Response papers

Poverty, minority economic discrimination, and domestic terrorism

Author(s): James A Piazza

Source: Journal of Peace Research , may 2011, Vol. 48, No. 3, Special Issue: New Frontiers of Terrorism Research (may 2011), pp. 339-353

Published by: Sage Publications, Ltd.

Stable URL: https://www.jstor.org/stable/23035431

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at https://about.jstor.org/terms

Sage Publications, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to Journal of Peace Research

This content downloaded from ������������132.174.250.188 on Fri, 25 Nov 2022 17:26:26 UTC������������

All use subject to https://about.jstor.org/terms

journal <>/

imcwe

Poverty, minority economic discrimination, and domestic terrorism

Journal of Peace Research 48(3) 339-353 © The Author(s) 2011 Reprints and permission: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0022343310397404

jpr.sagepub.com

(§)SAGE James A Piazza

Department of Political Science, The Pennsylvania State University

Abstract

Recognizing that the empirical literature of the past several years has produced an inconclusive picture, this study revi

sits the relationship between poverty and terrorism and suggests a new factor to explain patterns of domestic terrorism:

minority economic discrimination. Central to this study is the argument that because terrorism is not a mass phenom

enon but rather is undertaken by politically marginal actors with often narrow constituencies, the economic status of subnational groups is a crucial potential predictor of attacks. Using data from the Minorities at Risk project, I determine

that countries featuring minority group economic discrimination are significantly more likely to experience domestic terrorist attacks, whereas countries lacking minority groups or whose minorities do not face discrimination are signif

icantly less likely to experience terrorism. I also find minority economic discrimination to be a strong and substantive

predictor of domestic terrorism vis-a-vis the general level of economic development. I conclude with a discussion of the

implications of the findings for scholarship on terrorism and for counter-terrorism policy.

Keywords

discrimination, economic development, minorities, terrorism

Though it remains a popular thesis among policymakers that poverty causes terrorism, the empirical literature has been inconclusive regarding the link between socio economic factors and terrorism. Studies that use cross

national analysis to model the effects of macroeconomic

indicators on terrorism fail to show conclusively that impo

verished or underdeveloped countries experience higher rates of terrorism, or produce more terrorists, than do

1 See, for example, public statements linking poverty, poor education and unemployment to terrorism made by former Presidents Bush and

Clinton in the immediate aftermath of the 11 September 2001 terrorist attacks in the United States and British Prime Minister

Tony Blair in the wake of the 7 July 2005 suicide bomb attacks in London. More recently, in a January 2009 Stanford address, former President of Pakistan Pervez Musharraf described poverty and illiteracy as key motivators of global terrorism. Senior counter terrorism adviser to the Obama administration, John O Brennan,

poses a more nuanced relationship regarding poverty, among a host

of other factors, as a contributing factor to political grievances that

themselves propel terrorist activity. See Spencer S Hsu & Joby War rick, 2009, 'Obama's battle against terrorists to go beyond bombs and

bullets', Washington Post, 6 August.

middle or high-income countries (Abadie, 2006; Dreher & Gassebner, 2008; Krueger & Laitin, 2008; Piazza, 2006). The same has been found to be the case for regions within countries (Krueger, 2007; Piazza, 2009). Studies examining individuals likewise do not reveal a causal link between poverty, inequality, and terrorism. Empirical research has not found that terrorist perpetrators are more

likely than the average person to come from a lower socio

economic background or to be uneducated, unemployed, and economically distressed, and survey research has also not determined that economically deprived people are more likely to support terrorism (Berrebi, 2007; Fair & Haqqani, 2006; Krueger & Maleckova, 2003; Sageman, 2004).

Still, a handful of other studies prevent scholars from

confidently closing the book on the relationship between measures of economic deprivation and terrorism. Rather

than finding a consistent null result, this body of work

Corresponding author: [email protected]

This content downloaded from ������������132.174.250.188 on Fri, 25 Nov 2022 17:26:26 UTC������������

All use subject to https://about.jstor.org/terms

340 journal of Peace Research 48(3)

reveals a more complicated picture. On the one hand, Li & Schaub (2004) determine that economically developed OECD countries are less likely to experience international terrorist attacks than developing countries, while Bravo & Dias (2006) find the same for terrorism in

Eurasia. Burgoon (2006) demonstrates that social welfare spending reduces international terrorist attacks in some countries. Lai (2007) finds that countries with

higher levels of economic inequality experience higher levels of terrorism than more egalitarian societies. Bueno de Mesquita (2005) argues that the selection regimes used by terrorist movements, which favor higher socio-economic status recruits, obscure the fact that

larger pools of potential recruits are produced by poverty.

However, Blomberg & Hess (2008a) determine that economically developed countries are more likely than developing countries to experience terrorist attacks. Ross (1993) theoretically substantiates this empirical finding, noting that economically developed societies afford terrorists more targets, a greater availability of deadly weapons, and a mass transit and communication system to maximize the effectiveness of their attacks.

Yet other studies manage to find indicators of poverty

to be simultaneously negative and positive predictors of terrorism. Using dyadic analyses of source and target countries, Li (2009), Derin-Giire (2009), Blomberg & Hess (2008b), and Blomberg & Rosendorff (2006) find that increased income levels in countries reduce the

probability that their nationals will launch terrorist attacks abroad, but that countries with higher incomes, and higher levels of political democracy and economic openness, are more likely to be targeted by international terrorists. Taken together, these studies indicate a more complex relationship wherein economic underdevelop ment incubates terrorist movements and motivates them

to launch international attacks against developed coun tries because they feature developed, free media that are likely to cover attacks (Hoffman & McCormick, 2004), they are endowed with more numerous and lucrative targets (Sandler, 2005), and they are symbols of a non egalitarian status quo and a focus for political resentments (Crenshaw, 2007).

The end result is that there is little scholarly consensus

on the role that socio-economic factors play in determin

ing patterns of terrorism. This is a glaring deficit on more

than one front. First, it has contributed to a discovery lag

Though when they examine only developing countries, Blomberg & Hess (2008a) do find some evidence that indicators of economic

development programs are negatively associated with terrorism.

vis-a-vis other social science literatures on violence, such as the work on civil war (Collier, Hoeffler & Rohner, 2009; Fearon, 2008; Fearon & Laitin, 2003; Sambanis,

2004) and the fields of criminology and sociology (Fajnzylber, Lederman & Loayza, 2002; Hsieh & Pugh, 1993). Second, looking beyond to policy-oriented research, the failure of terrorism scholars to definitively

determine how, or whether, poverty and socio-economic inequality in countries precipitates terrorism handicaps evaluation ofa key element of post-September 11th United

States counter-terrorism that was inaugurated during the

Bush Administration: increased US bilateral development aid as a panacea for terrorism (Bluestein, 2002; Piazza, 2006). Coupled with similar ambiguities regarding other predictors of terrorism, this has left terrorism studies unable to articulate a clear counter-terrorism

policy recommendation.

Minority discrimination and terrorism

This study suggests that a heretofore overlooked factor may further elucidate the relationship between socio economic features of countries and the occurrence of

terrorist attacks: economic discrimination against minority groups. Though the experience of minority group discrimination has been identified as a factor that motivates and fuels terrorist campaigns in a host of quali tative studies of individual countries or individual terrorist

movements (see, for example, Bradley, 2006; Buendia, 2005; Ergil, 2000; Laqueur, 1999; O'Hearn, 1987; Van de Voorde, 2005; Whittaker, 2001), it has largely been ignored in the growing cross-national, time-series quanti tative literature investigating the root causes of terrorism.

Aside from control findings in studies focused on demo cratic rule (Eubank & Weinburg, 1994), political stability

(Lai, 2007), and national demographic composition as predictors of terrorism (Wade & Reiter, 2007), a cross national empirical investigation of minority economic status as a cause of terrorism has not been systematically

undertaken. This is striking, given the proliferation of cross-national empirical research on the causes of terrorism since 2001 (Young & Findley, 2011) and the prominence afforded to the individual experience of ethnic, racial or class discrimination as a predictor of aggressive behavior and future violent crime within the

sociology, social psychology, and criminology literatures

(Dubois et al., 2002; McCord & Ensminger, 2002; Simons et al., 2006).

There is some theoretical justification to suspect that a

causal link exists between minority economic discrimina

tion and domestic terrorist activity within countries and

This content downloaded from ������������132.174.250.188 on Fri, 25 Nov 2022 17:26:26 UTC������������

All use subject to https://about.jstor.org/terms

Piazza 341

an argument to be made that it should have a strong, substantive effect in comparison to general levels of pov erty. To establish this link, I borrow from Gurr's (1993) theory of relative deprivation, which integrates group motivations for political violence with the collective opportunities to do so. In Gurr's model, collective or social status disadvantages — when accompanied by repression on the part of the state - help to produce cohesive minority group identities within countries that

differentiate group members from larger society. These collective disadvantages, the sense of 'otherness' vis-a-vis the majority, and alienation from the state and mainstream

society facilitate the creation of long-term grievances within

afflicted subgroups. When these grievances are wedded to opportunities to mobilize, which, Gurr assumes, are condi

tioned by the size and demographic concentration of the group, political violence results. Though Gurr's model seeks to explain episodes of widespread mass political violence such as ethnic rebellions, riots, and civil wars rather than terrorist attacks, which are smaller in scale,

more sporadic, and executed by small groups rather than mass movements, I argue that his two intervening factors

in the relationship between relative deprivation and political violence — group grievance and organizational opportunity — are likewise key to understanding the cau sal link between minority economic discrimination and terrorism. I am partially assisted in this by Crenshaw (1981) and Ross (1993), both of whom argue that group grievances of marginalized subnational communities is the crucial root cause of terrorism. I add to this an argu ment that terrorist movements, as small organized actors led by elites that draw recruits from aggrieved subna tional communities, are instruments of mobilization

that allow group grievances to be channeled into violent activity.

Minority economic discrimination — which usually involves some combination of employment discrimina tion, unequal access to government health, educational or social services, formal or informal housing segregation, and lack of economic opportunities available to the rest of society — is a catalyst for the development of minority

group grievances, which are directed against the state, economic status quo, mainstream society, and the major ity population. Discrimination also reinforces social exclusion and the previously described sense of otherness

among afflicted minority group members. This leaves aggrieved minority populations alienated from the main

stream economic system, distrustful of state institutions

and authority and, thereby, more susceptible to radicali

zation and fertile ground for terrorist movements to recruit cadres, raise money, and plan and execute attacks.

Qualitative case studies of Northern Ireland (O'Hearn, 1987) and Latin America (Cleary, 2000) and survey research in Western Europe (Klausen, 2005) identify minority group experience of discrimination as a root source of minority community radicalization that is exploited by extremist movements and terrorist organiza

tions. Terrorist groups are crucial to the process here because, much like social movements or political organizations, they function as vehicles to organize and to channel minority group grievance into violent action.

In this way, they are agents of mobilization, overcoming collective action barriers that impair the larger aggrieved

minority community from acting upon their disaffection (Sandler, 2003). Discrimination also has an effect on the

'target side' of the relationship. States with aggrieved minority populations can find their counter-terrorism efforts hampered. Aggrieved communities are less likely to be cooperative with state counter-terrorism officials,

affording advantages to terrorist groups in their midst (Walsh & Piazza, 2010).

The relationship between discrimination and terrorism

can also work the other way. Societies with minority groups that do not face active economic discrimination, or where

the legacy of minority discrimination is addressed through remediation policies that level differences between minor

ity and majority populations, demonstrate that they can successfully integrate minorities into mainstream life. Minority communities in non-discriminatory societies are less likely to be radicalized or to be alienated from mainstream society, thereby making the terrorist group agenda less popular and stymieing terrorist group recruit ment. In his qualitative study of counter-terrorism responses in Northern Ireland, the Spanish Basque region, Italy against the Red Brigades, Uruguay against the Tupamaros, and Cyprus against EOKA, Hewitt (1984) credits the poor economic status of specific groups within the population, instead of the overall economic climate,

as a crucial element in fueling terrorist group recruitment and activities. In assessing the efficacy of counter-terrorism tools, Hewitt credits proactive economic

affirmative action for marginalized groups, for example education and housing subsidies of Catholics in Northern Ireland, with reducing the threat of terrorism. Minority

communities that are not aggrieved are also more likely to cooperate with state counter-terrorism officials. The qualitative counter-insurgency literature, recognizes this, noting that fostering a sense of mainstream system

legitimacy in the face of insurgent efforts to paint the

status quo as illegitimate is crucial to securing commu nity cooperation with security efforts (Hashim, 2006; Joes, 2004).

This content downloaded from ������������132.174.250.188 on Fri, 25 Nov 2022 17:26:26 UTC������������

All use subject to https://about.jstor.org/terms

342 journal of Peace Research 48(3)

Measurement issues round out the expectation that discrimination in particular is a key predictor of domestic

terrorism in countries and can help to explain the link between poverty and terrorism. A handful of scholars actually note, in asides, the problems posed by using indicators that measure nationwide socio-economic or

political statuses alone to predict the behavior of terrorist movements, which are small, narrow subnational entities

that typically operate within particular and limited geographic regions and social spaces of a country. Li extrapolates from the findings in Fearon & Laitin's (2003) study on the predictors of civil wars to depict terrorist groups as 'extremely marginal political actors' whose grievances are too narrow to be affected by main

stream political or social processes like democratic regime type or level of economic development (Li, 2005: 283). Looking outside of terrorism studies, Sambanis (2004) determines that while they are robust predictors of which

countries experience internal armed conflicts, aggregate country-level economic indicators are of little use in explaining which subgroups of citizens are likely to engage in political violence, making it very difficult

to assess the different opportunity costs for joining armed rebellions among different strata of country resi

dents. All of this highlights the value in examining more focused indicators, such as whether or not government policies or social conditions alienate subgroups from mainstream economic activity.

Hypotheses

I draw from my theoretical discussion three points that

lend themselves to empirical evaluation: that minority economic discrimination produces domestic terrorist activity; that absence of or remediation of economic discrimination suffered by minority groups reduces domestic terrorist activity; and that minority economic discrimination is an important explanatory factor for domestic terrorism alongside aggregate measures of economic development. I therefore test six hypotheses, the first two of which are:

HI: Countries with minority groups that experience economic discrimination will experience higher rates of domestic terrorism.

H2: Countries with minority groups that do not experience economic discrimination will experience lower rates of domestic terrorism.

As previously discussed, there are reasons grounded in theoretical reasoning (Crenshaw, 1981; Gurr, 1993; Ross, 1993), case studies of terrorist movements

(Bradley, 2006; Buendia, 2005; Cleary, 2000; Ergil, 2000; Klausen, 2005; Laqueur, 1999; O'Hearn, 1987; Van de Voorde, 2005; Whittaker, 2001), and some indi rect and trace cross-national empirical studies (Ellina & Moore, 1990; Eubank & Weinberg, 1994; Lai, 2007; Wade & Reiter, 2007) to expect that minority experi ence of economic discrimination might precipitate domestic terrorism. These first two hypotheses capture these expectations. It also stands to reason that if eco nomic discrimination against minorities precipitates domestic terrorism by enhancing group grievances and motivating organization, then public policies crafted to ameliorate the effects of minority economic discrimina tion should reduce domestic terrorism. Hewitt (1984)

provides some qualitative case evidence that this may be the case. Therefore, I also test the following hypothesis:

H3: Countries that have in place public policies to remediate the effects of ongoing or historical economic

discrimination against minorities will experience lower rates of domestic terrorism.

For the next two hypotheses, I retest the proposition that a country's level of economic development affects the probability that it will experience or produce terrorist

activity. My expectations about the observed relationship between level of development and terrorism are mixed,

given that while some studies find poverty to be a positive predictor of terrorism (Bravo & Dias, 2006; Burgoon, 2006; Li & Schaub, 2004) by aiding terrorist recruitment efforts, increasing public support for extre

mism, and damaging the legitimacy of the status quo, empirical findings by others (Li, 2009; Blomberg & Hess, 2008b; Blomberg & Rosendorff, 2006; Ross, 1993) suggest that wealthier countries are more likely to be targeted by terrorists because they are endowed with numerous targets, are more likely to developed free media outlets that will cover attacks, and are symbols of the political and economic status quo. Yet others find no significant relationship between level of economic development and terrorism (Krueger, 2007; Piazza, 2006). To address this controversy, I test two hypotheses:

H4: Countries with higher levels of economic develop ment will experience lower rates of domestic terrorism.

H5: Countries with higher levels of economic develop ment will experience higher rates of domestic terrorism.

Finally, motivated by Crenshaw's (1981) and Ross's (1993) discussions of permissive and precipitating root causes of terrorism, and by observations by Li (2005)

This content downloaded from ������������132.174.250.188 on Fri, 25 Nov 2022 17:26:26 UTC������������

All use subject to https://about.jstor.org/terms

Piazza 343

and Sambanis (2004) pointing out the difficulty in explaining small group activity using national indicators, I expect measurements of minority economic discrimina tion to have strong and substantive effects on patterns of terrorism in countries vis-a-vis national economic indica

tors. I therefore test my final hypothesis:

H6: Minority economic discrimination is a robust pre dictor of domestic terrorism compared to national eco nomic development indicators.

Analysis

To test these hypothesis, I use a set of zero-inflated negative binomial regression models on the incidence of domestic terrorism using a using a country-year data

base of 172 countries from 1970 to 2006. Owing to missing data for some cases, this yields a range of 2,961 to 3,088 observations, depending on the model. In all models, robust standard errors clustered on coun

try are calculated, and dispersion of observations is held

constant. My decision to use zero-inflated negative binomial estimators — rather than ordinary least squares, Poisson or standard negative binomial models — is rec ommended by several unique features of the dependent variable. First, it is an interval measurement that cannot

include negative values. Second, it is highly unevenly dis tributed across cases and years, resulting in temporal and spatial clustering with observational values that may not, in theory, be independent of one another. Finally, it con tains a large number of zero values in country-cases that

can be divided into two types: non-certain-zero types for countries that retain some probability of experiencing terrorist attacks in other observations; and certain-zero

types for countries that due to their nature do not experience terrorism at all (Brandt et al., 2000; Cameron & Trivedi, 1998; King, 1988). These elements - over dispersion and the possibility of two 'types' of zero values for the dependent variable - suggest the use of zero-inflated negative binomial techniques. This decision is buttressed by Vuong tests, included in the results, and

goodness of fit tests, published in the appendix, that recommend zero-inflated negative binomial estimations rather than negative binomial or Poisson tests. I do, how

ever, produce negative binomial and Poisson tests to check the robustness of the published, zero-inflated negative binomial estimations and find them to mirror the

core results.3 I am therefore confident that the findings of

3 See Appendix at http://www.prio.no/jpr/datasets.

the analysis are not dependent on my selection of the estimation technique.

Dependent variable The dependent variable used in the study is a country year count of domestic terrorist attacks derived from a

dataset developed by Enders, Sandler & Gaibulloev (2011). I opt to model domestic, rather than interna tional, terrorism because the literature 1 use to construct

my theoretical link between discrimination and terror ism - Gurr's (1993) relative deprivation model and Crenshaw's (1981) and Ross's (1993) group grievance models - presumes political violence is directed locally, is motivated by local conditions and involves local actors.

Empirical testing bears this out as well: identical regres sion models run on international terrorist attacks do not

demonstrate minority economic discrimination, or absence

or remediation of it, to be significant.4 Enders, Sandler & Gaibulloev (2011) derive their

count of domestic terrorist attacks occurring within countries by separating domestic from international terrorist events published in the widely used Global Terrorism Database (GTD), a publicly available, open source event-count database of aggregated domestic and international terrorist attacks from 1970 to 2008 built

and managed by the National Consortium for the Study of Terrorism and Responses to Terrorism, housed at the University of Maryland.5 Enders, Sandler & Gaibulloev undertake several steps to separate domestic and interna

tional attacks in GTD and to clean the data. They first purge the sum total of 82,536 events in GTD of doubt ful or mischaracterized attacks, eliminating approxi mately 16,000 incidents. They then use five criteria on the remaining events to sort domestic attacks - defined

4 See Appendix. Results produced using ITERATE (International Terrorism: Attributes of Terrorist Events) database published by Mickolus et al. (2009).

5 Access to the raw GTD database, along with descriptions of count methods and operationalization of terrorism, is available online at: http://www.start.umd.edu/gtd/. I wish to thank Walt Enders, Todd Sandler, and Khusrav Gaibulloev for allowing me to use their decomposed GTD database. GTD allows users to stipulate operational definition criteria for the inclusion of an event. The

Enders, Sandler & Gaibulloev (2011) decomposed terrorism dataset applies the following three criteria: For an act to be included as a terrorist event in the dataset, it must 'be aimed at a

political, economic, religious or social goal' [Criterion I] while intending to 'coerce, intimidate or convey some other message to a

larger audience' [Criterion II] while also 'including attacks against civilians but excluding attacks against military targets' [Criterion III] (Global Terrorism Database, 2009: 5).

This content downloaded from ������������132.174.250.188 on Fri, 25 Nov 2022 17:26:26 UTC������������

All use subject to https://about.jstor.org/terms

344 journal of Peace Research 48(3)

as incidents where the attack country venue matches the nationality of the perpetrators, excluding attacks on local

diplomatic targets or hostage situations involving multiple nationalities of victims — from international attacks. Enders, Sandler & Gaibulloev furthermore aid

their domestic-international decomposition technique by comparing the GTD international events with the international terrorist events published in the ITERATE database (International Terrorism: Attributes of Events)

by Mickolus et al. (2009) and making adjustments. The GTD dataset has noteworthy idiosyncrasies. For example, it used different coding procedures before 1998, and according to Enders, Sandler & Gaibulloev (2011), GTD under-counted transnational terrorist events prior to 1977 and over-counted events from 1991 to 1997. They assume that the domestic and inter national events they separate from GTD are plagued with

analogous measurement errors, and so they adjust both of them to the baseline ITERATE data. This produces a rea sonably accurate count of domestic and international GTD terrorist events

To produce the dependent variable I use in the study,

I aggregate the Enders, Sandler & Gaibulloev (2011) non-calibrated count of domestic terrorist incidents into

country-year units for the period 1970 to 2006. Multiple elements recommend a focus on domestic versus interna

tional terrorism. First, a study of predictors of domestic

terrorism stands to explain a more pervasive threat to security within countries. Abadie (2006) notes that while international terrorist attacks may generate more media attention, domestic terrorism is a far more frequent occurrence and accounts for the lion's share of all

terrorist activity in countries. Second, one can expect the

impact of minority economic discrimination on terrorist activity to be primarily manifested in domestic terrorism.

Though Enders, Sandler & Gaibulloev (2011) note that it is not unheard of for terrorist groups motivated by domestic grievances and local concerns to undertake international attacks to draw wider attention to their

goals - in the way that the FLN (National Liberation Front) of Algeria and the Palestine Liberation Organiza tion attacked international targets to highlight their national liberation struggles — this is the exception to the

rule. International attacks against third-country targets

are harder to justify to constituent audiences for terrorists

with domestic grievances, are more likely to invite third

party intervention, and are beyond the organizational and financial capacities of most local terrorist groups. For groups that have struck internationally to draw attention to their cause, international attacks remain rare

events over the course of the operational life of the

group, dwarfed in frequency by domestic attacks that directly target local assets.6

Minority economic discrimination variables To operationalize minority economic discrimination and policies to remediate discrimination in countries, I con struct a set of country-year dummy variables using the 'ECDIS/Economic Discrimination Index' variable

published by the Minorities at Risk Project (2009), housed at the Center for International Development and

Conflict Management at the University of Maryland.7 The ECDIS variable measures the degree to which members of groups designated as 'minorities at risk' (MARs) - ethnopolitical communities in countries that 'collectively suffer or benefit from systematic discrimina

tory treatment vis-a-vis other groups in society' (Minorities

at Risk Project, 2009: 1) — face economic discrimination as

a result of formal or informal governmental neglect, lack of

opportunities or social exclusion, and whether or not they are afforded affirmative remediation. ECDIS is coded in

the Minorities at Risk database as a five-point categorical measure coded in the following manner: 0 for countries

exhibiting no discrimination against minorities or for countries lacking a minority at risk group; 1 for countries

where minority groups suffer from poverty, high unem ployment and underemployment because of 'historical marginality, neglect or restrictions' but where government

policies are in place to remediate their status; 2 for countries

where minority groups face discrimination without reme dial government policies; 3 for countries where economic discrimination is due to current and ongoing social prac tices by dominant groups and where government policies either fail to remediate or are lacking; and 4 for countries

where both prevailing social practices and government policy conspire to restrict the economic wellbeing of the group (Minorities at Risk Project, 2009: 11).

The Minorities at Risk project reports data for all possible minority groups in a country and reports data by group. Most countries in the data - 123 out of 176, or 71%, constituting 66.2% of the total observa tions in the study — contain at least one designated minority at risk group, and the distribution of both

6 For example, using (non-GTD) data from Piazza (2009) it is evident that terrorist groups motivated by regime or policy change

objectives, which are assumed to be local rather than international concerns, very rarely commit international attacks: only 2.1% of attacks from these types of groups were launched against non co nationals for the period 1998 to 2006.

Data and codebook for the Minorities at Risk project can be found

at: http://www.cidcm.umd.edu/mar/data.asp.

This content downloaded from ������������132.174.250.188 on Fri, 25 Nov 2022 17:26:26 UTC������������

All use subject to https://about.jstor.org/terms

Piazza 345

MAR groups and experience of minority economic discrimination does not appear to be disproportionately featured in countries with low, or high, levels of economic development.8 To derive the dummy variables I use in this

study, I reshaped ECDIS into a country-year indicator and used a method employed by Lai (2007) and Caprioli & Trumbore (2003) whereby the highest measurement of discrimination across minority groups, if a country con tains more than one, is recorded. Observations for coun

tries containing no minorities at risk groups are also included in the analysis. I then constructed new dummy variables for each basic status that ECDIS has: 'minority economic discrimination', coded 1 for country-years indi

cating the presence of at least one Minority at Risk group

and where ECDIS has a value of 2, 3 or 4; 'MARs present but no minority economic discrimination', coded 1 for country-years containing at least one Minority at Risk group but where ECDIS has a value of zero, indicating that minorities do not suffer from economic discrimination;

'Remediation policy for minority economic discrimina tion', coded 1 for country-years containing at least one Minority at Risk group but where ECDIS has a value of 1, indicating that minorities either experience or

have a legacy of economic discrimination but where poli cies have been put into place to correct the effects discrim

ination; and finally a dummy variable titled 'No minorities

at risk present', which is simply coded 1 for observations in

countries where MAR groups are absent. It is also reason able that the effects of changes in minority group status on patterns of domestic terrorism might take time to regis

ter, so I also lag all of these MAR dummies by one period.

Indicators of economic development I use several independent variables to model the effects of

level of macroeconomic development on terrorist inci dents. Both are highly conventional (Nafeiger, 2006) and have been used to model terrorist activity in previous

studies. The first is the natural log of gross national

8 There is little to no evidence of correlation between the presence of

MAR groups and gross national income ip = -.185) or MAR groups and Human Development Index (p = -.117) or between minority experience of economic discrimination and gross national income (p - —.120) and Human Development Index (p - —.064). ' Most empirical studies of terrorism use a variant of either gross national product per capita (GNP) or gross national income (GNI) to operationalize overall level of economic development in a country. The human development index (HDI) is a more infrequently used measurement of development in empirical studies, but was found by Bravo & Dias (2006) to be a negative predictor of terrorism in Eurasian countries and to have no significant effect on terrorism by Abadie (2006).

income per capita, a commonly used indicator of a country's level of economic development, held at constant 2000 US dollars. Noting that gross national income measures only accumulation and consumption of wealth in a country as opposed to the impact of wealth

on quality of life or income inequality, I also include Human Development Index (HDI) country measures. HDI is published by the United Nations Development Program, and it combines measurements of gross national

product per capita, literacy rates, and life expectancy rates

into a single indicator intended to measure the standard of living that residents of a country enjoy. In the case of HDI, and also the Gini coefficient discussed as a covariate

below, I impute values for years in which data are missing -

both HDI and Gini are published less frequently than once

a year for some countries in the analysis — by just inserting

the most recent value. Like the minority economic discrim

ination variables, the economic development independent variables are also lagged one period in the models.

Controls

In addition, I include in all models a host of controls that

frequently appear in empirical studies of terrorism (Li, 2005; Wade & Reiter, 2007). To operationalize income inequality, I use the same measure used by Abadie (2006), Li (2005), and Li & Schaub (2004): national Gini coefficients. Derin-Gure (2009) found some evidence that

countries marked by high income inequality experienced more terrorism, so I expect Gini to be a positive predictor

of domestic terrorism. Eyerman (1998) argues that countries with large surface areas and large populations have higher policing costs and are therefore more likely to experience terrorism. I therefore include, in all models,

natural logs of national population and geographic area of all countries in the sample. Eyerman (1998) and Li (2005) also find the age of the current political regime to be a negative predictor of terrorism. I therefore control for

regime durability, which is calculated as the number of years the current regime has ruled, using data from the Polity IV project (Marshall & Jaggers, 2009). Finally, I control for political regime type using two variables: (1) political participation, which I measure by combining two individual components used in the Polity

IV index that indicate the level of free political participation

permitted by regimes — PARREG (regulation of political participation by the state) and PARCOMP (an index of the competitiveness of political participation); and (2) executive constraints, which I measure by averaging the Polity IV index components XRCOMP

This content downloaded from ������������132.174.250.188 on Fri, 25 Nov 2022 17:26:26 UTC������������

All use subject to https://about.jstor.org/terms

346 journal Peace Research 48(3)

Table I. Summary statistics

Variable Obs. Mean St. Dev. Min Max

Domestic Terrorist Incidents 3,287 8.3 32.9 0 523

Log Gross National Income per cap 3,282 7.5 1.6 3.7 14.1

Human Development Index 3,279 0.676 0.185 0.221 0.956 GINI Coefficient 3,293 42.2 9.2 23.0 84.8

Log Population 3,282 1.9 1.7 -2.8 7.1

Log Area 3,310 11.7 2.2 5.7 16.6 Durable 3,293 23.1 30.4 0 197

Political Participation 3,138 3.2 0.9 0.5 5.0 Executive Constraints 3,293 3.4 1.5 -8 6.0

Minority Economic Discrimination (Minority 3,310 0.37 0.48 0 1

at Risk Group Present, MAR ECDIS = 2, 3 or 4) MARs Present But No Minority Economic Discrimination 3,310 0.07 0.25 0 1

(Minority at Risk Group Present, MAR ECDIS = 0) Remediation Policy for Minority Economic Discrimination 3,310 0.09 0.29 0 1

(Minority at Risk Group Present, MAR ECDIS = 1) No Minorities at Risk Present 4,135 0.46 0.49 0 1

(competitiveness of executive recruitment), XROPEN (the

level of openness of the executive recruitment process), and

XCONST (the institutional constraints placed on the chief executive of the regime).101 expect political participation to

be a negative predictor of terrorism and executive con straints to be associated with higher levels of terrorism in

countries. All controls are also lagged one period in the models. Summary statistics for all variables used in the analysis are presented in Table I.

Results

For the analysis I run two sets of models, the results of which are published in Tables II and III. In the first set, I

separate the three variables measuring various aspects of minority economic discrimination into different models, along with covariates, to determine their effect on incidents of domestic terrorism in isolation from one another. The

results of these models are presented in Table II. In the sec

ond set of models, the results of which are presented in Table III, I examine the effects on terrorist incidents of

10 A full discussion of the operationalization of these variables can be

found at the Polity website: http://www.systemicpeace.org/polity/polity4.htm.

The temptation is to use the aggregate Polity score, but Vreeland (2008)

demonstrates that Polity, as well as the also commonly-used Freedom

House measures of political freedom and civil liberties, is built using indi

cators of political violence in addition to measurements of political prac

tices and institutions in countries. This would theoretically create

difficulties in interpretingresults, so herecommends individual participa tion (PARREG) and executive constraints (XCONST) instead of the

aggregate score.

having no MAR groups in a country and then include in the same model three out of the four MAR variables —

Minority Economic Discrimination, Remediation Policy for Minority Economic Discrimination, and No Minori ties at Risk Present — with MARs Present But No Eco

nomic Discrimination held out as a reference category. This permits me to see the effects of minority economic dis crimination on domestic terrorism in relation to other

MAR economic discrimination statuses. Note that for each

model reported in Tables II and III, the results of both equations run as part of the zero-inflated negative binomial

estimation technique are reported: (1) the count or non certain-zero results, which model the count of domestic ter

rorist attacks in countries retaining a probability of experi

encing domestic terrorism, thereby constituting the main

interpreted results of the study; and (2) the results of the

inflated logistical regression or certain-zero equation,

'' Adhering to convention, the zero-inflated negative binomial model results published in Tables II and III include the same covari

ates in the inflated, certain-zero equations as are in the count equa tions. However, Drakos & Gofas (2006), in their piece on underreporting bias in quantitative studies of terrorism, argue against

full specification of the inflated equation in zero-inflated negative modeling and recommend instead including only covariates associ ated with 'certain-zero' countries: regime type. They assume that certain-zero countries appear to be so in the data because they lack free media that would report on terrorist events. As a robustness

check, 1 fitted a set of zero-inflated models that include only the two

regime-type indicators in the certain-zero equations - political partic

ipation and executive constraints - and found these to produce the same results as the main published models. Results of these models are published in the Appendix.

This content downloaded from ������������132.174.250.188 on Fri, 25 Nov 2022 17:26:26 UTC������������

All use subject to https://about.jstor.org/terms

Table II. Zero-inflated negative binomial regression models for MAR economic discrimination and domestic terrorism, 1970—2006

(V (2) (3)_ ($ (V (Q_

Count Model (non-certain zero) Minority Economic Discrimination

1.301

(.221)***

1.192

(.221)***

MARs Present But No Minority Economic

-.539

(.312)*

-.653

(.299)*

Discrim. Remediation Policies for Econ. Discrim.

-.593

(.266)*

-.352

(.286)

Log Gross National Income per-cap

.298

(.087)**

.354

(.103)**

.376

(.102)***

Human Development Index

2.518

(.706)**

3.583

(.745)***

3.459

(.791)***

GINI Coefficient

.059

(.013)***

.070

(.014)***

.076

(.019)***

.088

(.018)***

.087

(.020)***

.099

(.019)***

Log Population

1.063

(.095)***

1.056

(.088)***

1.110

(.124)***

1.095

(.113)***

1.191

(.111)***

1.189

(.104)***

Log Area

-.436

(.100)***

-.433

(.098)***

-.485

(.122)***

-.479

(.115)***

-.506

(.118)***

-.508

(.115)***

Durable (Regime Age)

-.006

(.003)*

-.004

(.003)

-.006

(.004)*

-.005

(.003)

-.007

(.003)*

-.005

(.003)

Political Participation

-.285

(.135)*

-.225

(.137)

-.269

(.169)

-.195

(.167)

-.252

(.170)

-.179

(.167)

Executive Constraints

-.176

(.052)**

-.177

(.054)***

-.146

(.049)**

-.174

(.045)***

-.140

(.054)*

-.147

(.056)**

Constant

.487

(1.319)

.361

(1.352)

.743

(1.554)

.274

(1.485)

.054

(1.602)

-.259

(1.612)

Inflated. Logit (certain zero) Minority Economic Discrimination

-4.330

(1.563)**

-3.694

(.987)***

No Minority Economic Discrimination

-14.85

(1.50)***

-15.97

(2.06)***

Remediation Policies for Econ. Discrim.

.602

(1.089)

.793

(1.219)

Log Gross National Income per-cap

.615

(.417)

.309

(.208)

.374

(.270)

Human Development Index

-2.017

(3.755)

4.266

(2.760)

2.757

(3.816)

GINI Coefficient

.101

(.049)*

.053

(.047)

.018

(.036)

.022

(.036)

.076

(.049)

.085

(.064)

Log Population

-.354

(.506)

-.179

(.591)*

-.757

(.271)**

-.700

(.265)**

-.639

(.359)*

-.641

(.357)*

Log Area

-1.363

(.557)*

-1.388

(.568)*

-.276

(.245)

-.269

(.204)

-.399

(.315)

-.467

(.351)

Durable (Regime Age)

.052

(.020)**

.056

(.020)**

.008

(.006)

.008

(.006)

.012

(.010)

.016

(.010)

Political Participation

-.465

(.489)

-.192

(.351)

-.024

(.286)

.008

(.268)

.143

(.369)

.189

(.313)

Executive Constraints

-.388

(.184)*

-.257

(.124)*

-.222

(.089)*

-.328

(.138)*

-.109

(.129)

-.103

(.150)

Constant

3.880

(5.773)

11.450

(6.109)*

.404

(3.454)

-.105

(3.093)

-3.009

(3.925)

-2.108

(6.193)

Observations

2,964

2,964

2,964

2,957

2,964

2,957

Nonzero Observations

1,170

1,170

1,170

1,169

1,170

1,169

Zero Observations

1,794

1,794

1,794

1,788

1,794

1,788

Wald x2

203.96***

203.96***

140.00***

161.58***

146.13***

170.46***

Vuong z-Test 5.54*** 5.24*** 5.55*** 5.64*** 5.61*** 5.54*** Independent variables lagged one period. Robust standard errors clustered on country reported in parentheses. *p < -10; **/> < .01; ***/> < .000.

00 GN

This content downloaded from ������������132.174.250.188 on Fri, 25 Nov 2022 17:26:26 UTC������������

All use subject to https://about.jstor.org/terms

348 journal of Peace Research 48(3)

Table III. Zero-inflated negative binomial regression models for MAR economic discrimination and domestic terrorism, 1970 2006

(7) (9) (10)

Count Model (non-certain zero) Minority Economic Discrimination .932 .297)** 1.015 (.296)** Remediation Policies for Econ. Discrim. .103 .352) .402 (.348) No Minorities at Risk Groups Present -1.700 (.412)*** -1.525 (.303)*** -.923 .330)** -.726 (.329)* Log Gross National Income per-cap .328 (.195)* .277 .089)** Human Development Index 2.492 (.903)** 2.134 (.813)** GINI Coefficient .078 (020)*** .084 (.019)*** .057 .014)*** .062 (.015)*** Log Population .988 (.244)*** .969 (.144)*** 1.020 .108)*** .987 (.104)*** Log Area -.428 (.121)*** -.422 (.117)*** -.444 .098)*** -.432 (.096)*** Durable (Regime Age) -.007 (.007) -.004 (.003) -.006 .003)* -.004 (.003) Political Participation -.295 (.151)* -.249 (.148)* -.292 .135)* -.249 (.137)* Executive Constraints -.090 (.049)* -.103 (.052)* -.149 .051)** -.148 (.052)** Constant .525 (1-579) .800 (1.619) 1.263 1.348) 1.304 (1.440) Inflated Logit (certain zero) Minority Economic Discrimination 3.508 4.449) 2.832 (2.263) Remediation Policies for Econ. Discrim. 7.859 4.859) 6.782 (3.263)* No Minorities at Risk Groups Present .090 (2.761) .500 (1.119) 6.877 4.027)* 5.453 (2.864)* Log Gross National Income per-cap .088 (1.239) .521 .483) Human Development Index -.947 (3.837) -.429 (10.555) GINI Coefficient .088 (.136) .062 (.090) .066 .077) .020 (.063) Log Population -.695 (.420)* -.616 (.500) -.584 .732) -.580 (.773) Log Area -.307 (.958) -.305 (.513) -1.113 .669)* -.946 (.719) Durable (Regime Age) .019 (.011)* .020 (.009)* .042 .023)* .038 (.042) Political Participation -.177 (.787) -.063 (.352) -.507 .754) -.275 (.662) Executive Constraints -.152 (.199) -.134 (.100) -.318 .331) -.232 (.308) Constant -1.351 (6.494) .555 (7.648) -2.952 10.438) 2.903 (9.372) Observations 3,669 3,661 2,964 2 957 Nonzero Observations 1,279 1,275 1,170 1,169 Zero Observations 2,393 2,386 1,794 1,788

Wald X2 105.32*** 140.33*** 213.32*** 203.80***

Vuong z-Test 3.58*** ^ 9Q*** 5_54*** 5 24***

Independent variables lagged one period. Robust standard errors clustered on country reported in parentheses. *p < .10; **p < .01; ***p < .000. Reference category is states with Minority at Risk populations experiencing no discrimination.

which models the absence of terrorist attacks in countries

that theoretically should never experience terrorism. Because the count equation models events and the inflated

logit models absence of events, the signs of the coefficients

of the non-certain-zero and certain-zero equations are frequently opposite, or are not significant.

The main findings revealed in Table II are that minor

ity economic discrimination is a significant predictor of domestic terrorist events in countries and that absence

of and remediation of minority economic discrimination

are significant negative predictors of domestic terrorism.

The results also show that poverty is not a significant pre

dictor of domestic terrorism; on the contrary, countries

with higher levels of economic development experience

more domestic terrorism than do poorer countries. I briefly detail the specifics of the results: Across five of the six models, the three different indicators of minority economic discrimination statuses in the non-certain-zero

count equations are significant predictors of domestic terrorism in the expected direction, thereby supporting Hypotheses 1, 2, and, partially, 3, and are robust to the

inclusion of often highly significant covariates. In the count equations of Models 1 and 2, presence of minority

economic discrimination in countries is a significant, positive predictor of the likelihood that a country will experience domestic terrorism. In the count equations of Models 3 and 4, absence of economic discrimination

against minority groups is a significant negative predictor

This content downloaded from ������������132.174.250.188 on Fri, 25 Nov 2022 17:26:26 UTC������������

All use subject to https://about.jstor.org/terms

Piazza 349

of domestic terrorism, suggesting that countries that contain minority groups but that do not subject them to systematic economic disadvantages experience less domestic terrorism. Finally, the results of count equa tions in Models 5 and 6 produce mixed results. In Model 5, remediation policy for minority groups that experience or have experienced economic discrimina tion is a significant negative predictor of terrorism, but it is not found to be significant in Model 6. More information about remediation of economic discrimi

nation is provided in the next set of models. The results in Table III produce results consistent with those in Table II.

First, Models 7 and 8 show that countries that do not

contain MAR groups at all are significantly less likely to experience domestic terrorism. This is consistent with the previous finding that minority economic discrimina tion is a positive predictor of domestic terrorism, but begs the question of whether or not this relationship is overshadowed by the mere presence of sizeable minority communities in countries, regardless of their economic status. The answer to this question is found in Models 9 and 10. When the three minority economic discrimina tion variables are placed in the same models, they yield the same results as when they are run by themselves. Minority economic discrimination remains a significant positive predictor of domestic terrorism and is robust to

the inclusion of the dummy variable for absence of MAR

groups in countries, which itself remains significant and negative. In Models 9 and 10, however, remediation of minority economic discrimination is not significant at all, further eroding support for Hypothesis 3, that affir mative action policies to ameliorate minority economic discrimination are not associated with a reduction in domestic terrorist attacks.

The results of the count equations across all of the models also shed light on the perennial question of the relationship between poverty, economic development, and domestic terrorism. General level of economic devel

opment, operationalized by gross national income per capita and the Human Development Index, bears a sig nificant positive relationship with domestic terrorism across all models. This suggests that countries marked by high levels of economic development have a higher probability of experiencing domestic terrorist attacks than do poorer, less-developed countries. This finding supports the fifth hypothesis, and also confirms expecta tions by Li (2009), Blomberg & Hess (2008b), and Blomberg & Rosendorff (2006) that are consistent with the theoretical discussion by Ross (1993), that 'modernized' countries offer more targets to terrorists

and more effective means to plan, coordinate, and claim credit for terrorist attacks.

As previously noted, many of the covariates are also significant, thus increasing the robustness of the core findings. Most of these are significant in the direction expected, given the literature. The Gini coefficient is, as expected, a consistently significant, positive predictor of domestic terrorism in the count models, as is national

population. In their respective studies, Eyerman (1998) and Li (2005) found regime age (durable) and political participation to be negative predictors of terrorism, and

I mostly find the same. I also find population to be a pos itive predictor, as expected. The only two surprising findings among the controls is that area and executive constrains are significant negative predictors of terrorism

in some of the models, contradicting the findings of previous scholars. I have little in the way of explanation

for these unexpected findings and can only note that previous studies have examined the effects of area and exec utive constraints on international, rather than domestic, terrorism.

Substantive effects

To test the sixth hypothesis — that minority economic discrimination, or remediation of minority economic discrimination, is a robust factor for explaining domestic terrorism vis-a-vis aggregate level of economic develop ment in a country -1 calculate and compare the substan

tive effects of the main independent variables using Monte Carlo simulations. Table IV presents the results of these simulations.

The substantive effects portrayed in Table IV are first

difference effects of a unit change of the six main inde pendent variables on incidents of domestic terrorism per year while holding all other covariates to their appropri ate levels of measurement. Because of the different levels

of measurement of the independent variables themselves — the minority economic discrimination variables are binary categorical measures while the gross national income per capita and the Human Development Index are interval/continuous — I report the effects of standar

dized unit changes of the independent variables: changes from 0 to 1 for the minority economic discrimination variables and average quartile changes for the economic development indicators. Table IV reveals that the minor

ity economic discrimination variables have as large or larger substantive effects on domestic terrorist attacks

as aggregate economic development indicators, thus sup porting Hypothesis 6. Countries that feature economic discrimination against minority groups experience

This content downloaded from ������������132.174.250.188 on Fri, 25 Nov 2022 17:26:26 UTC������������

All use subject to https://about.jstor.org/terms

350 journal of Peace Research 48(3)

Table IV. Substantive effects, MAR economic discrimination and domestic terrorism, 1970—2006

Effects on domestic terrorism Variable Unit change attacks per year [95% Confidence interval]

Minority Economic Discrimination 0-1 +6.120 [5.153-7.182] MARs Present But No Minority Econ. Discrim. 0-1 -2.086 [-2.849- -1.209] Remediation Policies for Econ. Discrim. 0-1 -2.087 [-3.046- -0.947] No Minorities at Risk Groups Present 0-1 -4.322 [-4.987- -3.709] Log Gross National Income per-capita Quartile avg. +2.083 [1.311-3.005] Human Development Index Quartile avg. + 1.596 [1.143-2.110]

First difference substantive effects produced via Monte-Carlo simulations using Clarify (King, Tomz & Wittenberg, 2000).

around six more incidents of domestic terrorism per year,

holding all other covariates constant. This is the largest substantive effect on terrorism for all of the predictors tested in the analysis. However, the other indicators of economic discrimination have sizeable effects on domes

tic terrorism as well. Absence of minority economic dis crimination in countries that have minorities at risk

groups in their national populations, and policies aimed at remediation of past or ongoing minority economic discrimination reduce domestic terrorist attacks by 2.4 and 2 attacks per year, on average, while absence of any MAR groups in countries reduces terror attacks by 4.3 incidents per year. The effect of a country's overall

level of economic development on terrorism is also size able, but is not as large as the effect of minority economic

discrimination or absence of minority groups. For each quartile increase in the Human Development Index, countries are projected to experience only one and one-half more domestic attack per year. These results provide some empirical substantiation for Hypothesis 6: that minority group economic status is a significant and sizeable factor in predicting which countries will be plagued by domestic terrorism and that its potency as a predictor stands up well against national economic indicators.

Conclusion

There are two main conclusions produced by the study. The first is that discrimination 'matters'. The empirical results show that countries that permit their minority communities to be afflicted by economic discrimination make themselves more vulnerable to domestic terrorism

in a substantive way. The second main finding is that while aggregate poverty, or rather affluence, within soci ety does affect the amount of domestic terrorism a coun

try suffers, the overall economic status of a country has a smaller effect on terrorism than does the economic status

of a country's minority groups. There are both scholarly

and policy implications for these, albeit preliminary, findings. For scholars, these results underscore the potential limitations of relying solely on aggregate coun try indicators to evaluate which countries are most likely to experience terrorist activity. Rather, it shows that because we are seeking to explain the behavior of small groups representing often marginal subnational constitu encies, indicators of the political, economic, social, and cultural status of non-modal, subnational actors are

worthy of investigation. One can imagine several ways to apply this to future research on terrorism, but an immediate example might be the re-evaluation of regime-type indicators as predictors of terrorism versus

indicators of the status of political rights, or levels of political participation, enjoyed by minority groups within countries.

There are also potential implications for counter terrorism policy. As noted by Abadie (2006) and Piazza (2008), promotion of national economic development in poor countries and democratic and free market economic

reforms in politically and economically illiberal countries

as a means to reduce violent radicalism became a promi nent feature of US foreign policy under the Bush Admin istration. Elements of this policy framework remain in

place under President Obama (US State Department, 2009), as does the Millennium Challenge Account (MCA), created by President Bush in 2002 in the wake of the 9/11 attacks, which provides bilateral aid to impo verished countries conditioned upon their undertaking broad governance and economic restructuring programs. However, the results of this study suggest that counter

terrorism policymakers would be advised to use more spe cifically targeted measures to attack the socio-economic

roots of terrorism. We may, for example, temper our expectations that raising the US economic assistance budget for developing world countries by 50% from 2004 to 2005, or conditioning aid on reforms that improve fiscal responsibility, control inflation or liberalize

trade - all policy components of the Millennium Challenge

This content downloaded from ������������132.174.250.188 on Fri, 25 Nov 2022 17:26:26 UTC������������

All use subject to https://about.jstor.org/terms

Piazza 351

Account — will help to reduce the threat of terrorism. Instead, we might integrate other components of MCA, such as equalization of national public health and educa tion expenditures across social groups or strengthening and

universalizing the rule of law, which may more directly improve the economic status of minority and/or socially excluded and vulnerable groups - groups that if aggrieved are more likely to engage in terrorism.

Replication data The web Appendix and all replication materials for this study can be found at: http://www.prio.no/jpr/datasets.

References

Abadie, Alberto (2006) Poverty, political freedom and the roots of terrorism. American Economic Review

96(2): 159-177. Berrebi, Claude (2007) Evidence about the link between

education, poverty and terrorism among Palestinians. Peace Economics, Peace Science and Public Policy 13(1): article2.

Blomberg, S Brock & Gregory Hess (2008a) From (no) butter to guns? Understanding the economic role in transnational terrorism. In: Philip Keefer & Norman Loayza (eds) Terrorism, Economic Development and Political Openness. Cambridge: Cambridge University Press, 83-115.

Blomberg, S Brock & Gregory Hess (2008b) The Lexus and the olive branch. In: Philip Keefer & Norman Loayza (eds) Terrorism, Economic Development and Political Openness. Cambridge: Cambridge University Press, 116—147.

Blomberg, S Brock & B P Rosendorff (2006) A gravity model of globalization, democracy and transnational terrorism. Research Paper no. C06-6. University of Southern California Law School.

Bluestein, Paul (2002) Bush seeks foreign aid boost; Plan counters overseas critics. Washington Post 15(March): A10.

Bradley, John R (2006) Iran's ethnic tinderbox. Washington Quarterly 30(1): 181—190.

Brandt, Patrick T; John T Williams, Benjamin O Fordham & Brian Pollins (2000) Dynamic models for persistent event count time series. American Journal of Political Science 44(4): 823-843.

Bravo, Ana Belo Santos & Carlos Manuel Dias (2006) An empirical analysis of terrorism: Islamism and geo

political factors. Defense and Peace Economics 17(4): 329-341.

Buendia, Rizal (2005) The state—Moro armed conflict in

the Philippines: Unresolved national question or question of governance? Asian Journal of Political Science 13(1): 109—138.

Bueno de Mesquita, Ethan (2005) Conciliation, counterterrorism and patterns of terrorist violence. International Organization 59(1): 145-176.

Burgoon, Brian (2006) On welfare and terror. Journal of Conflict Resolution 50(2): 176-203.

Cameron, Adrian Colin & Pravin K Trivedi (1998) Regression Analysis of Count Data. Cambridge: Cambridge University Press.

Caprioli, Mary & Peter F Trumbore (2003) Ethnic discrimination and interstate violence: Testing the international impact of domestic behavior. Journal of Peace Research 40(1): 5—23.

Cleary, Matthew R (2000) Democracy and indigenous rebellion in Latin America. Comparative Political Studies 33(9): 1123-1153.

Collier, Paul; Anke Hoeffler & Dominic Rohner (2009)

Beyond greed and grievance: Feasibility and civil war. Oxford Economic Papers 61(1): 1-27.

Crenshaw, Martha (1981) The causes of terrorism. Com

parative Politics 13(4): 379-399. Crenshaw, Martha (2007) Terrorism and global security.

In: Chester A Crocker, Fen Osier Hampson & Pamela Aall (eds) Leashing the Dogs of War: Conflict Management in a Divided World. Washington, DC: United States Institute of Peace Press, 15-30.

Derin-Giire, Pinar (2009) Does terrorism have economic roots? Unpublished paper, Department of Economics, Boston University.

Drakos, Konstantinos & Andreas Gofas (2006) The

devil you know but are afraid to face: Underreporting bias and its distorting effects on the study of terrorism.

Journal of Conflict Resolution 50(5): 714—735. Dreher, Axel & Martin Gassebner (2008) Does proximity

to the United States cause terror? Economics Letters

99(1): 27-29. Dubois, David L; Carol Burk-Braxton, Lance P Swenson,

Heather D Tevendale & Jennifer L Hardesty (2002) Race and gender influences on adjustment in early adolescence: Investigation of an integrative model. Child Development 73: 1573-1592.

Ellina, Maro & Will H Moore (1990) Discrimination and

political violence: A cross-national study with two time

periods. Western Political Quarterly 43(2): 267-278. Enders, Walter; Todd Sandler & Khusrav Gaibulloev

(2011) Domestic versus transnational terrorism: Data, decomposition and dynamics. Journal of Peace Research 48(3): 319—338.

This content downloaded from ������������132.174.250.188 on Fri, 25 Nov 2022 17:26:26 UTC������������

All use subject to https://about.jstor.org/terms

352 journal of Peace Research 48(3)

Ergil, Dogu (2000) The Kurdish question in Turkey. Journal of Democracy 11(3): 122—135.

Eubank, William L & Leonard Weinberg (1994) Does democracy encourage terrorism? Terrorism and Political Violence 6(4): 417-435.

Eyerman, Joe (1998) Terrorism and democratic states: Soft targets or accessible systems. International Interactions 24(2): 151—170.

Fair, Christine C & Husain Haqqani (2006) Think again: Islamist terrorism. Foreign Policy January 30.

Fajnzylber, Pablo; Daniel Lederman & Norman Loayza (2002) Inequality and violent crime. Journal of Law and Economics 45(1): part 1.

Fearon, James D (2008) Economic development, insurgency and civil war. In: Elhanen Helpman (ed.) Institutions and Economic Performance. Cambridge, MA: Harvard University Press.

Fearon, James D & David D Laitin (2003) Ethnicity, insurgency and civil war. American Political Science Review 97(1): 75-90.

Gurr, Ted Robert (1993) Why minorities rebel: A global analysis of communal rebellion and conflict since 1945. International Political Science Review 14(2): 161—201.

Hashim, Ahmed S (2006) Insurgency and Counter Insurgency in Iraq. Ithaca, NY: Cornell University Press.

Hewitt, Christopher (1984) The Effectiveness of Anti-Terrorist Policies. Lanham, MD: University Press of America.

Hoffman, Bruce & Gordon McCormick (2004) Terror ism, signaling and suicide attack. Studies in Conflict and Terrorism 27(4): 243—281.

Hsieh, Ching-Chi & MD Pugh (1993) Poverty, income inequality and violent crime: A meta-analysis of recent

aggregate data studies. Criminal Justice Review 18(2): 182-202.

Joes, Anthony James (1992) Modern Guerilla Insurgency. Westport, CT: Praeger.

King, Gary (1988) Statistical models for political science event counts: Bias in conventional procedures and evi dence for the exponential Poisson regression model. American Journal of Political Science 32(3): 838-863.

King, Gary; Michael Tomz & Jason Wittenberg (2000) Making the most of statistical analyses: Improving interpretation and presentation. American Journal of Political Science 44: 341—355.

Klausen, Jytte (2005) The Islamic Challenge: Politics and Religion in Western Europe. Oxford: Oxford University Press.

Krueger, Alan B & David D Laitin (2008) Kto Kogo? A cross-country study of the origins and the targets of terrorism. In: Phillip Keefer & Norman Loayza (eds) Terrorism, Economic Development and Political Openness. Cambridge: Cambridge University Press, 148-173.

Krueger, Alan B & Jitka Maleckova (2003) Education, poverty and terrorism: Is there a causal connection? Journal of Economic Perspectives 17(4): 119—144.

Lai, Brian (2007) Draining the swamp: An empirical examination of the production of international terror ism. Conflict Management and Peace Science 24(4): 297-310.

Laqueur, Walter (1999) The New Terrorism. Oxford: Oxford University Press.

Li, Quan (2005) Does democracy produce or reduce transnational terrorist incidents? Journal of Conflict Resolution 49(2): 278-297.

Li, Quan (2009) Dyadic sources of transnational attacks. Unpublished manuscript prepared for the Terrorism and Policy Conference, University of Texas at Dallas, 21-22 May.

Li, Quan & Drew Schaub (2004) Economic globaliza tion and transnational terrorism. Journal of Conflict Resolution 48(2): 230—258.

Marshall, Monty G & Keith Jaggers (2009) Polity IV Project: Political Regime Characteristics and Transi tions, 1800—2008. Center for Systemic Peace, George Mason University (http://www.systemicpeace.org/ polity/polity4.htm). Accessed July 2009.

McCord, Joan & ME Ensminger (2002) Racial discrim ination and violence: A longitudinal perspective. In: D Hawkins (ed.) Violent Crime: Assessing Race and Ethnic Differences. New York: Cambridge University Press, 319-330.

Mickolus, Edward F; Todd Sandler, Jean M Murdock & Peter Flemming (2009) International Terrorism: Attri

butes of Terrorist Events, 1968—2007. Dunn Loring, VA: Vinyard Software.

Minorities at Risk Project (2009) Minorities at Risk Database. College Park, MD: Center for International Development and Conflict Management (http://www.cidcm.umd.edu/mar/data.aspx). Accessed July 2009.

Nafziger, Wayne (2006) Economic Development, edn. Cambridge: Cambridge University Press.

O'Hearn, Denis (1987) Catholic grievances: Comments. British Journal of Sociology 38: 94—100.

Krueger, Alan B (2007) What Makes a Terrorist? Economics and the Roots of Terrorism. Princeton, NJ: Princeton University Press.

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

This content downloaded from ������������132.174.250.188 on Fri, 25 Nov 2022 17:26:26 UTC������������

All use subject to https://about.jstor.org/terms

Piazza 353

Piazza, James A (2008) Do democracy and free markets protect us from terrorism? International Politics 45(1): 72-91.

Piazza, James A (2009) Economic development, unre solved political conflict and terrorism in India. Studies in Conflict and Terrorism 32(5): 406-419.

Ross, Jeffery Ian (1993) Structural causes of oppositional political terrorism: A causal model. Journal of Peace Research 30(3): 317-329.

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

Sambanis, Nicholas (2004) Poverty and the organization of political violence. Brookings Trade Forum, Globalization, Poverty and Inequality: 165—211.

Sandler, Todd (1995) On the relationship between democracy and terrorism. Terrorism and Political Violence 7(4): 1-9.

Sandler, Todd (2003) Collective action and transna tional terrorism. World Economy 26(6): 779—802.

Simons, Ronald L; Leslie Gordon Simons, Callie Harbin Burt, Holli Drummond, Eric Stewart,

Gene H Brody, Frederick X Gibbons & Carolyn Cutrona (2006) Supportive parenting moderates: The effect of discrimination upon anger, hostile view of relationships, and violence among African-American boys. Journal of Health and Social Behavior 47(December): 373—389.

US Department of State, Bureau for Democracy, Human Rights and Labor (2009) Advancing Freedom and Democracy Reports, May 2009 (http://www.state. gov/g/drl/rls/afdr/2009/frontmatter/122807. htm).

Van de, Voorde (2005) Sri Lankan terrorism: Assessing and responding to the threat of the Liberation Tigers of Tamil Eelam (LTTE). Policy Practice and Research 6(2): 181-199.

Vreeland, James R (2008) The effect of political regime on civil war: Unpacking anocracy. Journal of Conflict Resolution 52(3): 401—425.

Wade, Sarah Jackson & Dan Reiter (2007) Does democracy matter? Journal of Conflict Resolution 51(2): 329-348.

Walsh, James I & James A Piazza (2010) Why respecting physical integrity rights reduces terrorism. Compara tive Political Studies 43(4): 551—577.

Whittaker, David J (2001) The Terrorism Reader. New York: Routledge.

Young, Joseph & Michael Findley (2011) Promise and pitfalls of terrorism research. International Studies Review: forthcoming.

JAMES A PIAZZA, b. 1970, PhD in Politics (New York University, 1999); Associate Professor, Department of Political Science, The Pennsylvania State University (2010- ). Current research interests:

root causes of terrorism; state failure, human rights,

regime type, economic development and terrorism; terrorism in the Middle East, South Asia and Islamic

World. Research has appeared in Security Studies, Journal of Politics, International Studies Quarterly,

Comparative Political Studies, Terrorism and Political Violence, Studies in Conflict and Terrorism, and International Politics.

This content downloaded from ������������132.174.250.188 on Fri, 25 Nov 2022 17:26:26 UTC������������

All use subject to https://about.jstor.org/terms