will be assume d that terrorists are rational actors in the sense that they act in order to maximize their payoff and their overall utility. @ C.-(4&./"84&64*-(95"74*"-%'"#-9)>"47"2'**4*(65".&)"D'6E4&6'6"-4"2'**4*(65"F#2
Wagner R IV. INDEPENDENT VARIABLES Unemployment Rate (UNEMP_RATE) The unemployment rate provides a measure for the percentage of the available labor force that does not have a job but wants one It is a measure of the economic well being of a country and also provides a measure for the percentage of the population that has a lack of non violent activities with which to occupy their time and, therefore, may have a reduced opportunity cost of violence. Additionally, poor economic conditions create frust ration, which can make violence more likely. Unemployme nt is an economic condition which directly affect s people's lives and can cause emotional as well as economic distress, making it even more likely that disillusioned, unemployed citizens will turn to v iolence to fill their time and to give voice to their discontent. This would imply that unemployment rate is likely to have a positive effect on terrorist attacks (i.e. a higher unemployment rate is associated with more terrorist activity). It should be no ted, however, that wealth can also indicate a higher level of political participation which, in some cases, might lead to a greater likelihood of terrorist activity. Furthermore, the value being measured in this analysis is the number of attacks within a n ation, not the number of attacks perpetrated by a nation's citizens. Thus, it is also possible that if resentment or animosity toward more prosperous nations, which in some cases may also be more globally powerful nations, makes them more likely to be targ ets of terrorist attacks, t his value could be negative (i.e. a higher unemployment rate is associated with less terrorist activity). The data for unemployment rate will come from the Central Intelligence Agency World Factbook. T he data are from 2014, 2015, or 2016 d epending on the most recent data available for each nation Literacy Rate (LIT_RATE)
Wagner S Literacy rate will be used as a proxy variable to represent education levels in a nation. It is important to consider when interpreting this variable that, w hile it is true that a low literacy rate, and therefore undereducated populace, could be an indicator for the type of unrest and tension represented in the other independent variables, it is also possible that those who are more educated are more likely to be politically active. This means that a higher level of education may, in some cases, increase the terrorist activity in a nation. For this reason, the predicted outcome for literacy rate is unknown. In collecting data for this variable, it was discover ed that many developed nations do not report a literacy rate. In order to avoid excluding many developed nations from the data set, a dummy variable was used with 1 representing an above average literacy rate and 0 representing a below average literacy ra te, and with the assumption that these developed nations have an above average literacy rate. The average literacy rate was determined by using data for literacy rate from the Central Intelligence Agency World Factbook for the 102 nations for which it was available and calculating the mean The mean was 87.89 and all nations with a literacy rate of above 87.89 plus the twenty developed nations without data, were considered to have an above average literacy rate It should be noted that usually literacy rate is defined as the percentage of people age 15 an d over that can read and write, but some nations specify which language or languages a person must be able to read and write to be considered literate. The Economist Intelligence Unit's Democracy Index ( DEMOCR_INDEX) The Economist Intelligence Unit's index of democracy is a scoring of 165 independent states based on the extent to which they exhibit certain factors that indicate democracy and freedom within a nation. 3 These factors include free and fair competitive elections, the protection R T'U($+"V.W.0"12%'"3$4&45(6-"X&-'//(L'&$'"Y&(-Z6"(&)'O"47")'54$*.$>0?" !"#$%&'(')*+, 0"
Wagner [ of civil liberties, the quality of functioning of the government, democratic political culture, and political participation. Each factor is given an individual score, which is avera ged into the overall score. A higher score indicates a stricter adherence to democratic principles wi th ten as the highest score possible The most democratic country included in this analysis, according to this scale, is Norway, with a score of 9.93 and the least democratic is Turkmenistan, with a score of 1.83. The inclusion of this variable is vital in order to account for major political factor s that could affect terrorist activity in a nation. More democratic nations may provide citizens options for non violent ways for expressing their positions, but because of the protections of civil liberties in highly democratic nations, there may be fewer options for discovering and combatting terrorist activity. It is possible that nations with a middle range score (i.e. with a score of 4 6) may have the most tendency toward terrorist activity because there are both fewer avenues for expressing dissent and fewer opportunities to provide strict counter terrorism measures. Thus, it is uncertain what the influenc e of this variable will be. These data are from 2016. Fragile State s Index Indicator Refugees and Int ernally Displaced Persons (REFUG_IDPS) The Refugees and I nternally Displaced Persons (IDPs) indicator from the Fragile States Index represents the impact that the changes in population due to the influx of refugees or the migration of internally displaced persons can have on a nation. It should be noted that thi s is not meant to comment on wheth er people who are refugees are more or less likely to participate in terrorist activity than native citizens of a nation, but rather to measure how the societal pressures and tensions caused by an increased strain on resources, the mixing of differe nt cult ures and customs, and the potential introduction of new challenges, such as health issues does or does not
Wagner N affect the level of terrorist activity in a nation. The indicator also takes into account the way refugees are received, whether services and resour ces are provided to them, and how capable a nation is of receiving an influx of new persons. As with the other Fragile State Index indicators, a higher score represents a more fragile situation. It is likely that this indicator will be positively correlate d with terrorist activity because a higher score will represent more tension within a na tion as a result of the refugee or IDP situation therein. Fragile State s I ndex indicators are available from the Fund for Peace and are from 2016. Fragile State s Index Indicator External Intervention (EXT_INTERV) The External Intervention indicator measures the level of influence of external factors within a nation's internal affairs. This can manifest as political or military intervention in the state or econ omic aid or intervention of any sort. This includes intervention that is aiding or assisting the nation, as well as unfriendly attacks or covert operations. Furthermore, it takes into consideration whether a state is dependent on this outside aid. This variable is very important because, while so called "friendly" external intervention (i.e. intervention defined as "aid") is meant to help a nation, any external intervention, particularly when the receiving nation becomes dependent on the intervention, can create ill will toward the intervening government within the receiving nation or toward the nation of the receiving government for allowing the intervention It is highly likely that the consequences of aid would be different in each case, so the effect of this vari able is unknown. Note that a higher value of this indicator represents a higher level of external intervention. Fragile State s I ndex indicators are available from the Fund for Peace and are from 2016. Individuals Using the Internet (Percent of Population) (INTERNET_PERCENT)
Wagner H This measure of I nternet access an d usage (individuals using the I nterne t by a percent of population) was originally included to represent access to the incredible flow of information and instant communication provided by the I nterne t. In recent years, there has been an increase in the use of the Internet by terrorist organizations to convert people to their cause. On the other hand access to the I nternet might also represent a more informed population which is able to stay up to dat e with current events and make informed decisions for themselves, which could lead to either increased or decreased terrorist activity, depending on the person and the content they are accessing. However, when it was found to display multicollinearity with the Uneven Economic Development indicator (correlation of 0.84), it was considered a candidate for exclusion from the final regression. Regressions were run with both Internet and Uneven Economic Development and it was found that Internet had a significa nt effect on the dependent variable, while Uneven Economic Development did not. Thus, here Internet is included in the regression and is expanded to also represent by proxy the level of economic development in the nation, where a higher percentage of Inter net usage would represent a more developed nation and a smaller value would also indicate (by lack of Internet access for a larger share of the population) uneven development within that nation. A lower level of develo pment or the existence of uneve n econo mic development may be connected with a high level of economic dissatisfaction, making it easier for terrorist organizations to recruit by promoting violence as a means of changing the status quo or because those in poor economic conditions see little oppo rtunity cost associated with participating in terrorist activity. Due to the many factors to be considered related to this variable, its likely effect on terrorist activity is unknown. The data for Internet usage for 2016 will be obtained from the World Ba nk.
Wagner \ Fragile States Index Indicator Group Grievance (GROUP_GRIEV) Excluded from final regression The Fragile States Index Group Grievance indicator measures divisions amongst groups within a nation, taking into account social and political characterist ics that contribute to these divisions and how these divisions affect access to resources and services, as well as involvement in the political process. Some of these divisions are rooted in history, and the indicator also accounts for whether any specific groups are targeted, persecuted, or oppressed by state authorities or by other more powerful groups within the nation. A higher value of the Group Grievance indicator represents a more divided nation. The three nations with the highest Group Grievance sco re which were included in this analysis are Sudan (9.8), Israel (9.8), and Pakistan (9.7), and the three nations with the lowest Group Grievance score which were included in this analysis are Iceland (1.3), Ireland (1.6), and Sweden (1.7). It is likely tha t this variable w ould be positively correlated with number of terrorist attacks because the larger goals of terrorist organizations often include intimidating an opposing group, and because stronger divisions and feelings of exclusion or oppression within a nation often foster extremist sentiments and could help in the recruitment of terrorist agents. This variable was called into question because of the possibility of terrorist activity being included as a factor in the development of the indicator score. As the exact formula for the calculation of these indicators is not available, it is not entirely clear whether this is the case. However, "mass violence" is included in this score. Although it did not show a high enough correlation with terrorist attacks (correlation of 0.35) to initially cause concern, it has been eliminated from the regression in order to avoid any possible issues. A regression including the Group Grievance indicator is available in the appendix. Fragile States Index indicators are available from the Fund for Peace and are from 2016.
Wagner B GDP Per Capita (Purchasing Power Parity) (GDPC) Excluded from final regression Gross Domestic Product (GDP) is a measure of the total output of a nation's economy. In this analysis it was originally included to represent the level of economic success of a nation, as well as the standard of living of its citizens. Here, GDP per capita was chosen in order to allow for comparison between nations of vastly different siz es. GDP per capita was excluded from the final regression when terrorist attacks were adjusted from total attacks to attacks per 100,000 population to avoid dividing by the same denominator on both sides of the re gression. Unemployment i s used instead as a representation of the economic well being of the nations in the study. Fragile States Index Indicator Uneven Economic Development (UNEVEN_ECON_DEV) Excluded from final regression Originally, the Gini Index was going to be used to measure economic inequality in a nation. However, due to lack of consistent and recent data for this variable, it was replaced with the Fragile States Index indicator for Uneven Economic Development. The indicator takes into account inequality in the economy, perceived in equality in the economy, and opportunities to improve economic status. This variable was meant to s how whether tensions or unrest caused by actual or perceived high levels of inequality contribute to terrorist activity within a nation. Further, it was to b e used to represent a nation's level of economic deprivation. In situations of economic deprivation, there is a "discrepancy between what individuals think they deserve and what they actually receive in the course of the economic (distributive) process." 4 However, due to multicollinearity with the Internet usage variable (correlation of 0.84), it has been excluded from the final regression. A correlation matrix is available in the appendix. S #$%&'()'*+",*(')*($%+"'-"./0"12%'"3$4&45($6"47"2'**4*(65".&)"849&-'* : 2'**4*(65;"<"#9*='>0?"@AAB0
Wagner !A V. Summary Statistics Variable Count Mean Standard Deviation Minimum Maximum Range TERR_PER_ 100000 122 0.13 0.32 0 2.34 2.34 U N EMP_RATE 122 9.38 7.97 0.7 41.7 41 LIT_RATE 122 0.74 0.44 0 1 1 DEMOCR_ INDEX 122 5.97 2.13 1.7 9.5 7.8 REFUG_IDPS 122 4.80 2.15 1.4 10 8.6 EXT_INTERV 122 5.10 2.47 0.9 10 9.1 INTERNET_PE RCENT 122 57.03 26.38 4.32 98.24 93.92 VI. Regression and Results Regression Variable Coefficient Standard Error p value H0 (5%) Intercept 0. 50 0. 20 0.0 1 rejected UNEMP_RATE 0.002 0.00 4 0. 68 fail to reject LIT_RATE 0.04 0.0 9 0. 69 fail to reject DEMOCR_INDEX 0.004 0.02 0. 80 fail to reject REFUG_IDPS 0.03 0.02 0. 13 fail to reject EXT_INTERV 0.06 0.02 0.00 4 rejected INTERNET_PERCENT 0.005 0.002 0.007 rejected Regression Statistics R 0.4 1 R Squared 0. 17 Adjusted R Squared 0.1 3 Predicted R Squared 0.0 1 N 122 Unemployment Rate (UNEMP_RATE)
Wagner !! The variable for unemployment rate is insignificant. Because the p value is 0. 68 which is greater than 0.05, the null hypothesis that there is no linear relationship between unemployment rate and terrorist activity cannot be rejected Literacy Rate (LIT_RATE) T he variable for literacy rate i s insignificant Because the p value is 0. 69 which is greater than 0.05, the null hypothesis that there is no linear relationship between literacy rate and terrorist activity cannot be rejected The Economist Intelligence Unit's Democracy Index (DEMOCR_INDEX) The variable for The Economist Intell igence Unit's Democracy Index i s insignificant. Because the p value is 0. 80 which is greater than 0.05, the null hypothesis that there is no linear relationship between Democracy Index and terrorist activity cannot be rejected FSI Refugees a nd IDPs (REFUG_IDPS) The variable for the Fragile States Index Refugees and IDPs indicator is insi gnificant. Because the p value i s 0. 1 3 which is greater than 0.05, the null hypothesis that there is no linear relationship between the Refugees and IDPs i ndicator and terrorist activity cannot be rejected FSI External Intervention (EXT_INTERV) The Fragile States Index External Intervention indicator score has a statistically significant, positive effect on the number of terrorist attacks. Wit h a p value of 0.004 the null hypothesis that there is no linear relationship between these v ariables can be rejected with 99 % confide nce. The coefficient of 0.06 indicates that a one point increase in the External Intervention score corresponds to about 0.06 more terr orist attacks per 100,000 population Even with this result, it is difficult to say exactly why this relationship occurs. It could be that some
Wagner !@ aspect of external intervention, particularly when the intervention includes military components, incr eases or provokes terrorist activity, but it is just as likely that the conditions that cause a nation to need external intervention or aid also lead to increased terrorist activity. Internet The Internet usage in a nation also has a statistically signific ant effect on the number of terrorist attacks. The p value is 0.007, which is less than 0.01, so the null hypothesis that there is no relationship between Internet usage and terrorist activity can be rejected with 99% confidence. The coefficient is 0.005, so a one percentage point increase in Internet usage is associated with a 0.005 increase in the number of terrorist attacks per 100,000 population. This is consistent with the idea that access to Internet is correlated with opportun ities to be radicalized online and is inconsistent with the idea that lower Internet usage indicates less economic development or more inequality leading to more terrorist activity. It is important to note here that, like with unemployment rate, it is possible there is another f actor at work : nations with higher Internet usage may be more at risk to be targets of terrorist activity because of their wealth, global influence, or values. For example, western nations are often the foc us of anger from radical Islamist groups due to th eir lack of restrictions on speech and other personal freedoms, especially the freedoms of women, and these are all factors that are likely to be associated with more Internet usage or access. This is just one example of the many factors that may be includ ed in or closely associated with Internet usage; thus, it is difficult to fully interpret the results of this variable without further study. Alternate Regression Only Significant Variables
Wagner !R Alternate regressions run with only the significant variables ( External Intervention and Internet, with and without Group Grievance) show very little variation in the coefficients R squared, or adjusted R squared values VII. Conclusion This analysis provides insight into some of the economic, social, and political factors that contribute to the level of terrorist activity in a nation. However, with an adjusted R squared value of 0.13 it does not come close to explaining all of the facto rs involved in the creation and spread of terrorist organization s and activity. This is unsurprising for several reasons. First, it was assumed during thi s analysis that terrorists act rationally in order to maximize their utility and well being. However, it is highly likely that this assumption cannot be universally applied to all members of terrorist organizations In fact, many w ould argue that terrorists, particularly those engaged in attacks which they do not expect to survive, cannot be said to be usi ng rational thought in their decision making at all. The passionate feelings and emotion s involved in terrorist activity are perhaps the largest factors missing from this analysis. It is impossible to quantify the pull of religious devotion or the intense anger and hatred that often fuels the decision to commit an act of terrorism. This does not mean that the study of the factors that lead to terrorism is futile or unworthy of resources; to the contrary, any understanding that can be gained is important in fighting the long term battle against terrorism.
Wagner !S VII I Appendix i. Correlation Matrix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ii. Regression with Fragile States Index Group Grievance indicator a. Summary Statistics Group Grievance Variable Count Mean Standard Deviation Minimum Maximum Range GROUP_GRIEV 122 6.08 1.97 1.3 9.8 8.5 b. Regression Variable Coefficient Standard Error p value H0 (5%) Intercept 0. 74 0.2 2 0.0 009 rejected
Wagner ![ UNEMP_RATE 0.00 08 0.00 4 0. 83 fail to reject LIT_RATE 0.0 6 0.09 0. 47 fail to reject DEMOCR_INDEX 0.00 8 0.02 0. 61 fail to reject REFUG_IDPS 0.0 0 3 0.02 0. 87 fail to reject EXT_INTERV 0.0 5 0.02 0.0 1 rejected INTERNET_PERCENT 0.005 0.002 0.00 5 rejected GROUP_GRIEV 0.05 0.02 0.02 rejected c. Regression Statistics R 0.4 6 R Squared 0. 21 Adjusted R Squared 0.1 6 Predicted R Squared 0.0 5 N 122 d. Interpretation FSI Group Grievance (GROUP_GRIEV) When it is included in the regression, t here is a positive correlation between a nation's Fragile States Index Group Grievance indicator and the number of terrorist attacks in a nat ion. The coefficient is 0.05 indicating that a one point increase in the Group Grievance indicator score le ads to an increase of about 0.05 terro rist attacks per 100,000 population. This value is statistically significant because, with a p value of 0.0 2 the null hypothesis that there is no linear relationship between Group Grievance indicator score and terrorist activity can be rejected with 95 % c onfidence. This confirms the hypothesis that the tensions created by the factors considered in the Group Grievance score contribute significantly to the level of terrorist activity in a nation. It is important to note the potential issues noted earlier rel ated to the possible inclusion of terrorist activity in the calculation of this indicator. iii. Nations Included in the Study and the Terrorist Attacks per 100,000 Population (mean of 2014 2016 )
Wagner !N /ABCDEC % (8(: % /AFGHEC % (8(@ % /HFGDIEDC % (8((; % /HJGDEC % (8(9 % /KLIHCAEC % (8(= % /KLIHEC % (8(' % /MGHBCENCD % (8(; % OCPHCED % '8=< % OCDFACQGLP % (8'= % OGACHKL % ( % OGAFEKJ % (8(; % OPKICD % (8(= % ORAESEC % ( % ORLDEC%CDQ%TGHMGFRSEDC % (8(: % OHCMEA % (8((' % OKAFCHEC % (8(; % +CJGHRRD % (8;: % +CDCQC % (8(' % +PEAG % (8(< % +PEDC % (8((' % +RARJBEC % (8@' % +RJRHRL % ( % +RLIC%#EUC % ( % +HRCIEC % ( % +VWHKL % (8;> % +MGUP%#GWKBAEU%X+MGUPECY % (8(@ % *GDJCHZ % (8(@ % *RJEDEUCD%#GWKBAEU % (8((@ % "UKCQRH % (8((? % "FVWI % (8=9 % "A%6CASCQRH % ( % "[KCIRHECA%)KEDGC % ( % "LIRDEC % (8(9 % 5EDACDQ % (8(< % 5HCDUG % (8(= % )CBRD % ( % )GRHFEC % (8(? % )GHJCDV % (8(9 % )PCDC % (8((; % )HGGUG % (8;< % )KCIGJCAC % ( % )KEDGC % (8((> % TCEIE % (8((@ % TRDQKHCL % (8((< % TKDFCHV % (8(' % 1UGACDQ % (8'( % 1DQEC % (8(< % 1DQRDGLEC % (8(' % 1HCD % (8(' % 1HGACDQ % (89@ % 1LHCGA % '89? % 1ICAV % (8(' % \CJCEUC % (8(' % \CWCD % (8((= % \RHQCD % (8(? % ]CMCZPLICD % (8((? % ]K^CEI % (8(; % ]VHFVMLICD%X]VHFVM% #GWKBAEUY % (8(= % 0CRL%X0CR%&*#Y % (8(@ % 0CISEC % ( % 0GLRIPR % (8(; % 0EBGHEC % (8((< % 0EIPKCDEC % ( % 0K_GJBRKHF % ( % .CUGQRDEC % (8'' % .CQCFCLUCH % (8((@ % .CACVLEC % (8(= % .CAE % (89@ % .CAIC % (8'9 % .CKHEIEKL % ( % .G_EUR % (8((> % .RAQRSC % (8((: % .RDIGDGFHR % (8(9 % .RHRUUR % (8(((: % .RMCJBE[KG % (8'; %
Wagner !H -CJEBEC % ( % -GIPGHACDQL % (8(; % -G^%`GCACDQ % (8(' % -EUCHCFKC % (8(' % -EFGH % (8'' % -EFGHEC % (8@= % -RH^CV % ( % &CZELICD % (8<@ % &CDCJC % (8((> % &CWKC%-G^%)KEDGC % ( % &CHCFKCV % (8;= % &GHK % (8(@ % &PEAEWWEDGL % (8?@ % &RACDQ % (8((; % &RHIKFCA % ( % aCICH % (8(' % #GWKBAEU%Rb%IPG%+RDFR % (8(> % #RJCDEC % ( % #KLLEC%X#KLLECD% 5GQGHCIERDY % (8(@ % 6CKQE%/HCBEC % (8;9 % 6GHBEC % ( % 6EDFCWRHG % ( % 6ARSGDEC % ( % 6RKIP% /bHEUC % (8(@ % 6RKIP%]RHGC % (8((' % 6WCED % (8((? % 6HE%0CDZC % (8(= % 6KQCD % (8=' % 6KHEDCJG % ( % 6^CMEACDQ % ( % 6^GQGD % (8': % 6^EIMGHACDQ % (8((= % !CNEZELICD % (8(; % !PCEACDQ % (89( % !EJRH 7 0GLIG%X"CLI%!EJRHY % ( % !HEDEQCQ%CDQ%!RBCFR % (8(; % !KDELEC % (8'9 % !KHZGV % (8== % !KHZJGDELICD % (8((? % ,ZHCEDG % '8'> % ,DEIGQ%/HCB%"JEHCIGL % (8((< % ,DEIGQ%]EDFQRJ % (8'? % ,DEIGQ%6ICIGL % (8('@ % ,HKFKCV % (8(' % ,MBGZELICD % (8((' % 4EGIDCJ % ( % cGJGD % ;8@= % !"# $%&%'()*+,-. Global Terrorism Database: https://www.start.umd.edu/gtd/contact/ The World Bank Population: https://data.worldbank.org/indicator/SP.POP.TOTL The World Bank GD P per Capita: https://data.worldbank.org/indicator/NY.GDP.PCAP.PP.CD?name_desc=false
Wagner !\ The CIA World Factbook Unemployment Rate: https://www.cia.gov/library/publications/the world factbook/rankorder/2129rank.html The CIA World Factbook Literacy Rate: https://www.cia.gov/library/publications/the world factbook/fields/2103.html The Economist Democracy Index: https://www.eiu.com/public/to pical_report.aspx?campaignid=DemocracyIndex2016 The Fund for Peace Uneven Economic Development, Group Grievance, Refugees and IDPs, and External Intervention indicators: http://fundforpeace.org/fsi/exce l/