Variables Used in Analysis Many variables are not available for all countries over the entire sample period. As such, in all instances, each variable used in the study is the sample average over the available dates. Domestic Terrorism: This variable corresponds to the average number of domestic terrorist incidents within a country over the 1970?1992 and 1994–2010 periods, respectively. In some places in the analysis, the terrorism variable indicates the cumulative number of attacks during the respective periods. We use the decomposition from the ESG (2011) data set. In addition, we use only incidents with at least one casualty (i.e., murdered or injured individual). For all domestic incidents, the nationalities of the perpetrators and victims match the location of the incident. Transnational Terrorism: For the GTD and ITERATE data, this variable corresponds to the average number of transnational terrorist incidents that occurs within a country during the 1970?1992 and 1994–2010 periods, respectively. In some places in the analysis, the terrorism variable indicates the cumulative number of attacks during the respective periods. We use only incidents with at least one casualty. For the GTD data, we employ the decomposition from the ESG data set. For the ITERATE data, we can measure terrorism by either the location of the incident or the nationality of the perpetrator. For all transnational incidents, the location, nationalities of the perpetrators, and nationalities of the victims are not all the same. Freedom House and POLITY: The Freedom House (2012) and POLITY (Marshall and Jaggers 2012) variables measure political freedom. The Freedom House variable measures political rights and civil liberties on a scale from 1 to 7, with lower values indicating more freedom. If the sum of these two indices is less than or equal to 5, the country is deemed to have a high degree of political and civil freedom. In this case, we set the dummy variable for Freedom House equal to one; otherwise, we set the dummy variable equal to zero. Our POLITY measure is a dummy variable equal to zero if the overall POLITY score is less than 7, and equal to unity otherwise. Higher POLITY values indicate more democratic principles characterize the country. Rule of Law: This variable indicates the strength and impartiality of the legal system along with the popular observance of the law. This is an index that varies between 0 and 6, with higher values indicating more respect for law and order. The data source is International Country Risk Guide 2012 http://www.prsgroup.com/icrg.aspx. Ethnic Tension and Religious Tension: Each of these indices varies from 0 to 6, where smaller values indicate more tension. Ethnic Tension refers to racial, nationality, or language divisions, while Religious Tension indicates religious suppression and persecution. The data source is International Country Risk Guide 2012 http://www.prsgroup.com/icrg.aspx. The remaining variables are from the World Bank’s (2012) World Development Indicators. The annual values for each variable are averaged over two time periods: 1970–1992 and 1994–2010. Not all of the years are available for some countries (as shown in Table A2). In such circumstances, we averaged all of the available years. Area: The area of country i. Education: World Banks’ (2012) measure of the number of individuals receiving secondary education. (Indicator code: SE.SEC.ENRL) Gini Coefficient: measurement of income inequality. Larger values reflect greater income inequality. (Indicator code: SI.POV.GINI). lgdpi: The natural logarithm of real per capita GDP in country i (in constant 2000 US dollars). Unfortunately, real per capita GDP figures are not available for Afghanistan. lpopi: The natural logarithm of population in country i. Unemployment: The overall unemployment rate in a country as a percent of total labor force. Table A2: Availability of the Raw Data Variable No. of countries Education (1970–1992) 163 Education (1994–2010) 172 Per capita GDP (1970–1992) 158 Per capita GDP (1994–2010) 167 Ethnic Tension (1984–1992) 126 Ethnic Tension (1994–2010) 138 Gini (1970–1992) 73 Gini (1994–2010) 142 Population (1970–1992) 172 Population (1994–2010) 172 Religious Tension (1984–1992) 126 Religious Tension (1994–2010) 138 Unemployment (1970–1992) 113 Unemployment (1993–2010) 148 The first eight variables on the file ALLDATACOMBINED.XLS are: Series Obs Mean Std Error Minimum Maximum D_LOC1 166 0.38 0.79 0.00 6.55 D_LOC2 166 0.32 0.86 0.00 8.06 C_LOC1 166 0.63 1.38 0.00 10.68 C_LOC2 166 0.41 0.99 0.00 8.94 D_NAT1 166 0.22 0.45 0.00 3.14 D_NAT2 166 0.19 0.46 0.00 3.17 C_NAT1 166 0.33 0.69 0.00 4.73 C_NAT2 166 0.23 0.53 0.00 3.44 These variables are from the ITERATE data set The preface D stands for death incidents and C stands for casualty indidents The suffix 1 stands for the first sample period and 2 stands for the second sample period LOC refers to incidents bt locatin and NAT refers to incidents classified by nationality The next eight variables are from the GTD data set Series Obs Mean Std Error Minimum Maximum DDEATH1 166 2.40 8.07 0.00 57.18 DDEATH2 166 2.66 10.06 0.00 94.59 TDEATH1 166 0.37 0.83 0.00 5.23 TDEATH2 166 0.31 1.14 0.00 12.60 DCAS1 166 3.20 9.78 0.00 67.41 DCAS2 166 3.78 13.22 0.00 119.90 TCAS1 166 0.59 1.28 0.00 8.56 TCAS2 166 0.45 1.38 0.00 14.66 The preface D refers to domestic and T to transnational The suffix 1 stands for the first sample period and 2 stands for the second sample period The remaining variables are Series Obs Mean Std Error Minimum Maximum MIDEA 166 0.11 0.32 0.00 1.00 EDU1 163 1628928.85 5339881.81 1250.67 52967283.67 EDU2 172 2717287.07 9350489.05 2644.78 83621455.67 GDP1 158 5122.31 7753.83 129.81 45791.68 GDP2 167 7423.87 10793.68 135.90 56867.07 GINI1 72 38.51 11.62 19.40 61.33 GINI2 142 40.73 9.21 24.70 64.30 POP1 172 25986905.68 96740636.81 17827.39 997235217.39 POP2 172 35936309.75 131848231.40 45513.94 1278276111.11 AREA 172 722652.17 1939304.19 28.00 16381145.56 UNEMP1 113 8.91 6.88 0.50 43.50 UNEMPL2 148 9.37 7.50 0.60 59.50 RT1 121 4.30 1.40 0.90 6.00 RT2 133 4.68 1.18 1.06 6.00 ET1 121 3.54 1.52 0.14 6.00 ET2 133 4.17 1.20 0.89 6.00 LAW1 122 3.16 1.55 0.58 6.00 LAW2 135 3.91 1.26 1.06 6.00 FH_PR1 162 4.16 1.96 1.00 7.00 FH_PR2 166 3.57 2.09 1.00 7.00 FH_CL1 162 4.10 1.73 1.00 6.90 FH_CL2 166 3.59 1.75 1.00 7.00 POLITY1 153 0.25 0.42 0.00 1.00 POLITY2 154 0.44 0.47 0.00 1.00 The suffix 1 stands for the first sample period and 2 stands for the second sample period All variables are defined in detail in the online supplement. MIDEA is a Mideast dummy variable EDU refers to educational level GDP is the GDP variable GINI is the ini coefficient POP is polulation Area is area unemp1 is the unemployment rate RT is religious tension ET is ethnic tension LAW is the Rule of Law FH_PR and FH_CL are the Freedom House measures POLITY is the POLITY measure