Simple question

What is the central research question of this article?
Why is that question difficult to answer/why is the answer not immediately obvious?
What do the author(s) do to answer that question?

Requirements: very short 2-3 sentences for each question

For each article, answer those questions.

DISCUSSION PAPER SERIES
IZA DP No. 14970
Income and Terrorism:
Insights from Subnational Data
Michael Jetter
Rafat Mahmood
David Stadelmann
DECEMBER 2021
DISCUSSION PAPER SERIES
IZA DP No. 14970
Income and Terrorism:
Insights from Subnational Data
Michael Jetter
University of Western Australia and IZA
Rafat Mahmood
University of Western Australia and Pakistan Institute of Development Economics
David Stadelmann
Universität Bayreuth
DECEMBER 2021
Any opinions expressed in this paper are those of the author(s) and not those of IZA. Research published in this series may
include views on policy, but IZA takes no institutional policy positions. The IZA research network is committed to the IZA
Guiding Principles of Research Integrity.
The IZA Institute of Labor Economics is an independent economic research institute that conducts research in labor economics
and offers evidence-based policy advice on labor market issues. Supported by the Deutsche Post Foundation, IZA runs the
world’s largest network of economists, whose research aims to provide answers to the global labor market challenges of our
time. Our key objective is to build bridges between academic research, policymakers and society.
IZA Discussion Papers often represent preliminary work and are circulated to encourage discussion. Citation of such a paper
should account for its provisional character. A revised version may be available directly from the author.
IZA – Institute of Labor Economics
Schaumburg-Lippe-Straße 5–9
53113 Bonn, Germany
Phone: +49-228-3894-0
Email: publications@iza.org
www.iza.org
IZA DP No. 14970
DECEMBER 2021
ABSTRACT
Income and Terrorism:
Insights from Subnational Data
To better understand potential relationships between income and terrorism, we study data
for 1,527 subnational regions in 75 countries between 1970 and 2014. Results consistently
imply an inverted U-shape that remains robust to accounting for a comprehensive set of
region-level covariates, region- and time-fixed effects, as well as estimating an array of
alternative specifications. The threat of terrorism systematically rises as low-income polities
become richer, peaking at an income level of about US$12,800 per capita (in constant
2005 PPP US$), but then falls consistently above that level. This pattern emerges for
domestic and transnational terrorism alike. Peaks in the income-terrorism relationship differ
by perpetrator ideology. Thus, alleviating poverty per se may first exacerbate terrorism,
contrary to much of the proposed recipes advocated since 9/11.
JEL Classification:
D74, O11
Keywords:
subnational income, subnational terrorism, domestic terrorism,
transnational terrorism, terror group ideology
Corresponding author:
Michael Jetter
University of Western Australia
8716 Hackett Drive
Crawley 6009, WA
Australia
E-mail: mjetter7@gmail.com
“We won’t win the war against terror without addressing the problem of poverty.”
(Wolfensohn, then-President of the World Bank, 2002).
1
Introduction
In the aftermath of the 9/11 attacks twenty years ago, US President George W. Bush, US
Secretary of State John Kerry, British Prime Minister Tony Blair, along with other prominent
politicians, policymakers, and commentators explicitly linked terrorism to poverty (Bush, 2002;
Krueger, 2007; Sterman, 2015; Easterly, 2016).
However, cross-country research has produced ambiguous and sometimes contradictory evidence for a potential relationship between income and terrorism. Table 1 summarizes the
corresponding quantitative literature, illustrating the substantial uncertainty of whether and, if
so, how income connects with terrorism. In that branch of research, aggregating variables at
the national level to then explore systematic relationships with indicators of terrorism has been
common, largely because of data availability and convention.
In the following pages, we propose that our understanding of the income-terrorism nexus
sharpens substantially once we zoom in to the subnational level, i.e., studying Balochistan, California, Catalonia, and Île-de-France instead of Pakistan, the United States, Spain, and France.
Two basic observations motivate this refocus. First, terror attacks often cluster regionally within
a country, rather than being spread out uniformly. For example, in the United Kingdom from
1970 to 2014, we identify striking di↵erences between Northern Ireland (1,544 attacks) and the
North (four attacks). Similarly, while the Chilean O’Higgins region was completely spared of
terror attacks over that entire time period, the metropolitan region of Santiago su↵ered 1,612
attacks, ranking the region fourth worldwide. And second, income levels across regions within
a country often di↵er more than incomes across countries. For example, the average income
of Moscow exceeds the average income of Sicily, even though Italy is on average approximately
three times richer than Russia. Such substantial within-country heterogeneities are lost when
studying country-level aggregates.
Our approach matches subnational (regional) data on GDP/capita (from Gennaioli et al.,
2014) with subnational data on terror attacks (from START, 2017) for 1,527 regions across 75
1
countries between 1970 and 2014. These sample countries are statistically representative of the
global relationship between income and terrorism. Our unit of analysis constitutes the secondlargest administrative unit in the respective nation, i.e., a federal state, county, or province,
depending on the country. Our main estimation results and interpretations hold constant potential confounders associated with (i) population size, (ii) regions hosting a country’s capital
city, (iii) oil production, (iv) period-fixed, and (v) region-fixed e↵ects. Region-fixed e↵ects prove
particularly powerful as they account for unobservable time-invariant di↵erences across regions,
such as geographical attributes that often correlate with terrorist activity (e.g., mountainous
terrain or ruggedness) and unique histories of ethnic and religious conflict or colonization experiences. These fixed e↵ects also reasonably control for certain societal and environmental
aspects that only change slowly over time within a given region, such as fractionalization and
polarization along ethnic or religious dimensions.
Our empirical results lend firm support to a nonlinear relationship between income and
terrorism that follows an inverted U-shape. This is consistent with the cross-country findings
by Enders and Hoover (2012) and Enders et al. (2016) who posit that low-income polities
lack the resources terrorist organizations need, while high-income polities can a↵ord e↵ective
counterterrorism measures. Our findings suggest that, as incomes in poor regions increase,
terrorism becomes substantially more likely until an estimated peak of approximately US$12,800
(in constant 2005 PPP US$). For perspective, 63% of all observations in our sample would fall
under that threshold. After that, economic growth is associated with a decline in terrorism.
Importantly, we find this nonlinear pattern for domestic and international terrorism alike. Our
analysis helps reconciling the di↵erent findings of Table 1.
We also look into the ideologies of perpetrators to explore whether di↵erent types of terrorism
follow di↵erent income-related patterns. Illustrating the generality of our main findings, the
inverted U-shape independently emerges for all identifiable ideologies with (i) Islamist, (ii) leftwing, (iii) right-wing, (iv) separatist, and (v) other religious groups. Interestingly, religious
terrorism peaks at income levels that are lower than those for left- or right-wing terrorism – a
relationship that was proposed by Enders et al. (2016) but, to our knowledge, remained untested
since. The consistency with which this pattern emerges across regions around the world for
over 45 years suggests a systematic inverted U-shape link between income and terrorism that
2
Table 1: Overview of the quantitative literature linking GDP/capita to terrorism (based on
Gosling, 2017).
Statistically insignificant
Statistically significant
negative
Statistically significant
positive
Abadie (2006)
Azam and Delacroix (2006)
Blomberg and Hess (2008b,a)a
Basuchoudhary and Shughart (2010)
Azam and Thelen (2008)
Berman and Laitin (2008)
Blomberg and Hess (2008b)
Campos and Gassebner (2013)
Braithwaite and Li (2007)
b
Blomberg and Rosendor↵ (2006)
a
Burgoon (2006)b
Eyerman (1998)
Crenshaw et al. (2007)
Bravo and Dias (2006)
Dreher and Fischer (2010, 2011)
Li (2005)
Kurrild-Klitgaard et al. (2006)
Gassebner and Luechinger (2011)
Li and Schaub (2004)
Neumayer and Plümper (2009)
Testas (2004)
Piazza (2007, 2008a, 2011)c,b
Goldstein (2005)
b
Koch and Cranmer (2007)
Krueger and Laitin (2008)
Plümper and Neumayer (2010)
Krueger and Malečková (2003)
Walsh and Piazza (2010)
Piazza (2006, 2008b)
b
Tavares (2004)
Sambanis (2008)
Notes: a Blomberg and Hess (2008b) find a negative (positive) association with ‘low (lower) income’ countries.
b
GDP/capita constitutes one component of a composite indicator, such as the Human Development Index or the
Government Capability Index.
3
transcends time, ideology, and space.
Overall, our study contributes to a wider understanding of terrorism determinants, while
particularly informing the debate on the link between income and terrorism. We combine existing
data sources at subnational levels to introduce an integrated database that allows us to gain
more refined insights into the problem. Beyond terrorism, this paper also informs the literature
on the impact of economic growth on non-economic variables, as well as the benefits and costs
associated with that development process (e.g., see Bloom and Canning, 2000, Friedman, 2010,
and Gürlük, 2009).
Section 2 begins by positioning the theoretical backgrounds on income and terrorism. Section
3 introduces our data and sources, followed by our methodology in Section 4. Section 5 details
our empirical findings, and Section 6 concludes.
2
Theoretical Background
The political and scholarly debate that followed 9/11 inextricably linked poverty to terrorism
(Pilgrim, 2015; Odede, 2015; Haggar, 2021). The underlying hypothesis is grounded in existing
work on civil conflict (Abadie, 2006), civil war (Collier and Hoe✏er, 2004; Miguel et al., 2004),
and political coups (Alesina et al., 1996). As another form of political violence, terrorism has
been suggested to follow a similar logic: Poverty brings grievances that may motivate terrorism
(Piazza, 2007).
Nevertheless, two decades after 9/11, the corresponding empirical evidence remains inconclusive. Cross-country studies have produced negative, positive, and null results – an artefact we
illustrate in Table 1. Similarly, individual-level studies have failed to establish a systematic correlation between poverty and terrorism (Hassan, 2002; Krueger and Malečková, 2003; Sageman,
2004; Berrebi, 2007; and Benmelech et al., 2012).
Theoretically, the inconclusive link between income and terrorism may be owed to an incomplete functional form that conceals nonlinearities (Enders and Hoover, 2012; Enders et al.,
2016). While very low-income polities do not o↵er sufficient human and monetary resources
to support terrorism, high-income societies may be able to employ e↵ective counterterrorism
strategies (Lai, 2007; Enders et al., 2016). From a sociological perspective, Maslow’s (1943)
4
hierarchy of needs implies political and societal prospects only gain relevance once basic physiological needs are met. Thus, ideological and political considerations may not constitute primary
objectives in impoverished societies, i.e., political violence in the form of terrorism could play less
of a role. Also, economic grievances are less likely to arise in richer countries where governments
can leverage more substantial funds to address concerns of their citizenry (Lai, 2007).
Consequently, ceteris paribus, terrorism, whether domestic or transnational, may peak at
medium incomes. A handful of cross-country studies support this perspective (Lai, 2007; Freytag
et al., 2011; De la Calle and Sánchez-Cuenca, 2012). Further, Enders et al. (2016) suggest the
peak of terrorism may have changed over time, owing to the shift from left-wing ideologies that
were concentrated in relatively wealthy countries to religious fundamentalists that predominantly
live in the developing world. We will also explore this hypothesis and provide evidence for it
using our regional data.
3
Data
3.1
Subnational Income Levels
We derive data on region-level income from Gennaioli et al. (2014) who record real GDP/capita
(in constant 2005 PPP US$) in five-year intervals for subnational units in a global sample.1
As comprehensive data on terrorism start in 1970, we consider observations from 1970 to 2010,
producing a maximum of nine observations per region and an average of six observations per
region. Table 2 documents summary statistics of all variables in our main analysis, while Table
A1 summarizes the variables used in additional analyses. Table A2 shows full data coverage for
each country and year.
Consistent with the literature, we employ the natural logarithm of GDP/capita (e.g., see
Freytag et al., 2011, Enders and Hoover, 2012, Enders et al., 2016, and Krieger and Meierrieks,
2019). Using GDP/capita levels (sans logarithm) instead, produces consistent results (see Table
A5). To allow for nonlinearities, we follow Enders and Hoover (2012) and Enders et al. (2016)
to incorporate a squared term of that variable.
1
Gennaioli et al. (2014) collect data on subnational population and income levels primarily from national
statistics agencies. Data are scaled such that the population-weighted sum of subnational GDP equates to national
GDP recorded in the Penn World Tables or, when unreported there, in the World Development Indicators.
5
Table 2: Summary Statistics for main variables at the subnational (regional) level for 1,527
regions (n=8,383 for all variables). Variables in Panel A come from Gennaioli et al.
(2014), while variables in Panel B come from START (2017).
Variable
Mean
(Std. Dev.)
Min.
(Max.)
Description
Ln(GDP/capita)i,t
12,429
(12,334)
189
(166,007)
GDP/capita in 2005 PPP US$ (we apply
the natural logarithm)
Population size (in thousands)i,t
2,823
(8,367)
10
(196,243)
Population (we apply
the natural logarithm)
Capitali
0.05
(0.22)
0
(1)
=1 if hosts country’s capital
Oili,t
10.26
(20.56)
0
(89.38)
Cumulative oil and gas
production (per capita; we
apply the natural logarithm)
Terror attacksi,t,…,t+4
7.40
(46.50)
0
(1,479)
# of terror attacks in t, .., t + 4
Domestic attacksi,t,…,t+4
6.27
(43.15)
0
(1,461)
# of non-transnational terror
attacks in t, .., t + 4
Transnational attacksi,t,…,t+4
1.13
(10.90)
0
(705)
# of transnational terror
attacks in t, .., t + 4
Islamist attacksi,t,…,t+4
0.46
(13.73)
0
(1,136)
# of terror attacks by Islamist
groups in t, .., t + 4
Leftist attacksi,t,…,t+4
2.73
(22.74)
0
(987)
# of terror attacks by Leftist
groups in t, .., t + 4
Rightist attacksi,t,…,t+4
0.13
(1.21)
0
(44)
# of terror attacks by right-wing
groups in t, .., t + 4
Separatist attacksi,t,…,t+4
1.59
(19.29)
0
(1,230)
# of terror attacks by separatist
groups in t, .., t + 4
Religious non-Islamist attacksi,t,…,t+4
0.40
(8.74)
0
(674)
# of terror attacks by religious,
non-Islamist groups in t, .., t + 4
Panel A: Independent variables
Panel B: Dependent variables
6
Figure 1 visualizes the global coverage of our sample. African regions remain under-represented
with notable omissions including Iraq and Afghanistan – two of the countries most a↵ected by
terrorism. As such selection issues may threaten the generalizability of our findings, we carefully compare global country-level results for all years with those from studying our sample
countries and years. These estimations produce consistent coefficients, which suggests that our
interpretation is unlikely to su↵er from misrepresentation issues (see Table A3).
Figure 1: Regional Sample Coverage.
3.2
Subnational Terrorism
For data on terrorism, we employ the well-known Global Terrorism Database (GTD). Accessing
information on the location of each terror attack allows us to assign each attack to a particular
within-country region. Appendix B explains this procedure in detail. We then aggregate attacks
over five-year intervals and merge the data with Gennaioli et al.’s (2014) data. For example,
GDP/capita for Catalonia in 1970 is matched with terror attacks in Catalonia between 1970
and 1974.
Our main dependent variable measures the number of terror attacks, which constitutes the
7
most commonly employed measure in the literature. Additional estimations distinguish between
domestic and transnational attacks.2 Figure 2 plots GDP/capita against the number of terror
attacks. Panel A considers all terrorism, while Panels B and C distinguish between domestic
and transnational terrorism. Although these graphs do not incorporate potentially confounding
factors yet, they do imply a nonlinear relationship between regional income and terrorism in the
form of an inverted U-shape.
3.3
Further Covariates
Our estimations include a list of region-level covariates that may independently be associated
with terrorism. Following the literature, we incorporate population size, oil production (to control for resource-curse-related dynamics; see Tavares, 2004, and Sambanis, 2008), and a binary
indicator for hosting the nation’s capital (because of a potential concentration of cultural, political, and religious targets).3 As the data on educational attainment feature several missing values
in our sample period, we do not include that in our main regressions. Including that variable
produces consistent results though for a smaller sample (see Table A5). Further, accounting for
lagged terror attacks also leaves our main conclusions unchanged (see Table A5).
A major advantage of our subnational data structure comes from combining the withincountry variation for each period with the panel dimension of repeated information for each region. Our data allows us to account for region-fixed e↵ects to hold time-invariant, region-specific
particularities constant. This accounts for prevalent correlates of terrorism, such as geography
and terrain, unique historical features pertaining to civil conflict, civil war, colonization, and
others, as well as other long-term cultural, economic, and political artefacts. Year-fixed e↵ects
absorb any time-specific global developments that may independently correlate with terrorism.
Nevertheless, it is important to note which factors our analyses are unable to account for.
In particular, unobservable aspects that inform terrorism and do change within a region over
2
We code international attacks using the GTD classification which closely matches that of Enders et al. (2011).
Specifically, we code transnational attacks as IN T AN Y = 1 in the GTD i.e., either the attack is logistically
or ideologically transnational, or the nationality of the targets or victims di↵ers from the location of the attack.
All other attacks (IN T AN Y = 0 in the GTD) are coded as domestic in our main specifications. Employing
alternative definitions of domestic attacks produces consistent results (available upon request). Considering a
binary indicator for experiencing any attacks (to alleviate concerns about under-reporting in particularly lowincome regions) or predicting attacks/capita (to explicitly acknowledge the role of population size; Jetter and
Stadelmann, 2019) produces consistent findings (see Table A4).
3
We multiply oil production by international oil prices following Brückner et al. (2012).
8
0
Terror attacks in t…t+4
5
10
15
20
Panel A: GDP/capita and terror attacks
6
8
10
Ln(Subnational GDP/capita)
95% confidence interval
12
Terror attacks
Panel C: GDP/capita and
transnational terror attacks
Domestic terror attacks in t…t+4
0
5
10
15
Transnational terror attacks in t…t+4
0
1
2
3
20
Panel B: GDP/capita and
domestic terror attacks
6
8
10
Ln(Subnational GDP/capita)
95% confidence interval
12
6
Domestic terror attacks
8
10
Ln(Subnational GDP/capita)
95% confidence interval
12
Transnational terror attacks
Figure 2: Subnational GDP/capita and terror attacks, displayed by kernel-weighted local polynomial smoothing along with 95% confidence intervals.
9
time can influence our derived coefficients associated with income levels. For example, changes
in regional governance, changes in regional ethnic polities, or changes in within region inequality
are only incorporated to the extent that they are correlated with our observables of population
size, oil production, hosting the country’s capital, educational attainment, and lagged terror
attacks.
4
Empirical Methodology
4.1
Main Specification
Our main empirical strategy employs a negative binomial regression model in line with the
literature (Walsh and Piazza, 2010; Young and Dugan, 2011; Young and Findley, 2011; Piazza,
2013; Gaibulloev et al., 2017) because the dependent variable constitutes a non-negative count
variable and exhibits overdispersion. For region i and year t, we estimate:
Attacksi;(t,…,t+4) =
where
1
and
2
0+
1 Ln(GDP/capita)i;t +
2
2 Ln(GDP/capita)i;t + X i;t 3 +
i+
t + i;t ,
(1)
represent our main coefficients of interest. Note that observations do not
overlap, as for each region we employ an observation for, say, 1970-1974, another for 1975-1979,
and so on. We begin with a linear form assuming
2 = 0 and then relax this assumption allowing
for nonlinearity in accordance with Figure 2 and Enders and Hoover (2012), as well as Enders
et al. (2016). X i,t constitutes the matrix of control variables introduced in Section 3.3;
t capture region- and year-fixed e↵ects; and
4.2
i and
i;t represents an error term.
Potential Sources of Endogeneity
Endogeneity pertaining to reverse causality and omitted variables remains a threat to identifying causal relationship in the associated literature. First, reverse causality implies regions
(or countries) may become poorer because of terrorism. Aggregating the dependent variable
over years t to t + 4, while measuring independent variables in year t alleviates such concerns.
Predicting terrorism in t + 1 until t + 4, thereby not leaving any overlap between the dependent
and independent variables, produces consistent results (see Table A5). To further acknowledge
potential path dependency, additional specifications account for terror attacks in the previous
10
five years, producing consistent results (see Table A5). In sum, reverse causality is unlikely to
pose a systematic threat to the interpretation of our results.
Second, omitted variables, i.e., unobservable factors could influence both regional income
levels and terrorism. We control for a list of notable confounders in our main estimations
and additional robustness tests incorporate educational attainment levels leading to consistent
results (see Table A5). As discussed, region-fixed e↵ects account for any statistical variation in
terrorism owed to time-invariant regional cultural, ethnic, language, or religious heterogeneity.
For example, cultural heritage, religious denominations or language may di↵er geographically
within a country (e.g., across regions in the United Kingdom, Switzerland, or Tanzania) –
something that country-fixed e↵ects are not able to absorb, while region-fixed e↵ects are better
positioned to do so.
Similarly, geographical characteristics within a country often vary, and any potential association between poverty and terrorism may di↵er along such dimensions. For instance, Colombia’s
more hospitable regions happen to be wealthier (e.g., Bogotá or Medellı́n) than the difficult-toaccess rainforest regions. Region-fixed e↵ects capture substantially more unobservable, terrorismrelevant variation than country-fixed e↵ects in the traditional cross-country literature are able
to. For instance, if a region di↵ers systematically from the country average in terms of terrain or climate, but also in the de facto implementation of law and order, region-fixed e↵ects
capture such heterogeneity. Importantly, region-fixed e↵ects also implicitly account for countryfixed e↵ects, i.e., any country-level heterogeneity relevant for terrorism is accounted for, such as
historical events or colonial ties.
5
Regional Income and Terrorism
5.1
Main Results
Table 3 reports our main regression results. Column (1) considers a univariate regression that
only employs a linear term of GDP/capita to predict terror attacks. The respective coefficient
is negative and statistically significant at the 1% level (p-value of 0.000). Conclusions from this
specification would support many politicians’ (e.g., George Bush’s) responses to 9/11 in the
association between income levels and terrorism.
11
However, upon allowing for nonlinearity in column (2), that conclusion changes, suggesting
an inverted U-shape: GDP/capita becomes a positive predictor, while its squared term emerges
as a negative predictor (p-values of 0.016 and 0.006). The fourth row from the bottom reports the
GDP/capita level at which the income-terrorism relationship is suggested to peak, corresponding
to US$2,826.
Table 3: Main results, predicting terror attacks for region i in years t,…,t + 4 in a negative
binomial regression framework.
Ln(GDP/capita)i,t
(1)
(2)
(3)
(4)
(5)
Domestic
terrorism
(6)
International
terrorism
-0.351⇤⇤⇤
(0.094)
3.131⇤⇤
(1.294)
4.690⇤⇤⇤
(1.287)
3.517⇤⇤⇤
(0.379)
3.677⇤⇤⇤
(0.416)
5.472⇤⇤⇤
(0.638)
-0.197⇤⇤⇤
(0.072)
-0.282⇤⇤⇤
(0.072)
-0.186⇤⇤⇤
(0.021)
-0.202⇤⇤⇤
(0.024)
-0.298⇤⇤⇤
(0.036)
X
X
X
X
X
X
X
Ln(GDP/capita)2i,t
Control variablesa and
time-period-fixed e↵ects
Region-fixed e↵ects
GDP/capita at maximum
Nb
# of regionsb
# of time periods
8,383
1,527
9
2,826
4,087
12,763
8,969
9,713
8,383
1,527
9
8,383
1,526
9
5,351
863
9
5,055
802
9
3,357
517
9
Notes: Standard errors clustered at the regional level are displayed in parentheses for columns (1) – (3) while columns (4)
– (6) report standard errors based on the observed information matrix, using the option vce(oim) in STATA. ⇤ p < 0.10, ⇤⇤ p < 0.05, ⇤⇤⇤ p < 0.01. a Control variables include the logarithm of population size, a binary indicator for the location of the capital city, and the natural logarithm of oil produced. b The decline in the number of observations in columns (4)-(6) stems from the introduction of region-fixed e↵ects, where regions with no terror attacks are dropped automatically. Columns (3) and (4) first add the covariates introduced in equation (1) and time-periodfixed e↵ects, before also accounting for region-fixed e↵ects. The inverted U-shape persists, while the suggested peak rises to US$12,763. This value roughly corresponds to regions such as Quintana Roo (Mexico) in 1980 or Kaliningrad (Russia) in 2010. It is important to recall that the specification in column (4) exploits within-region variation only, i.e., we only compare the same region to itself at di↵erent income levels. Thus, the derived coefficients do not rely on any cross-regional di↵erences, not even within the same country. A corollary of that statistical 12 artefact is that a low-income region is suggested to experience rising likelihoods of terrorism as its GDP/capita levels increase; but as soon as GDP/capita levels surpass the peak for that same region, terrorism diminishes, everything else equal. Columns (5) and (6) delineate between domestic and transnational terrorism, acknowledging the often-proposed distinction between these types of terrorism and their underlying dynamics (Enders and Hoover, 2012; Enders et al., 2016). Our results are consistent: In both cases, we derive statistical significance at the one percent level for both coefficients of interest, as well as the signs suggested by our benchmark estimation from column (4). Domestic terrorism peaks at a level of GDP/capita that is lower than that for transnational terrorism, but the corresponding di↵erence remains small (a conclusion that also emerges from Figure 3).4 Figure 3 visualizes the suggested relationships from columns (4)-(6). The peaks of the inverted-U shape are comparable for domestic and transnational terrorism, which implies a universal nonlinearity of the relationship between income and terrorism. Interestingly, the slope of the relationship di↵ers to some degree, as transnational terrorism appears to be more responsive to GDP/capita in quantitative terms. 5.2 Robustness Checks We conduct a large series of alternative specifications to test the validity of these results. In particular, we implement alternative estimation techniques and measures of terrorism by (i) calculating bootstrapped standard errors, (ii) applying Poisson and Ordinary Least Square (OLS) methods, (iii) considering alternative measures of terrorism with attacks per year, terror per capita, a binary indicator for experiencing any terrorism, and deaths from terrorism. Across all these specifications, the inverted U-shaped relationship prevails with remarkable consistency (see Table A4). Table A5 documents regression results from (i) considering levels of GDP/capita (i.e., not applying the natural logarithm), (ii) controlling for years of educational attainment at the regional level, (iii) controlling for terror attacks in the past five years, (iv) using an alternative time frame for our outcome variable (from t+1 to t+4), and (v) considering annual GDP/capita 4 This result is also consistent with a narrative of strict security measures across borders encouraging perpetrators to target foreign entities at home (Enders et al., 2016). 13 2500 2000 1500 1000 500 0 Change in number of terror attacks in % GDP/capita and terrorism 0 2 4 6 8 10 12 14 16 Ln(Regional GDP per capita) 0 2 4 6 8 10 12 14 16 Ln(Regional GDP per capita) 2500 2000 1500 1000 500 0 500 1000 1500 2000 2500 Change in number of terror attacks in % GDP/capita and transnational terror attacks 0 Change in number of terror attacks in % GDP/capita and domestic terror attacks 0 2 4 6 8 10 12 Ln(Regional GDP per capita) Figure 3: Visualizing regression results from columns (4)-(6) of Table 3. 14 14 16 data as reported in Gennaioli et al. (2014) without adjusting observations to conform with our five-year panel structure. Again, results remain consistent. 5.3 Terror Group Ideologies Finally, we explore the link between poverty and terrorism for di↵erent group ideologies. Prior cross-country research has suggested the role of income levels may vary depending on a group’s ideological background (Enders et al., 2016). Consistent with the common distinctions, we delineate between Islamist, left-wing, right-wing, ethnic/separatist, and religious non-Islamist groups (e.g., see Kis-Katos et al., 2014). Table 4 provides further support for a universal nonlinearity when distinguishing between these categories, as the inverted U-shaped pattern emerges across all five group ideologies.5 These results prevail when delineating between domestic and transnational terrorism (Tables A7 and A8). Notably, the corresponding peaks di↵er in terms of magnitude, although moderately. This finding supports the theoretical proposition that peaks in terrorism di↵er with perpetrator ideology (e.g., Enders et al., 2016): The peak of terrorism associated with Islamist and other religious ideologies occur at income levels that are lower than those for left-wing or right-wing ideologies. 6 Conclusion This paper analyzes the relationship between income and terrorism at the subnational (regional) level. Using data for 1,527 subnational entities from 1970 to 2014, all results provide firm support for an inverted U-shape in how regional income levels link to regional terror attacks. This result prevails once we account for a comprehensive set of covariates, as well as region- and yearfixed e↵ects; when delineating between domestic and transnational terrorism; and even when distinguishing between terror group ideology. Contrary to the post-9/11 claims of poverty being a monotonically positive predictor of terrorism, these results suggest poverty alleviation can 5 We extend Kis-Katos et al.’s (2014) code beyond 2008 to include newer terrorist organizations that conducted ten or more attacks. Nevertheless, limiting our analysis to 2008 produces consistent results (available on request). Table A6 reports results for a stricter definition of group identity in which a group is considered Islamist if their main identity is religious and their religious identity is Islam. 15 Table 4: Distinguishing by group ideology, predicting the number of terror attacks for subnational region i in years t,...,t + 4 in a negative binomial regression framework. (1) Islamist (2) Left-wing (3) Right-wing (4) Ethnic/ separatist (5) Religious non-Islamist Ln(GDP/capita)i,t 7.075⇤⇤⇤ (1.378) 4.829⇤⇤⇤ (0.661) 10.399⇤⇤⇤ (1.760) 5.519⇤⇤⇤ (0.782) 10.039⇤⇤⇤ (1.687) Ln(GDP/capita)2i,t -0.408⇤⇤⇤ (0.081) -0.269⇤⇤⇤ (0.037) -0.569⇤⇤⇤ (0.098) -0.308⇤⇤⇤ (0.044) -0.584⇤⇤⇤ (0.095) Control variablesa X X X X X Time-period- and region-fixed e↵ects X X X X X GDP/capita at maximum 5,827 7,910 9,302 7,781 5,405 N # of regionsb # of time periods 841 145 9 3,060 441 9 1,456 191 9 2,036 306 9 806 107 9 Group identity: Notes: Standard errors based on the observed information matrix, using the option vce(oim) in STATA), are displayed in parentheses. ⇤ p < 0.10, ⇤⇤ p < 0.05, ⇤⇤⇤ p < 0.01. a Control variables include the logarithm of population size, a binary indicator for the location of the capital city, and the natural logarithm of oil produced. potentially lead to more terrorism for countries that are currently to the left of the average peaks we derive. Naturally, we advise caution in the interpretation of these findings since, similar to most of the cross-country literature, our analysis is not able to fully resolve all empirical challenges. For example, unobservable factors that change within a subnational region over time may still be able to bias the coefficients we derive. Nevertheless, the subnational data structure allows us to substantially alleviate these endogeneity concerns , especially when compared to the crosscountry literature. Our most complete estimations exploit within-region variation only, i.e., any time-invariant di↵erences across regions (even within the same country) are filtered out. Carefully structuring corresponding time sequencing by using contemporaneous GDP/capita levels to predict subsequent terrorism further addresses threats from reverse causality. Results also remain consistent when accounting for lagged terror levels. In sum, the fact that the inverted U-shape emerges in virtually all settings provides what we believe to be the strongest empirical evidence to date for a systematic, universal link between 16 income levels and terrorism. We hope these insights can inform national and regional policymakers, as well as inspire further research into a topic that has informed substantial political and societal decisions since 9/11. 17 References Abadie, A. (2006). Poverty, political freedom, and the roots of terrorism. American Economic Review, 96(2):50–56. Alesina, A., Özler, S., Roubini, N., and Swagel, P. (1996). Political instability and economic growth. Journal of Economic Growth, 1(2):189–211. Azam, J.-P. and Delacroix, A. (2006). Aid and the delegated fight against terrorism. Review of Development Economics, 10(2):330–344. Azam, J.-P. and Thelen, V. (2008). The roles of foreign aid and education in the war on terror. Public Choice, 135(3-4):375–397. Basuchoudhary, A. and Shughart, W. F. (2010). On ethnic conflict and the origins of transnational terrorism. Defence and Peace Economics, 21(1):65–87. Benmelech, E., Berrebi, C., and Klor, E. F. (2012). Economic conditions and the quality of suicide terrorism. The Journal of Politics, 74(1):113–128. Berman, E. and Laitin, D. D. (2008). Religion, terrorism and public goods: Testing the club model. Journal of Public Economics, 92(10-11):1942–1967. Berrebi, C. (2007). Evidence about the Link Between Education, Poverty and Terrorism among Palestinians. Peace Economics, Peace Science, and Public Policy, 13(1):1–38. Blomberg, S. B. and Hess, G. D. (2008a). The Lexus and the olive branch: Globalization, democratization and terrorism. Terrorism, Economic Development, and Political Openness. Blomberg, S. B. and Hess, G. D. (2008b). From (no) butter to guns? Understanding the economic role in transnational terrorism. Terrorism, Economic Development, and Political Openness. Blomberg, S. B. and Rosendor↵, B. P. (2006). A gravity model of globalization, democracy and transnational terrorism. USC CLEO Research Paper, (C06-6). Bloom, D. E. and Canning, D. (2000). 287(5456):1207–1209. The health and wealth of nations. Science, Braithwaite, A. and Li, Q. (2007). Transnational terrorism hot spots: Identification and impact evaluation. Conflict Management and Peace Science, 24(4):281–296. Bravo, A. B. S. and Dias, C. M. M. (2006). An empirical analysis of terrorism: Deprivation, Islamism and geopolitical factors. Defence and Peace Economics, 17(4):329–341. Brückner, M., Ciccone, A., and Tesei, A. (2012). Oil price shocks, income, and democracy. Review of Economics and Statistics, 94(2):389–399. Burgoon, B. (2006). On welfare and terror: Social welfare policies and political-economic roots of terrorism. Journal of Conflict Resolution, 50(2):176–203. 18 Bush, G. W. (2002). United States of America: Remarks by Mr. George W. Bush President at the International Conference on Financing for Development. March 22, 2002. Campos, N. F. and Gassebner, M. (2013). International terrorism, domestic political instability, and the escalation e↵ect. Economics & Politics, 25(1):27–47. Collier, P. and Hoe✏er, A. (2004). Greed and grievance in civil war. Oxford Economic Papers, 56(4):563–595. Crenshaw, E., Robison, K., and Jenkins, J. C. (2007). The “roots” of transnational terrorism: A replication and extension of Burgoon. In Annual Meetings of the American Sociological Association, New York City, NY (August 2007). De la Calle, L. and Sánchez-Cuenca, I. (2012). Rebels without a territory: An analysis of nonterritorial conflicts in the world, 1970–1997. Journal of Conflict Resolution, 56(4):580– 603. Dreher, A. and Fischer, J. A. (2010). Government decentralization as a disincentive for transnational terror? An empirical analysis. International Economic Review, 51(4):981–1002. Dreher, A. and Fischer, J. A. (2011). Does government decentralization reduce domestic terror? An empirical test. Economics Letters, 111(3):223–225. Easterly, W. (2016). The war on terror vs. the war on poverty. The New York Review of Books. Enders, W. and Hoover, G. A. (2012). The nonlinear relationship between terrorism and poverty. American Economic Review, 102(3):267–72. Enders, W., Hoover, G. A., and Sandler, T. (2016). The changing nonlinear relationship between income and terrorism. Journal of Conflict Resolution, 60(2):195–225. Enders, W., Sandler, T., and Gaibulloev, K. (2011). Domestic versus transnational terrorism: Data, decomposition, and dynamics. Journal of Peace Research, 48(3):319–337. Eyerman, J. (1998). Terrorism and democratic states: Soft targets or accessible systems. International Interactions, 24(2):151–170. Freytag, A., Krüger, J. J., Meierrieks, D., and Schneider, F. (2011). The origins of terrorism: Cross-country estimates of socio-economic determinants of terrorism. European Journal of Political Economy, 27:S5–S16. Friedman, B. M. (2010). The moral consequences of economic growth. Vintage. Gaibulloev, K., Piazza, J. A., and Sandler, T. (2017). Regime types and terrorism. International Organization, pages 1–32. Gassebner, M. and Luechinger, S. (2011). Lock, stock, and barrel: A comprehensive assessment of the determinants of terror. Public Choice, 149(3-4):235. Gennaioli, N., La Porta, R., De Silanes, F. L., and Shleifer, A. (2014). Growth in regions. Journal of Economic Growth, 19(3):259–309. 19 Goldstein, K. B. (2005). Unemployment, inequality and terrorism: Another look at the relationship between economics and terrorism. Undergraduate Economic Review, 1(1):6. Gosling, T. L. (2017). States of terror: Regional income and terrorism. Unpublished manuscript. Gürlük, S. (2009). Economic growth, industrial pollution and human development in the Mediterranean region. Ecological Economics, 68(8-9):2327–2335. Haggar, K. E. (2021). President Al-Sisi Fights Terrorism by Eliminating Informal Settlements. Daily News Egypt. Hassan, N. (2002). An arsenal of believers. Le Débat, (3):134–143. Jetter, M. and Stadelmann, D. (2019). Terror per capita. Southern Economic Journal, 86(1):286– 304. Kis-Katos, K., Liebert, H., and Schulze, G. G. (2014). On the heterogeneity of terror. European Economic Review, 68:116–136. Koch, M. T. and Cranmer, S. (2007). Testing the “Dick Cheney” hypothesis: Do governments of the left attract more terrorism than governments of the right? Conflict Management and Peace Science, 24(4):311–326. Krieger, T. and Meierrieks, D. (2019). Income inequality, redistribution and domestic terrorism. World Development, 116:125–136. Krueger, A. (2007). What makes a terrorist? It’s not poverty and lack of education, according to economic research by Princeton’s Alan Krueger Look elsewhere. The American (Washington, DC), 1(7):16–22. Krueger, A. B. and Laitin, D. D. (2008). Kto kogo?: A cross-country study of the origins and targets of terrorism. Terrorism, economic development, and political openness, pages 148–173. Krueger, A. B. and Malečková, J. (2003). Education, poverty and terrorism: Is there a causal connection? Journal of Economic Perspectives, 17(4):119–144. Kurrild-Klitgaard, P., Justesen, M. K., and Klemmensen, R. (2006). The political economy of freedom, democracy and transnational terrorism. Public Choice, 128(1-2):289–315. Lai, B. (2007). “Draining the swamp”: An empirical examination of the production of international terrorism, 1968 - 1998. Conflict Management and Peace Science, 24(4):297–310. Li, Q. (2005). Does democracy promote or reduce transnational terrorist incidents? Journal of Conflict Resolution, 49(2):278–297. Li, Q. and Schaub, D. (2004). Economic globalization and transnational terrorism: A pooled time-series analysis. Journal of Conflict Resolution, 48(2):230–258. Maslow, A. H. (1943). A theory of human motivation. Psychological Review, 50(4):370. Miguel, E., Satyanath, S., and Sergenti, E. (2004). Economic shocks and civil conflict: An instrumental variables approach. Journal of political Economy, 112(4):725–753. 20 Neumayer, E. and Plümper, T. (2009). International Terrorism and the Clash of Civilizations. British Journal of Political Science, 39(4):711. Odede, K. (2015). If you really want to fight terrorism, start by fighting child poverty. The Guardian. Published on August 21, 2015. Piazza, J. A. (2006). Rooted in poverty?: Terrorism, poor economic development, and social cleavages. Terrorism and Political Violence, 18(1):159–177. Piazza, J. A. (2007). Draining the swamp: Democracy promotion, state failure, and terrorism in 19 Middle Eastern countries. Studies in Conflict & Terrorism, 30(6):521–539. Piazza, J. A. (2008a). Incubators of terror: Do failed and failing states promote transnational terrorism? International Studies Quarterly, 52(3):469–488. Piazza, J. A. (2008b). Do democracy and free markets protect us from terrorism? International Politics, 45(1):72–91. Piazza, J. A. (2011). Poverty, minority economic discrimination, and domestic terrorism. Journal of Peace Research, 48(3):339–353. Piazza, J. A. (2013). Regime Age and Terrorism: Are New Democracies Prone to Terrorism? International Interactions, 39(2):246–263. Pilgrim, S. (2015). Poverty and injustice feeding terrorism, French minister says. France 24 News. Published on February 10, 2015. Plümper, T. and Neumayer, E. (2010). The friend of my enemy is my enemy: International alliances and international terrorism. European Journal of Political Research, 49(1):75–96. Sageman, M. (2004). Understanding terror networks. University of Pennsylvania Press. Sambanis, N. (2008). Terrorism and civil war. Terrorism, Economic Development, and Political Openness, pages 174–206. START (2017). Global Terrorism Database. National Consortium for the Study of Terrorism and Responses to Terrorism (START). Retrieved from http://www.start.umd.edu/gtd. Sterman, D. (2015). Don’t Dismiss Poverty’s Role in Terrorism Yet. Time, 4. Tavares, J. (2004). The open society assesses its enemies: Shocks, disasters and terrorist attacks. Journal of Monetary Economics, 51(5):1039–1070. Testas, A. (2004). Determinants of terrorism in the Muslim world: An empirical cross-sectional analysis. Terrorism and Political Violence, 16(2):253–273. Walsh, J. I. and Piazza, J. A. (2010). Why respecting physical integrity rights reduces terrorism. Comparative Political Studies, 43(5):551–577. Wolfensohn, J. D. (2002). Fight terrorism by ending poverty. New Perspectives Quarterly, 19(2):42–44. 21 Young, J. K. and Dugan, L. (2011). Veto players and terror. Journal of Peace Research, 48(1):19–33. Young, J. K. and Findley, M. G. (2011). Promise and pitfalls of terrorism research. International Studies Review, 13(3):411–431. 22 Appendix A Table A1: Summary Statistics for additional variables. Variable N Mean (Std. Dev.) Min. (Max.) Description Panel A: Dependent variables Terror attackst1,..,t+4 8,353 5.91 (36.84) 0 (1160) # of terror attacks in t1, .., t + 4 Terror attacks per yeart,..,t+4 8383 1.48 (9.30) 0 (295.80) # of terror attacks per year in t, .., t + 4 Terror attacks per capitat,..,t+4 8383 4.06 (29.14) 0 (1021.87) # of terror attacks per million of population in t, .., t + 4 Terror attacks (Y/N)t,..,t+4 8383 0.34 (0.47) 0 (1) Any terror attack (Y/N) in t, .., t + 4 Killed in terror attackst,..,t+4 8,383 12.15 (138.49) 0 (10,335) # of people killed in terror attacks in t, .., t + 4 Islamists alt attackst,..,t+4 8,383 0.35 (13.01) 0 (1135) # of terror attacks by groups with prime identity Islam in t, .., t + 4 Left alt attackst,..,t+4 8,383 1.96 (19.56) 0 (986) # of terror attacks by groups with prime identity Left in t, .., t + 4 Right alt attackst,..,t+4 8,383 0.11 (1) 0 (36) # of terror attacks by groups with prime identity Right in t, .., t + 4 Ethnic/sep alt attackst,..,t+4 8,383 1.37 (13.62) 0 (674) # of terror attacks by groups with prime identity Ethnic/separatists in t, .., t + 4 Religious Non-Islamist alt attackst,..,t+4 8,383 0.02 (0.47) 0 (25) # of terror attacks by groups with prime identity Religious but not Islam in t, .., t + 4 Islamists domestic attackst,...,t+4 8383 0.37 (13.27) 0 (1123) # of domestic terror attacks by Islamist groups in t, .., t + 4 Islamists transnational attackst,...,t+4 8383 0.09 (1.95) 0 (120) # of transnational terror attacks by Islamist groups in t, .., t + 4 Left domestic attackst,...,t+4 8383 2.29 (21.20) 0 (877) # of domestic terror attacks by Left-wing groups in t, .., t + 4 Left transnational attackst,...,t+4 8383 0.44 (4.15) 0 (171) # of transnational terror attacks by Left-wing groups in t, .., t + 4 Right domestic attackst,...,t+4 8383 0.10 (0.88) 0 (34) # of domestic terror attacks by Right-wing groups in t, .., t + 4 Right transnational attackst,...,t+4 8383 0.04 (0.60) 0 (35) # of transnational terror attacks by Right-wing groups in t, .., t + 4 Ethnic/Sep domestic attackst,...,t+4 8383 1.09 (16.85) 0 (1218) # of domestic terror attacks by ethnic or separatist groups in t, .., t + 4 Ethnic/Sep transnational attackst,...,t+4 8383 0.5 (8.6) 0 (673) # of transnational terror attacks by ethnic or separatist groups in t, .., t + 4 Religious Non-Islamists domestic attackst,...,t+4 8383 0.24 (4.23) 0 (242) # of domestic terror attacks by religious (non-Islamist) groups in t, .., t + 4 Religious Non-Islamists transnational attackst,...,t+4 8383 0.16 (7.58) 0 (673) # of transnational terror attacks by religious (non-Islamist) groups in t, .., t + 4 GDP/capita 8,383 12,429.07 (12,334.17) 188.97 (166,007.3) GDP per capita in 2005 PPP US$ Education 6,940 7.34 (3.23) 0.67 (13.76) Years of educational attainment 23 Table A2: Country-years that appear in the regional data set. 1970 Albania Argentina Australia Austria Bangladesh Belgium Benin Bolivia Bosnia and Herzegovina Brazil Bulgaria Canada Chile China Colombia Croatia Czech Republic Denmark Ecuador Egypt, Arab Rep. El Salvador Estonia Finland France Germany, East Germany, West Greece Guatemala Honduras Hungary India Indonesia Iran, Islamic Rep. Ireland Italy Japan Jordan Kazakhstan Kenya Korea, Rep. Kyrgyz Republic Latvia Lesotho Lithuania Macedonia Malaysia MeXico Mongolia Morocco Mozambique Nepal Netherlands Nicaragua Nigeria Norway Pakistan Panama Paraguay Peru Philippines Poland Portugal Romania Russian Federation Serbia Slovak Republic Slovenia South Africa Spain Sri Lanka Sweden Switzerland Tanzania Thailand Turkey Ukraine United Arab Emirates United Kingdom United States Uruguay Uzbekistan Venezuela Vietnam Total X X 1975 X X X 1980 1985 1990 X X X X X X X X X X X 1995 X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X 23 X X X 26 X X X X X X X X X X X X X X X X X X X X X 33 30 24 X X X X X X X X X X X X X X X X X X X X 50 X X X X X X X X X 2000 2005 X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X 65 X 75 X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X 2010 Total X 3 5 6 9 4 5 3 7 1 9 5 9 8 9 9 3 4 9 3 3 3 4 7 9 5 9 9 3 4 4 7 5 3 6 8 8 3 5 1 6 3 3 3 4 4 8 7 5 4 4 2 4 3 2 6 8 4 3 9 7 5 8 4 4 1 4 4 9 7 5 6 9 7 7 6 3 5 5 9 3 3 3 5 438 X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X 71 X 65 Table A3: Comparison between our sample and global database. All country-years Sample countries Sample country-years All country-years Sample countries Sample country-years 123.770 (102.559) 501.310⇤⇤ (199.062) 1016.198⇤⇤ (476.781) 1242.816⇤ (674.529) -30.266⇤⇤ (11.929) -61.777⇤⇤ (29.856) -71.963⇤ (36.388) 405.002⇤⇤ (181.363) 886.987⇤ (447.070) 1046.545⇤ (625.057) -24.032⇤⇤ (10.849) -53.575⇤ (28.050) -59.246⇤ (33.376) 96.308⇤⇤ (37.302) 129.210⇤⇤⇤ (45.979) 196.271⇤⇤ (78.383) -6.234⇤⇤ (2.402) -8.202⇤⇤⇤ (2.938) -12.717⇤⇤ (5.065) 1,624 581 419 Panel A: Dependent variable - Terror attacks Ln(country GDP/capita)i,t 23.989 90.724 (17.017) (71.510) Ln(country GDP/capita)2i,t Panel B: Dependent variable - Domestic terror attacks Ln(country GDP/capita)i,t 26.003 84.391 (15.974) (68.627) 125.255 (99.771) Ln(country GDP/capita)2i,t Panel C: Dependent variable - Transnational terror attacks Ln(country GDP/capita)i,t -2.015 6.333 (2.625) (6.148) -1.485 (13.021) Ln(country GDP/capita)2i,t N 1,624 581 419 Notes: Country GDP per capita is in constant 2010 US dollars and is obtained from United Nations Statistical Database. All regressions control for Country-FE and Year-FE. Standard errors clustered at country level are displayed in parentheses.⇤ p < 0.10, ⇤⇤ p < 0.05, ⇤⇤⇤ p < 0.01. Table A4: Robustness checks using the main specification of Column 6 in Table 3. Estimation Method: (1) NBREGa Dependent Variable: (2) Poisson (3) OLS Subnational terrort,..,t+4 (4) OLS (5) OLS (6) Logit (7) NBREG Subnational terror per yeart,..,t+4 Subnational terror per capitat,..,t+4 Subnational terror (Y/N)t,..,t+4 Subnational killed in terrort,..,t+4 Ln(GDP/capita)i,t 3.517⇤⇤⇤ (0.480) 15.651⇤⇤⇤ (2.593) 72.161⇤⇤⇤ (14.090) 14.432⇤⇤⇤ (2.818) 21.552⇤ (11.092) 7.604⇤⇤⇤ (1.023) 4.454⇤⇤⇤ (0.489) Ln(GDP/capita)2 i,t -0.186⇤⇤⇤ (0.027) -0.963⇤⇤⇤ (0.153) -4.434⇤⇤⇤ (0.841) -0.887⇤⇤⇤ (0.168) -1.264⇤⇤ (0.616) -0.439⇤⇤⇤ (0.060) -0.261⇤⇤⇤ (0.028) Control variablesb X X X X X X X Time period- and regionfixed e↵ects X X X X X X X 5,351 5,351 8,383 8,383 8,383 4,644 3,808 N Notes: Standard errors are displayed in parentheses. Column (1) reports bootstrapped standard errors, columns (2) - (4) report robust standrad errors clustered at the regional level, and columns (5) and (6) report standard errors based on the observed information matrix, using the option vce(oim) in STATA. ⇤ p < 0.10, ⇤⇤ p < 0.05, ⇤⇤⇤ p < 0.01. a Specification (1) reports bootstrapped standard errors (100 reps). b Specifications (11) and (12) consider subnational GPD/capita and controls for the years reported in the Gennaioli et al. (2014) dataset without any adjustment. b Controls include logged subnational population, logged value of oil production, and a binary indicator for location of capital in the region. 25 Table A5: Further robustness checks using the main specification of Column 6 in Table 3. Estimation Method: (1) NBREG (2) NBREG (3) NBREG (4) NBREG (5) NBREGb (6) NBREGb Subnational terrort1,..,t+4 Subnational terrort,..,t+4 Subnational terrort1,..,t+4 Dependent Variable: Subnational terrort,..,t+4 Ln(GDP/capita)i,t 3.708⇤⇤⇤ (0.479) 3.520⇤⇤⇤ (0.481) 3.647⇤⇤⇤ (0.396) 3.555⇤⇤⇤ (0.425) 3.456⇤⇤⇤ (0.439) Ln(GDP/capita)2i,t -0.212⇤⇤⇤ (0.027) -0.181⇤⇤⇤ (0.027) -0.194⇤⇤⇤ (0.022) -0.187⇤⇤⇤ (0.024) -0.180⇤⇤⇤ (0.025) X X X X X Control variablesc X GDP/capitai,t 0.020⇤⇤⇤ GDP/capita2i,t -0.000⇤⇤⇤ (0.000) (0.005) X Education Terror attackst N X 5,..,t 1 Time period- and regionfixed e↵ects X X X X X X 5,351 4,280 3,882 5,084 4,624 4,411 Notes: Standard errors based on the observed information matrix, using the option vce(oim) in STATA, are displayed in parentheses. ⇤ p < 0.10, ⇤⇤ p < 0.05, ⇤⇤⇤ p < 0.01. a Specification (1) reports bootstrapped standard errors (100 reps). b Specifications (5) and (6) consider subnational GDP/capita and controls for the years reported in the Gennaioli et al. (2014) dataset without any adjustment. c Controls include logged subnational population, logged value of oil production, and a binary indicator for location of capital in the region. 26 Table A6: Predicting terror attacks in period t, ..., t + 4 perpetrated by groups with various identities considering only prime identity of the group, building on the specification in column(6) Table 3. (1) Islamistsalt (2) Leftalt (3) Rightalt (4) Ethnic/Sepalt (5) Religious Non-Islamistsalt Dependent variable: Subnational terrort,..,t+4 Ln(GDP/capita)i,t 7.993⇤⇤⇤ (1.676) 8.201⇤⇤⇤ (0.875) 9.708⇤⇤⇤ (1.926) 5.455⇤⇤⇤ (0.816) 11.591⇤⇤⇤ (3.825) Ln(GDP/capita)2i,t -0.461⇤⇤⇤ (0.099) -0.487⇤⇤⇤ (0.050) -0.526⇤⇤⇤ (0.106) -0.304⇤⇤⇤ (0.046) -0.649⇤⇤⇤ (0.215) Control variablesa X X X X X Time period- and regionfixed e↵ects X X X X X 5,820.83 4,489.36 9902.67 7,932.22 7,554.45 648 2,393 1,349 1,897 291 GDP/capita at the maximum N Notes: Standard errors based on the observed information matrix, using the option vce(oim) in STATA, are displayed in parentheses. ⇤ p < 0.10, ⇤⇤ p < 0.05, ⇤⇤⇤ p < 0.01. a Controls include logged subnational population and a binary indicator for location of capital in the region. 27 Table A7: Displaying results for domestic terror attacks in period t, ..., t + 4 perpetrated by groups with various identities, building on the specification in column(6) Table 3. (1) Islamists (2) Left (3) Right (4) Ethnic/Sep (5) Religious Non-Islamists Dependent variable: Subnational domestic terrort,..,t+4 Ln(GDP/capita)i,t 5.678⇤⇤⇤ (1.886) 8.890⇤⇤⇤ (0.837) 11.952⇤⇤⇤ (1.996) 6.800⇤⇤⇤ (1.100) 7.816⇤⇤⇤ (2.025) Ln(GDP/capita)2i,t -0.349⇤⇤⇤ (0.115) -0.525⇤⇤⇤ (0.048) -0.665⇤⇤⇤ (0.111) -0.408⇤⇤⇤ (0.064) -0.478⇤⇤⇤ (0.115) Control variablea X X X X X Time period- and regionfixed e↵ects X X X X X 3,410.69 4,753.64 7994.16 4,160.26 3,553.66 548 2,459 1,308 1,321 614 GDP/capita at the maximum N Notes: Standard errors based on the observed information matrix, using the option vce(oim) in STATA, are displayed in parentheses. ⇤ p < 0.10, ⇤⇤ p < 0.05, ⇤⇤⇤ p < 0.01. a Controls include logged subnational population, value of oil production, educational attainment, and a binary indicator for location of capital in the region. 28 Table A8: Displaying results for transnational terror attacks in period t, ..., t + 4 perpetrated by groups with various identities, building on the specification in column(6) Table 3. (1) Islamists (2) Left (3) Right (4) Ethnic/Sep (5) Religious Non-Islamists Dependent variable: Subnational transnational terrort,..,t+4 Ln(GDP/capita)i,t 11.148⇤⇤⇤ (2.349) 5.139⇤⇤⇤ (1.166) 10.879⇤⇤ (4.585) 10.217⇤⇤⇤ (1.289) 23.163⇤⇤⇤ (3.982) Ln(GDP/capita)2i,t -0.616⇤⇤⇤ (0.135) -0.278⇤⇤⇤ (0.064) -0.515⇤⇤ (0.243) -0.576⇤⇤⇤ (0.073) -1.298⇤⇤⇤ (0.222) Control variablesa X X X X X Time period- and regionfixed e↵ects X X X X X 8,507.48 10,329.98 38,643.58 7,107.63 7,499.36 571 2,044 509 1,396 453 GDP/capita at the maximum N Notes: Standard errors based on the observed information matrix, using the option vce(oim) in STATA, are displayed in parentheses. ⇤ p < 0.10, ⇤⇤ p < 0.05, ⇤⇤⇤ p < 0.01. a Controls include logged subnational population, value of oil production, and a binary indicator for location of capital in the region. 29 Appendix B: Data Preparation for Regional Income Levels and Terror Attacks The original data on GDP/capita constitute an unbalanced panel in which intervals between observations for each entity varies. We adjust the data to construct a panel containing five-yearly data by matching the observation on a subnational entity for a year to the closest year in the five-yearly panel. For example, the Albanian Berat region reports GDP/capita for 1990, 2001, and 2009. We assign these observations to the years 1990, 2000, and 2010. Keeping reported years in their original format produces consistent findings (see Table A5). To match regional GDP/capita data with terror attacks, we follow a three-step matching process: First, we match the subnational entity listed in the GTD with the corresponding entity in the Gennaioli et al. (2014) database. This data-merging mechanism itself remains imperfect because subnational entities listed in the GTD are not standardized in terms of spellings or changes in the boundaries of the entities over time. Second, for those observations, we match the respective information from the GTD by hand to the regional level identified by Gennaioli et al. (2014). Some terror attacks lack information on the subnational entity but feature more disaggregated geographical identifiers (e.g., city-level). Third, we exploit the geographical coordinates of the remaining attacks to match them to a subnational region. The terror incidents that remained unmatched after the three steps were discarded. Overall, these steps allow us to match 92% of all 107,221 attacks listed in our sample period. 30 Terrorists versus the Government STRATEGIC INTERACTION, SUPPORT, AND SPONSORSHIP KEVIN SIQUEIRA TODD SANDLER School of Economic, Political and Policy Sciences University of Texas at Dallas This article focuses on the strategic interaction between a terrorist group and a government as both vie for grassroots support. When terrorists and the government act contemporaneously, the equilibrium outcome depends on the effectiveness of the government’s countermeasures and the ability of the government to curb popular support of the terrorists through public spending. In two alternative scenarios, the authors establish that leadership may improve both adversaries’ well-being while reducing terrorism. The leader changes in the two cases, with the weaker player going first to the advantage of both players. State sponsorship and franchising of terrorists augment violence as both adversaries expend more effort. Sponsors can offset some strategic limits to violence that competition for supporters offers. Keywords: noncooperative game; leader-follower; terrorism; counterterrorism; state sponsorship Since the modern epoch of transnational terrorism beginning in 1968, terrorist groups have come and gone, with the overwhelming majority lasting less than a year (Hoffman 1998). Why has the Palestine Liberation Organization (PLO) endured for decades, while the Combatant Communist Cells (CCC) in Belgium lasted from just October 1984 to December 1985? The contention of this article is that at least three factors play a role: the responsiveness of grassroots supporters, the effectiveness of the government’s counterterrorism campaign, and the terrorist group’s ability to attract outside sponsorship. Without outside sponsorship, terrorist organizations must gain the active backing of a base of supporters in the population while simultaneously fending off government actions to limit the organizations’ effectiveness. This grassroots support takes the form of contributions and political allegiance. Targeted governments also face a dilemma: they can apply stringent counterterrorism policy (i.e., the stick) or more accommodative actions (i.e., the carrot) to AUTHORS’ NOTE: This research was partially supported by the U.S. Department of Homeland Security through the Center for Risk and Economic Analysis of Terrorism Events (CREATE) at the University of Southern California, grant number N00014-05-0630. However, any opinions, findings, and conclusions or recommendations are solely those of the authors and do not necessarily reflect the views of the Department of Homeland Security. JOURNAL OF CONFLICT RESOLUTION, Vol. 50 No. 6, December 2006 878-898 DOI: 10.1177/0022002706293469 © 2006 Sage Publications 878 Siqueira, Sandler / TERRORISTS VERSUS THE GOVERNMENT 879 reduce the terrorists’ base of support (Frey 2004). Rigorous counterterrorism may not only sow the seeds of discontent through callous actions (e.g., the French execution of two Front de Libération National terrorists in Algeria in 1956 [Hoffman 1998, 61]) but also divert funds from social programs that may assist potential terrorist supporters. The latter diversion is especially worrisome because these as-yet uncommitted individuals may then be won over by the terrorists. In cases where social services are inadequate, the terrorists can seize an opportunity to win over supporters (Berman 2003). If, however, the government does not take counterterrorism actions, then the terrorists can operate with impunity until the costs to the government of not conceding to terrorist demands outweigh the benefits of holding firm. Given their options, the terrorists and the government face not only a test of wills over who will gain the upper hand in a nonconventional conflict but also a competition over who will win over a base of potential supporters. The latter offers financial resources and loyalty to the side with the most attractive actions and political agenda. Although the literature on insurgencies and civil wars, on occasion, includes the role of potential supporters in formal models (e.g., Azam 2002; Grossman 1995; Mason 1996), terrorism models traditionally ignore the essential position of grassroots supporters. Even the nontechnical literature often leaves out the contest for supporters waged by the government and terrorists—for example, DeNardo (1985) views political-fringe groups as resorting to terrorism to overcome their small numbers and inability to mobilize popular support. When a contest between the government and the terrorist group is depicted in the literature, this contest is over something other than the base of potential supporters. For example, the contest can be over the distribution of rents (Kirk 1983), the fate of hostages (Lapan and Sandler 1988; Selten 1988), the securing of concessions (Bueno de Mesquita 2005a; Kydd and Walter 2002), or the effectiveness of countermeasures (Enders and Sandler 1993). The primary purpose here is to fill this void by characterizing the competition for popular support between a terrorist organization and a government.1 We also seek answers to other questions, such as the role of adversarial leadership and terrorist sponsorship. Since the late 1970s, sponsorship has been a key issue in terrorism. This sponsorship can take at least three forms: Diaspora support (e.g., funding for the Tamil Tigers in Sri Lanka and the Irish Republican Army in Northern Ireland), state sponsorship (e.g., Libya’s support of the downing of Pan Am flight 103), and outside financing (e.g., al-Qaida franchising of terrorist groups) (Byman et al. 2001). We begin by incorporating various assumptions about the relative strengths of adversarial interests. At the outset, we assume that the government and terrorist group 1. Our analysis differs from that of Rosendorff and Sandler (2004), who investigate how heavyhanded actions of a government may lead to terrorist recruitment and large-scale events. This theme of government-induced radicalism also characterizes de Figueiredo and Weingast (2001), where moderate terrorist supporters are incited to violence by harsh antiterrorism measures. We are interested here in support and sponsorship rather than recruitment of operatives, which is also why our analysis is different from Bueno de Mesquita (2005b). The latter interesting article addresses why terrorist operatives are educated (owing to screening) and why economic downturns may augment the level of terrorism. Our analysis also differs from some of the literature (e.g., Kydd and Walter 2002; Bueno de Mesquita 2005a) because we do not divide the terrorists into extreme and moderate camps. 880 JOURNAL OF CONFLICT RESOLUTION move simultaneously without the knowledge of their opponent’s activity level. We then use the resulting equilibrium to ascertain whether one of these adversaries may profit by costlessly precommitting to an action. Two scenarios are identified. In the first, the terrorist group prefers to scale back operations when seizing the initiative owing to a fickle support base, whose allegiance decreases with a government’s offensive. The terrorists’ initiative to curb their campaign protects them from being overwhelmed by effective countermeasures. In the second scenario, the terrorist group prefers to move after government countermeasures foster grassroots support for the group by those who react negatively to the government’s neglect of their interests. Governmental countermeasures increase popular support for the terrorists by keeping a segment of the population impoverished as social programs are either curtailed or put out of reach. This scenario characterizes the Palestinian situation and the ability of Hamas to build up popular support by providing social services and attacking the alleged oppressors (Hilsenrath 2005). Israeli counterterrorism measures, which eventually closed off Israel to many Palestinian workers, limited labor-associated health care benefits and other public services (e.g., public transport) for Palestinians (Berman 2003; Hilsenrath 2005). As such, the Israeli government effectively traded off antiterrorist actions for public goods, thereby leaving an opening for others to provide these goods. Another instance is southern Lebanon, where Hezbollah has gained grassroots support among a neglected constituency. Despite the bolstering of terrorist support, a government may be willing to preempt the terrorists because such actions avoid even greater losses associated with the terrorists seizing the initiative. The two scenarios indicate that strategic precommitment by an adversary curbs terrorist activities. Nevertheless, the terrorist group gains because it can better adjust its militant campaign based on resources garnered from supporters. An outside party, bent on more militant action, will be disappointed by the potentially low level of hostility. If the outside sponsor can command and direct resources to the terrorist group, then the financier can limit the moderation in violence that stems from efforts to win popular support and exercise strategic precommitment. Thus, al-Qaida’s actions to franchise new groups and to bolster old ones (e.g., Abu Sayyaf in the Philippines) serve an insidious purpose by curtailing inherent checks on violence. The body of this article contains four sections. The first section presents the model in terms of its three agents: the terrorists’ support base, the terrorist group (or its leader), and the government. In the next section, we investigate gains from leadership by the terrorist group and the government under two alternative assumptions about the tolerance of the terrorists’ support base for government neglect. The third section introduces a fourth player—a financier or sponsor of the terrorist group. Concluding remarks follow in the final section. STRATEGIC PLAYERS: SUPPORT BASE, TERRORIST GROUP, AND THE GOVERNMENT Each of the three elements is presented in a simple form. Both the terrorist group and the government try to capture the loyalty of a larger proportion of a potential Siqueira, Sandler / TERRORISTS VERSUS THE GOVERNMENT 881 pool of supporters, who are assumed to be risk neutral. The timing of the game is as follows: in the first stage of the game, the government and the terrorist group act simultaneously while taking the best response of their counterpart as given. The government chooses the amount of public services and the level of counterterrorism, while the terrorists determine the magnitude of their campaign and the consumption of a good (unrelated to terrorism). In the second stage of the game, the potential pool of supporters decides its allegiance to a side while taking the first-stage equilibrium activity levels of the two adversaries as given. We use backward induction to ascertain the subgame perfect equilibrium by first finding the partition of supporters and then determining the first-stage choices of the government and terrorists. THE POTENTIAL SUPPORT FOR TERRORISM We let p denote the total number of potential supporters of the terrorist group. These supporters are uniformly indexed by δ on the unit interval [0,1], where individuals who actively support the group display a higher index number than nonsupporters. An active supporter gains a net payoff, H, from the expected payoffs associated with the terrorist group’s success or failure and the supporter’s preference for the group (denoted by δ ) and loses his or her fixed contribution of σ to the terrorists. We denote the probability that the terrorist group fails (succeeds) in meeting its goals as π (1−π). Moreover, the terrorist supporter obtains h(a; s) from a terrorist success, s, and h(a; f) from a terrorist failure, f, where variable a denotes the terrorists’ level of action. Active supporters of terrorism not only care about the group’s success but also about its militancy. These actions also place greater costs on the targeted government. A more sustained terrorist campaign may assuage supporters’ anger from either frustration or being ignored. Greater terrorist actions may raise the payoff from success. Obviously, h(a; f) < h(a; s). We also assume that the marginal gain from success exceeds the marginal gain associated with failure, so that ha(a; f) < ha(a; s). Moreover, payoff function h is assumed to be strictly increasing and strictly concave in a, so that payoffs display diminishing returns. Terrorist supporters benefit to some extent from government spending, g, on public goods, intended for nonsupporters of the terrorists. Owing to imperfect exclusion, terrorist supporters gain θ− b(g) where − b(g) is increasing and concave in g, and θ ∈(0,1) represents the degree to which government spending benefits terrorist supporters. The government keeps θ smaller than 1 by limiting the supply of g to areas where terrorists are known to hide or train. Another reason for θ < 1 is that some terrorist supporters view government-provided goods as tainted and refuse to partake. The supporters’ payoff function equals − H(e,a) = π (e,a)h(a; f) + [1−π (e, a)]h(a;s) + θ b(g) + δ − σ, (1) where e represents the counterterrorist action of the government, which increases the terrorists’ likelihood of failure, πe > 0. In contrast, terrorist efforts serve to decrease
the terrorists’ likelihood of failure, πe < 0. We also assume that πee < 0, πaa > 0, and
882
JOURNAL OF CONFLICT RESOLUTION
πea > 0. Thus, the failure probability is strictly concave in e and strictly convex in a.
The assumed sign of πea reflects the technology of conflict that exists between the
adversaries as well as the government’s advantage over the terrorists: increased
action by the terrorists raises the ability of the government to increase terrorists’ failure as they expose themselves to greater risk and governmental scrutiny. The inability of terrorists to increase operations without risking greater jeopardy from
counterterrorism reflects the view of terrorism as asymmetric warfare, where a weak
opponent gains an advantage by remaining hidden (White 2003, 286). Greater terrorist activity exposes not only operatives but also infrastructure.
For those who opt not to support the terrorists, we represent their preferences by2
B(g) = −
b(g) + (1−δ),
(2)
where −
b reflects expenditure on public projects that benefits those who do not
actively support the terrorist group. Our normalization of the coefficient of −
b to equal
1 is consistent with the government’s actions to reduce supplies of g to terrorist-ridden areas, so that nonsupporters gain relative to supporters. In (2), those individuals
with a higher δ (i.e., a greater allegiance to the terrorists) receive less utility when
withholding their active support for the terrorists.
Let δ̂ represent the potential supporter who is indifferent from actively supporting
the terrorist group and not supporting it. To find δ̂, we equate (1) and (2) and solve
δ̂ =
b(g) − π(e, a)h(a; f ) − [1 − π(e, a)]h(a; s) + 1 + σ
,
2
(3)
where b(g) = (1−θ )−
b(g) represents non-supporter-specific public goods gains, not
received by terrorist supporters due to imperfect exclusion. Equation (3) determines the
equilibrium level of support for the terrorist group for given values of a, g, and e. Thus,
those individuals with a δ greater than δ̂ will support the group, while those with a δ
smaller than δ̂ will not.
To determine the impact of changes in terrorist and government activity on the terrorist support base, as reflected by the indifferent supporter, we must first specify the
relationship between the two variables under direct control of the government (i.e., g
and e). This is given by the government’s budget constraint, g+αe = β, where α represents the per unit costs of government effort directed against the terrorist group, and
β represents the (fixed) budgeted amount of money allocated to countering terrorism.
Our model acknowledges that g may curb terrorism by helping the welfare of those
who may be swayed by the terrorists’ political agenda. After solving the constraint for
2. Our specification assumes that the terrorist group lures supporters with increased terrorist activity rather than through coercion, in the form of retribution for a nonsupporter. Furthermore, we assume
that nonsupporters of the terrorists do not free ride on terrorist activities. If, however, we were to allow
for nonsupporter free riding, then this would reduce terrorist support levels. Less grassroots support
enhances terrorists’ need for outside sponsorship. To simplify the exposition, we ignore the imperfect
excludability of terrorist actions. Also, we believe that most nonsupporters do not gain from violence that
could be directed at them.
Siqueira, Sandler / TERRORISTS VERSUS THE GOVERNMENT
883
g and substituting the result in (3), we perform comparative statics and obtain the following results:

∂ δ̂
1
= − πa h(f ) − h(s) + πha (f ) + (1 − π)ha (s) < 0, ∂a 2 (4) ∂ δ̂ 1 = − αb + πe h(f ) − h(s) 0, ∂e 2 (5) where we henceforth suppress a in h(a; f), h(a; s), and ha(.). Associated second-order partials of δ̂ necessary for subsequent results, are gathered in Appendix A. Equation (4) shows that increased terrorist activity, a, augments the terrorist group’s support by lowering the threshold that makes a potential supporter indifferent between fostering and not fostering the group’s efforts. Given our assumptions, the sign of ∂ δ̂/ ∂e is ambiguous. If, however, αb′ > −πe[h(f)−h(s)], then ∂ δ̂/ ∂e
< 0. That is, the impact of a change in e (via a reduction in government social spending) for those who do not actively support the terrorist group exceeds the counterterrorism influence of e on terrorist failure. Hence, by expending more effort to increase the likelihood of a terrorist failure, the government reduces spending on nurturing its political base of terrorist nonsupporters. This spending decision can then enhance the support base of the terrorists by lowering the threshold level of their supporters. A smaller θ in (1) or (3) makes for more non-supporter-specific public good gains and, therefore, a greater likelihood of this scenario. When counterterrorism does more to increase the terrorists’ support base than to curb terrorism, such measures become counterproductive. This adverse outcome is through the government’s budget constraint and the limited impact of counterterrorism. Examples would be the rise of Hamas and Hezbollah, where constituencies were left with limited social services as government attention was either focused on counterterrorism or elsewhere. By not providing public goods to the Palestinians, Israel gave Hamas an opportunity to attract supporters. Other cases would include nationalist-separatist situations (e.g., Algeria during its fight for independence in the 1950s and early 1960s), where a government ignored the needs of the general population as resources were redirected to the antiterrorism campaign. This scenario also implies increased alienation of a segment of the population and counterterrorism actions with limited effectiveness. Henceforth, this case is equated with a strong support base for the terrorists. If, however, αb′ < −πe[h(f)−h(s)], then ∂δ̂/ ∂e > 0. Now, an increase in government
countermeasures, which directly target the terrorist organization, results in an
increase in the threshold level for active terrorist support. This represents a narrowing
of the base of active support for the terrorist organization since only those individuals with a relatively high δ will remain engaged in actively financing the terrorists
group. In this case, government resources may be more effective in limiting terrorist
success than in providing for the needs of potential supporters of the terrorists. If the
884
JOURNAL OF CONFLICT RESOLUTION
general population is already well off, then the opportunity cost of reduced social
programs is small. For example, αb′ was probably small in the 1970s and 1980s in
Europe during the era of left-wing terrorism. This is consistent with θ near 1 in value
when public goods are effectively nonexcludable. The government would have had a
difficult time excluding terrorist supporters during this period since there were few
factors, such as location, to identify then. Hence, effective countermeasures (πe is
large) eventually reduced support for these nihilistic terrorist groups (e.g., the BaaderMeinhof group), thereby raising potential supporters’ thresholds. Thus, we characterize this case as a fragile support base for the terrorists.
THE TERRORIST ORGANIZATION
Let the interests of the terrorist organization be represented by the following utility function:
W = w(a) + x,
(6)
where w is increasing and concave in a. This specification is consistent with the benefits to the organization being derived from the furtherance of some militant activity
a. In addition, the terrorist organization gains utility from the consumption of good
x, unrelated to terrorism. This interpretation of (6) is consistent with Stern’s (2003)
findings that, despite their claim to further strictly religious goals, many terrorist
groups serve a mix of sacred and profane interests. The specification also agrees
with the rule of a secular leader who fulfills not only the interests of supporters but
also his own goals or self-interest. In (6), the probability of terrorist success (or failure) does not directly enter the objective function of the terrorist organization. This
does not mean that the organization is uninterested in outcomes; it indirectly cares
about success through its influence on the mobilization of supporters, felt through
the organization’s budget constraint. If we also allow success (failure) to directly
affect W in (6), our results are somewhat changed.3 Because today’s transnational
terrorism is predominately driven by fundamentalist terrorists, who are more interested in the deed than in achieving some goal (Hoffman 1998), we rely on the objective in (6) in the text. This objective is also apropos of nihilistic leftist terrorists of
the 1970s and 1980s whose goals were poorly defined.
To account for the costliness of terrorist activities and operations, we represent the
cost of operations by a strictly increasing, convex function k(a). The terrorist organization obtains σ from each active supporter and must spend c on administering a sup− = σ − c > 0, so that a net contribution is
porter, where it is reasonable to suppose that σ
3. An alternative way of specifying the terrorists’ objective is to write it as
W = [1−π (e, a)]r + π (e, a) −r + x = r −π (e, a) (r − −r) + x = w(e, a) + x,
r represent the (constant) benefits of terrorist success and failure (with r > −r ) and
for which r and −
r ). The signs on the partials of w(e, a) follow from −π (e, a). For further details
w(e, a) ≡ r−π (e, a) (r − −
on this alternative specification, see footnote 5.
Siqueira, Sandler / TERRORISTS VERSUS THE GOVERNMENT
885
derived. We denote the terrorist group’s fixed endowment by y. Thus, the terrorist
organization’s budget constraint is
− (1 − δ̂) p,
x + k(a) = y +σ
(7)
where expenditures are on the left side (assuming the unit price of x is 1), and net
income is on the right side of (7). Given that 1− δ̂ represents the proportion of potential supporters from population p who actively finances the group in equilibrium, the
term σ− (1− δ̂) p represents the net amount of contributions received by the group
from its active base. Solving (7) for x and substituting the result into (6), we represent the terrorist group as maximizing its utility:
− (1− δ̂) p − k(a),
W = w(a) + y + σ
(8)
given the level of counterterrorism, e, of the government. The first-order condition is4
–p ∂ δ̂ − k = 0.
w − σ
∂a
(9)
In (9), the first two terms depict the marginal benefits of increased terrorist activity
derived from the organization’s activity and the larger net contributions from supporters, while the third term is the marginal costs of terrorist activity. Thus, the terrorists
choose their actions to equate the associated marginal benefits and marginal costs.
To further characterize the terrorist group’s behavior in terms of how it responds
to the government’s counterterrorism efforts, we implicitly differentiate (9) to determine the slope of the terrorist group’s best-response curve:
–p ∂ δ̂
σ
∂a∂e
< 0. 2 --p ∂ δ̂ − k w − σ ∂a 2 2 ∂a = ∂e (10) The terrorist group’s best-response curve will be downward sloping because the denominator in (10) is negative (see footnote 4) and ∂2δ̂/ ∂a∂e is positive (Appendix A). The best response for the terrorist group to an increase in the government’s counterterrorism efforts is to decrease its terrorist activity. To determine whether the group’s welfare is increasing or decreasing in government efforts, we differentiate W with respect to e: --p ∂ δ̂ . We = −σ ∂e (11) --p ∂ δ̂ − k < 0. Since we have assumed ∂2δ̂/ ∂a2 > 0
4. The second-order condition requires w − σ
∂a 2
(in Appendix A), the second-order condition is satisfied.
2
886
JOURNAL OF CONFLICT RESOLUTION
When the threshold level for active support of the terrorist group is increasing in
e, We < 0, so that the group welfare falls with greater countermeasures. If, however, government countermeasures decrease the threshold level and induce more active support for the movement (∂δ̂/ ∂e < 0), then We > 0.5
THE GOVERNMENT
The government’s preferences with respect to the terrorist problem are as follows:
U(π, g, n) = π (e, a) + u(β − αe, pδ̂)
(12)
where the government’s budget constraint has been substituted for g. In (12),
pδ̂ denotes n, the number of potential supporters who chooses not to actively aid the
terrorist organization. The government gains satisfaction from the likelihood of a terrorist failure and the welfare of nonsupporters. The subutility function, u(g, n), has
the following properties: ug > 0, un > 0, ugg < 0, unn < 0, and ugn > 0. For a given level
of terrorist activity, the government’s problem is then to choose the counterterrorism
effort—and thus the spending on potential terrorist supporters—to maximize (12).
The resulting first-order conditions is given by6
∂ δ̂
= 0,
∂e
(13)
∂ δ̂
= αug .
∂e
(14)
πe − αug + un p
which, upon rearrangement, gives
πe + un p
ˆ ∂e > 0, then the marginal benefits of increased counterterrorist measures from
If ∂δ̂/
more terrorist failure and reduced support equal the associated marginal costs from
not nurturing potential terrorist supporters. If, however, ∂δ̂/ ∂e < 0, then un p ( ∂δ̂/ ∂e) 5. If w(e, a) replaces w(a) in the terrorists’ objective function and we normalize −r to zero (see footnote 3) so that the probability of success directly influences the terrorists’ welfare, then (11) becomes − p(∂δ̂ / ∂e) since w = −π r. The first and second terms on the right-hand side respectively We= −πe r − σ e e represent the counterterrorism impact on terrorist welfare of the security effect on π and of the mobilizing effect on terrorist support. The security effect is always negative, but the mobilizing effect may be of either sign. Consequently, there are three cases. If ∂δ̂ / ∂e > 0, then We < 0. When, however, ∂δ̂ / ∂e < 0, the sign of We depends on the relative strengths of the two opposing effects. If the mobilizing effect is small relative to the security effect, then We is still negative. For a sufficiently large mobilizing effect, we have We > 0. This is consistent with a large base of potential supporters for the terrorists, which is the
worst-case scenario for the government. In all cases, the impact of government countermeasures on the
mobilization of terrorist support remains the key determinant of We. Regardless of the specification, the
results hinge on two scenarios later displayed in Figures 1 and 2.
6. The second-order condition,
πee + α2 ugg − 2αugn p
∂ δ̂
+ unn p2
∂e

∂ δ̂
∂e
2
+ un p
∂ 2 δ̂
< 0, ∂e2 is always satisfied for ∂δ̂/ ∂e > 0. If ∂δ̂ / ∂e < 0, then the second-order condition holds provided that the value of the middle term is outweighed by the absolute value of the sum of the other four terms. Siqueira, Sandler / TERRORISTS VERSUS THE GOVERNMENT 887 represents an additional marginal cost of counterterrorism efforts as government’s neglect of potential supporters of the terrorists augments their support base. From an implicit differentiation of (13), we have the slope of the government’s best-response function: ∂e = ∂a ∂ 2 δ̂ ∂ δ̂ ∂ δ̂ ∂ δ̂ − p2 unn − pun ∂a ∂a ∂e ∂e∂a , 2 ∂ δ̂ ∂ 2 δ̂ ∂ δ̂ 2 2 + unn p πee + α ugg − 2αugn p + un p 2 ∂e ∂e ∂e −πea + αpugn (15) which is assumed to be positive.7 Finally, differentiating the government’s objective function with respect to a shows that government utility is decreasing in a. That is, Ua = πa + pun ∂ δ̂ < 0. ∂a (16) With a on the vertical axis and e on the horizontal axis, the direction of increasing utility for the government is down and to the right, where terrorist attacks fall. STRATEGIC INTERACTION The Nash equilibrium of the first-stage, simultaneous-move game is obtained by solving (9) and (13) for the values of e and a. This, in turn, determines at which point a potential supporter is indifferent about actively supporting the terrorist group or not.8 There are two scenarios of interest based on whether the sign of ∂δ̂/ ∂e is positive or negative.9 We examine each case separately. SCENARIO 1: ∂δ̂/ ∂e > 0 —THE CASE OF A FRAGILE SUPPORT BASE
This case supposes that counterterrorism can limit terrorist success without
unduly alienating potential terrorist supporters by neglecting their needs (i.e.,
αb′ < −πe[h(f)−h(s)]), so that an increase in e raises the threshold level for active support of the terrorist group and results in less support for the movement. This implies 7. The denominator in (15) is negative when the second-order condition is satisfied, so that the sign of the whole expression depends negatively on the sign of the numerator. When ∂δ̂/ ∂e > 0, all terms in
the numerator are negative, and the government’s best-response function is upward sloping. If ∂δ̂/ ∂e < 0, the sign of (15) remains positive, so long as the expression associated with this term remains relatively small compared to the rest of the numerator. 8. An equilibrium of this first-stage game exists in pure strategies if (1) the government’s and terrorist group’s strategy sets are convex, closed, and bounded; (2) their payoff functions are continuous; and (3) their payoff functions are concave (Dasgupta and Maskin 1986). Condition (1) is satisfied since the adversaries’ activities are constrained by their budget sets, while condition (3) holds because utility functions are strictly concave. Condition (2) is satisfied provided that δ̂ remains interior, between 0 and 1. Equilibrium is unique since the players’ payoff or utility functions are strictly concave. 9. Without a specific payoff function, our condition for the sign of ∂δ̂/ ∂e is not just a function of the model’s parameters but is endogenously dependent on e. 888 JOURNAL OF CONFLICT RESOLUTION a IPt’ BRg IPt N L’ BRt IPg L 0 Figure 1: e Government and Terrorists Strategic Interaction: Scenario 1 that the iso-payoff contour for the terrorist group resembles that of a backward “C” in Figure 1, where contours representing higher levels of payoffs lie to the northwest, where there is less counterterrorism spending for each level of terrorist effort. The best-response curves (BR) and iso-payoff contours (IP) for the terrorist group and the government, along with the resulting Nash equilibrium (point N), are depicted in Figure 1. Superscript t and g refer to the terrorists and government, respectively. The nature of this first scenario and equilibrium is that potential terrorist supporters will not offer their support when faced with greater antiterrorism effort. This occurs even though more spending on counterterrorism implies less spending on g to raise the welfare of terrorist nonsupporters. The terrorist group accounts for its fragile support base when determining its optimal level of militant activity. The characteristics of the scenario are more likely to arise when the marginal benefit of government spending with respect to g is quite low, which can result if g levels are already high so that b′(g) is small. This marginal benefit is also small when potential Siqueira, Sandler / TERRORISTS VERSUS THE GOVERNMENT 889 supporters are unresponsive to changes in g, which can occur if the public goods’ benefits are received equally by terrorist supporters. Scenario 1 is also likely to arise when government counterterrorism policy is particularly effective, in which the marginal impact of government actions on the probability of terrorist failure is large. Leftist terror campaigns of the 1970s and 1980s best fit the first scenario. The Red Brigades in Italy, Direct Action in France, the CCC in Belgium, and the Red Army Faction in West Germany needed to limit their brutality to woo supporters. The general public never “bought into” the terrorists’ ill-defined nihilistic goals that threatened most people’s structured lives. These groups’ potential support base was very fragile. As targeted governments redirected some spending with mostly nonexcludable benefits to counterterrorism, there was no groundswell of support for the terrorists. More important, the governments’ counterterrorism measures were quite effective in infiltrating groups and gaining strategic intelligence. By the end of the cold war, these four groups had either been wiped out or disbanded (Hoffman 1998). Other left-wing European groups suffered similar fates. Figure 1 also offers some insights into the players’ behavior and their willingness to act strategically.10 If, for example, players had the ability to precommit and obtain a higher payoff in the resulting leader-follower equilibrium, then the terrorist group would prefer to move first and the government would prefer to move second. This is depicted by the shift in the iso-payoff contour from IPt to IPt′, with the leaderfollower equilibrium at L, where IPt′ is tangent to the government’s best-response curve. Compared to the Nash equilibrium at N, the terrorist group and the government are better off at L. The terrorists are on a higher utility level at L since iso-payoff contours to the west represent improved terrorists’ payoffs. For the government, the iso-payoff curve through L (not drawn) is also a better outcome than the one through N. If, however, the government leads, then the associated leaderfollower equilibrium is at L′ (with the government’s iso-payoff curve tangent to BRt). At L′ the terrorists are worse off than at L or N, and the government is not as well off as at L. For this game, the terrorists have a clear first-mover advantage, and the government has a clear second-mover advantage. At L in Figure 1, the terrorists’ actions and the government’s countermeasures are smaller than at the Nash equilibrium. The outcome is surprising since the terrorists lower their own level of violence when assuming a leadership role. Even though their reduced campaign hurts the terrorists’ ability to meet their goals and obtain grassroots support, the terrorists are still better off because of a significant drop in the government’s effective antiterrorist efforts. Owing to their weak support base and the formidable threat of the government, terrorists gain more in this scenario by seizing the initiative and turning down the heat. De-escalation by both the terrorist group and government makes both parties better off. Contrary to intuition, the government does not necessarily have to move first and preempt the terrorist group to arrive at a superior outcome. The terrorists prefer to limit their violence to operate under the 10. Eaton (2004) provides a similar analysis. However, his study covers symmetric social dilemmas in which both players’ best-response curves possess the same slope, and players’ iso-payoff contours have a similar orientation to their own axis. 890 JOURNAL OF CONFLICT RESOLUTION BRg a IPt BRt L’ N L IPg IPg’ 0 Figure 2: e Government and Terrorists Strategic Interaction: Scenario 2 government’s radar. Many of the left-wing groups reduced their campaigns in the late 1980s when the targeted governments became more adept at rooting them out and supporters were harder to attract (Hoffman 1998). SCENARIO 2: ∂δ̂/ ∂e < 0 —THE CASE OF A STRONG BASE If a dollar of government spending is more effective in providing excludable benefits to nonsupporters of terrorism than in curbing terrorism (i.e., αb′ > −πe[h(f)−h(s)]),
then an increase in e lowers the threshold level for active support for the terrorist group
and enhances the backing for the group. This then reverses the convexity of the terrorists’ iso-payoff curves, which are now C-shaped in Figure 2, where curves to the southeast represent greater levels of well-being (recall the discussion surrounding (11)). The
best-response curve for the terrorists is still downward sloping. For the government, its
Siqueira, Sandler / TERRORISTS VERSUS THE GOVERNMENT
891
iso-payoff curves and best-response path are the same as in the first scenario. In Figure
2, the Nash equilibrium, N, corresponds to where respective iso-payoff curves for the
terrorists and the government achieve a maximum for the best-response choice of its
adversary.
A study of Figure 2 indicates that the government prefers to assume the initiative in
the second scenario and move first. By so doing, the government obtains a higher payoff at the associated leader-follower equilibrium at L as IPg shifts to IPg′—the government’s iso-payoff level through L is better than that through N. This strategic move
increases its counterterrorism measures, while it results in a decrease in terrorist attacks.
Because the increase in counterterrorism has a larger (adverse) impact on nonactive,
potential supporters of the terrorist group than on the terrorists, the government’s offensive actually augments terrorists’ welfare. This follows insofar as the terrorists’ iso-payoff curve (not shown) through L is better than that through N. By seizing the initiative,
the government avoids a worse outcome in which the terrorists take the leadership role
and increase their attacks at point L′, where IPt (not displayed) is tangent to BRg. At L′
the government is not only worse off than at N or L, b…

Turn in your highest-quality paper
Get a qualified writer to help you with

“ Simple question ”

Get high-quality paper

Guarantee! All work is written by expert writers!

Still stressed from student homework?

Get quality assistance from academic writers!

Order now