Reading I NEED HELP ON MY HOMEWORK

The Charitable Activities of Terrorist Organizations

Author(s): Pierre-Emmanuel Ly

Source: Public Choice , Apr., 2007, Vol. 131, No. 1/2 (Apr., 2007), pp. 177-195

Published by: Springer

Stable URL: https://www.jstor.org/stable/27698091

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

Springer is collaborating with JSTOR to digitize, preserve and extend access to Public Choice

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

Public Choice (2007) 131 177-195 DOI 10 1007/slll27 006 9112 2

ORIGINAL ARTICLE

The charitable activities of terrorist organizations

Pierre-Emmanuel Ly

Received 10 March 2006 / Accepted 3 October 2006 Springer Science + Business Media B V 2006

Abstract Violent groups sometimes invest significant resources in social work, notably in the form of chanties and NGOs The present paper models a terrorist group's chanties as a means to advertise its cause in order to raise popular support The analysis explains how different types of organizations anse m equilibrium, depending on government policies Then, the interaction between a purely terrorist group and an independent local NGO is examined It is shown that a purely terrorist group always invests in more attacks than an integrated terrorist-chanty organization Furthermore, the latter may have more NGO activity than a separate local NGO

Keywords Terrorism NGOs Chanties

1 Introduction

The war on terror launched by George W Bush in 2001 has generated interest in the pro files of organizations and individuals who engage m terrorist activities Anti-terrorist poli cies worldwide have exhibited particular efforts to detect and destroy the organizations' financial resources These efforts have revealed the highly diversified nature of terrorists' asset portfolios In particular, it is now well known that some violent groups receive money from charitable and humanitarian organizations Several important international Non Gov ernmental Organizations (NGOs) had to cease their activities because they were suspected of transferring funds to terrorist groups For example, the Holy Land Foundation for Relief and Development was closed in December 2001 when the FBI discovered it was funding the Palestinian Hamas (Levitt, 2003) Levitt (2002) reports that "the US government estab lished that funds were used to support schools and indoctiinate children to grow into suicide bombers" Then, the author criticizes the European Union for drawing "a fallacious distinction between the nonviolent activities of terrorist groups and the terror attacks that

P-E Ly Universit de Toulouse 1 (ARQADE), 21 All es de Bnenne, 31000 Toulouse France email pierremly @ yahoo fr

; Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

178 Public Choice (2007) 131 177-195

they carry out" This debate is taking place in the context of the fight against terrorism fi nancing Indeed, some groups classified as terrorist undertake very efficient relief operations and development programs for the poor, and in addition to the fund-raising motive, these programs enhance the welfare of thousands of people in need In his work on Islamic NGOs, Ghandour (2002) reports that The Palestinian Hamas1 devoted 95% of its budget to maintain an important network of local NGOs, which makes it one of the most important actors in the local development community Furthermore, he argues that in Algeria, after an earthquake in October 1989, NGOs affiliated to the Islamic Salvation Front provided aid to the victims

more efficiently than the government Such examples show that an important purpose of these charitable initiatives is to gain support from the local population, and that the analysis has to go beyond the issue of money laundering

This paper considers charitable investments by terrorist groups as a way for them to advertise their ideals among potential sympathizers Indeed, charities not only provide a conduit for money laundering, they also truly benefit people in need As a result, those who at least partly share the goals of the terrorist group are likely to be more willing to make their contribution to the fight The analysis uses a simple theoretical model where the terrorists' chanties are assumed to have an advertising effect on people's preferences Then, it is shown how different types of activist organizations may arise, ranging from those specializing in charity to those which only produce terrorist attacks The interaction between terrorists and local chanties is also examined

The term "terrorist" for the purpose of this paper will refer to any violent, primarily domestic group in a country, that is not part of the government and has a political agenda Such groups are often termed "terrorist" by some while being classified simultaneously as rebels, guerrilla fighters, etc The paper is partly inspired by the wider conflict literature where groups are often termed rebels The basic idea presented here also comes from Che Guevara (1960)'s theory of guerrilla warfare, where he underlines the necessity for a rebel group to help the local population, not only to get food and clothes from them but also because the guerrilla is supposed to represent the people's aspirations

To sum up, the term used in the present analysis will be "terrorists", since the paper's motivation comes from the observation that groups often classified as "terrorist" provide charitable services for the poor, and the organization's objective function is modelled fol lowing recent economics models of terrorism However, the paper aims at providing more general insights on various kinds of violent groups for which support from the population may matter

The model presented here belongs both to the literature on the economics of conflict and to the strand of formal papers dealing more specifically with terconsm Azam (2005) assumes that agents care about the welfare of their descendants who might benefit from a public good

with some probability The latter can be increased by engaging in bombing, and agents are willing to give up some of their consumption today and contribute resources in favor of the next generation

Bueno de Mesquita (2005) presents a model where a terrorist group screens applicant volunteers and only recruits the best ones The author argues that people with little educa tion, few economic opportunities and anti-government feelings, are more likely to volunteer The model establishes a link between poor economic conditions and mobilization in favor

1 The Hamas won democratic elections in the Palestinian Territories in January 2006 While the group's change in status raises issues that are outside the scope of this paper, the election is clear evidence of the group's popular support The information provided about Hamas in the rest of the paper is based on its activities prior to being elected

ti Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

Public Choice (2007) 131 177-195 179

of terrorist groups, while explaining why some recent data (Krueger & Maleckova, 2002) suggested that actual terrorists are relatively less poor and more educated than the average

member of the population According to the author, this is due to the screening of applicants by terrorist organizations, which makes the latter recruit people whom they think will be more qualified to follow instructions As a result, even if poor people have lower opportunity costs of participation and may support the terrorists, the screening process selects those who are more educated (and richer) to ensure the efficiency of attacks

Bueno de Mesquita and Dickson (2005) analyze a signaling model where a terrorist group cares about the level of support it gets from the population By engaging in a form of countert erronsm that may or may not be damaging to the local population, the government signals the extent to which it cares about its people's welfare After observing the government's policy, people decide whether to support the terrorist organization Then, in this model, the latter tries to provoke hard government crackdowns in order to increase people's involvement in terrorism Thus, in their paper, signaling comes from the government's crackdowns while, in the present paper, the ("advertising") signal comes from the terrorists, since their charitable activities allow them to promote their ideas among potential sympathizers

Farm and Arce (2005) propose a dynamic model where the number of new terrorist recruits

depends positively on the level of social support for terrorism Like in the present paper, the terrorists benefit from popular support, but the latter is assumed to depend positively on past levels of terrorist activity and on some exogenous underlying support However, the analysis in Bueno de Mesquita and Dickson (2005) shows that the population might also blame the terrorist group for the increased damage it suffers Therefore, the positive link between violence and social support for terrorism assumed in Farm and Arce (2005) seems questionable without further explanations, since the sign of this correlation clearly depends on various factors

Recent papers have investigated the case of extremist groups that contribute to local pub lic goods Berman (2003) adopts a club good approach to model groups like the Palestinian

Hamas He argues that "exclusion from access to the local public good [ ] is a common, viable form of enforcement" of various kinds of prohibitions and sacrifices However, even if such requirements do exist, extremist groups also provide their charitable services with out requiring contributions in return Berman and Laitin (2005), who use Berman (2003)'s club good model to rationalize the use of suicide bombing, acknowledge that their analy sis does not explain cases where public goods are provided in a nonexcludable way The present analysis is therefoie a useful complementary step toward the understanding of such phenomena

As mentioned above, the idea that politically motivated groups provide services to the population to advertise their cause is not limited to terrorism, and can be applied to many insurgent or revolutionary movements Ferrero (2004) models a two-product political cooper ative that produces both revolution and reform In his paper, reform consists of the production of a commercial good in addition to the group's political good Commercial production helps to build trust in the cooperative among customers, which improves the credibility of revolu tionary promises and thus enhances people's willingness to pay for the political good This is similar to the idea of advertising presented in the present paper The author uses his model to explain the "divergent national development within the socialist movements of the 20th century", some of which having focused on reform more than others Another interesting paper is Azam (2006) which aims to explain why some rebellion leaders are regarded as thugs while others are considered heroes The author points out the example of the Entrean People's Liberation Front, which "was going out of its way in order to gain popular support in the liberated areas by engaging in various forms of social spending"

&\ t Sponger

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

180 Public Choice (2007) 131 177-195

This paper also relates to the economics literature on NGOs The organizations' major characteristic in this literature is their relatively higher altruism and commitment to the beneficiaries compared to private for-profit firms and the government This is the basis of the analysis in papers like, Besley and Ghatak (2001), Jack (2001), and Hopkins and Scott ( 1998), among others There is also an older literature on the nonprofit sector and chanties, for

which (Rose Ackerman, 1996) provides a good survey A signaling explanation for charitable contributions can be found in Glazer and Konrad (1996), where people voluntarily contribute to private chanties to signal their wealth In the present paper, the motivation of the terrorists to invest in chanties comes from their willingness to contribute to public goods and to advertise for their cause

Section 2 presents some examples that have motivated the paper In particular, they suggest that violent groups often care about popular support and make significant efforts to increase it In Section 3, the basic model is developed It starts by assuming that the terrorists' chanties allow them to advertise for their cause It is shown that agents' voluntary contributions to the terrorist group may increase or decrease as a result of a higher level of terrorists' chanties Furthermoie, studying the terrorist group's investment decisions shows how different types of organizations arise as a result of exogenous parameters representing the government's development and counter-terror policies

Then, in Section 4, a violent group (specialized in terrorism) and a local NGO (specialized in charity) are modeled as separate, independent entities, whose respective activities impact each other's costs More precisely, it is assumed that tenonst activity increases the cost of NGO operations, while charity work by the NGO has an advertising effect that may benefit the terrorist group by increasing its support base but also by reducing its costs The Nash equilibrium level of NGO and terrorist activity is analyzed, and potential effects of international aid to the local NGO are discussed Finally, the analysis goes back to the case of an integrated terrorist-chanty, as described in Section 3 It is shown that the integrated group (maximizing the joint surplus of the NGO and the terrorists) always produces less terrorist attacks than a terrorist group acting independently of local NGO activities The comparison is more ambiguous for the level of charity, so that the joint surplus maximizing level of NGO activities may be higher or lower in the Nash equilibrium

2 Evidence of violence/charity mix by activist groups

There exist many examples of organizations which, in addition to their violent activities, devote significant time and resources to take care of the poor in a more or less institutionalized way In his theory of guerrilla warfare, Che Guevara (1960) underlines the social role of "guerrilleros" He argues that rebel groups must represent the aspirations of the people, and that the fighters' mission includes taking care of the local population, notably in rural areas He explains that rebel groups should always try to help villagers to solve their social and economic problems Good relationships with the locals should be established so as to initiate voluntary contributions to the guerrilla Such contributions can be donations of food and clothes, allowing some wounded fighters to stay at one's house, or even volunteering as a soldier The author insists on the importance ofthat help from local people for the rebellion to survive and to enhance its efficiency Such a theory of guerrilla warfare exhibits the main idea of the present paper, namely, that rebels (or whatever such groups are termed) can get closer to the people to signal that they care about them, thereby initiating more popular support for the fight

A Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

Public Choice (2007) 131 177-195 181

Closer to the inteipretation in terms of chanties and NGOs, extremist Islamic political groups have been investigated by Ghandour (2002) in his work on Islamic NGOs The Hamas is one of the most important Palestinian political organizations It is classified as "tenonst" and has been involved in 115 attacks between 1989 and 2002, including 40 suicide bombings 2 Bueno de Mesquita (2005) points out the very high popularity of this group, which was ecently confirmed by the electoral victory of the group's political wing

The Lebanese Hezbollah is a very popular group which has gained people's support thanks to various activities (Ghandour, 2002) First, it has been known as the major military force that finally made the Israeli aimy leave southern Lebanon Second, it is a political party

with elected representatives at the Lebanese Parliament Third, it has a wide netwoik of chanties and NGOs In her study of Islamic NGOs in Beirut, Fawwaz (2004) even defines the Hezbollah as one of the major Islamic NGOs She criticizes the development literature that defines NGOs as necessarily non-political organizations, and she further aigu s that political ideology enhances the efficiency of NGOs such as those affiliated to the Hezbollah

According to Ghandour (2002), sustained efforts put in social welfare activities have allowed groups such as Hamas and Hezbollah to gain and maintain an impoitant support base in local populations, while groups that focus on violence have not managed to establish such legitimacy

In Sri Lanka, The Liberation Tigers of Tamil Eelam (LTTE) is another example of a group classified as terrorist that is also well known for its investment in NGO activities, as underlined by Berman and Laitin (2005)

Finally, it is important to point out that violent political groups in developing countries aie not the only users of charitable investments as a strategic advertising tool Rich governments involved in war zones in poor countries may also use foieign aid, channeled by international development organizations, to improve their political influence Burnett (2004),3 a former employee of the World Food Program, reports his experience in Somalia "[W]e distributed 50-kilogram sacks of grain emblazoned with the American flag and 'Gift of the People of the United States of America ' Somalis readily accepted the aid but it was clear that our professed neutrality was suspect " Further, he reports words of the US Secretary of State Colin Powell, to NGOs in 2001 "just as surely as our diplomats and military, American NGO's are out theie serving and sacrificing on the front lines of freedom NGO's are such a force multiplier foi us, such an important part of our combat team" Investigating such cases would necessanly lead to a discussion of the importance of NGOs' neutrality in places where they opeiate, but this is outside the scope of the present paper

The next section develops a simple foi mai model to show how different types of violent organizations can arise

3 The model

Assume that a tenonst group is involved in two activities, organizing attacks against the government and providing public goods to the local population in the form of NGOs oi chanties It invests T directly in tenonst attacks and s in chanty sei vices The cost function of the group's tenonst wing is given by CA(T, s), while that of its charitable wing is given

1 Source The Institute for Counter Tenonsm at the Interdisciplinary Institute at Hertzhya, http //www ict org ii

1 John S Burnett New York Times, August 4, 2004, available at http //www globalpolicy org/ngos/credib /2004/0804fire htm

A Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

182 Public Choice (2007) 131 177-195

by CNis, T) These cost functions assume that the group's terrorist and charitable activities affect each other, in the way described as follows

First, it seems reasonable to assume that the existence of terrorist activity increases the

cost of charity operations, so that C^T(s, T) > 0 Indeed, chanties whose ideology resembles that of violent groups aie more likely to face investigations, and temporary or permanent shut

downs Moreover, this certainly impacts their fund raising abilities negatively Furthermore, CN is, T) is assumed increasing and convex in s, as well as increasing and convex in T

Second, the existence of the group's chanties is likely to facilitate the organization of attacks Thanks to its network of chanties, the group can rely on a more favorable environment and even use the chanties as place to hide fighters or weapons Then, it seems realistic to assume that Cfs(T, s) > 0 In addition, CAiT, s) is assumed increasing and convex in T, as well as decreasing and convex m s

Furthermore, the terrorist group is altruistic, in the sense that it cares about the total quantity

of development programs available to the population, g + s, g being the part provided by the government (whose behavior is taken as exogenous) These programs may include the provision of various private and public goods, such as credit, education and health centers But the key feature here is that the benefit from these programs is a public good (as in Besley & Gathak, 2001), so that the terrorist group derives utility not only from its own contribution but also from the government's Formally, this part of the group's utility function is given by Ojvig + s), where the parameter 6j reflects the group's valuation for public goods, and i>( ) is an mci easing and concave function

Then, let the increasing and concave function B(T) reflect the damage caused by the terrorists' direct attacks As in Bueno de Mesquita (2005), the group also caies about popular support for its cause, b This comes from sympathizers in the population who each voluntarily

contribute time b, in favor of the group's interests, and b = ^T,lbl is the total contribution These contributions in time may be thought of as various types of activities, i anging from participating in demonstrations to hiding weapons and wounded fighters in one's house, or even training to commit future attacks Then, the overall impact of the terrorist gioup is the sum of its direct attacks and popular support, given by B(T) + \ /b, where the positive parameter i/r captures the relative importance of popular support for the group In addition, it is affected negatively by the government's counter terror operations K, so that the net terrorist impact of the group is co(K)(B(T) + \ rb), where coiK) is a decieasing function

To sum up, the terrorist group's net utility function is given by

VT = 0Tvig + s) + cuiK)iBiT) + irb) - CA(T, s) - CNis, T) (1)

Assume that the population of potential teironst sympathizers is made of identical indi viduals who dei ive utility from their private consumption c and norn then contribution b, to the terrorist group which claims to repiesent their interests Each agent's time endowment is L and by spending some time / working, she dei ives an income (g + s)l Recall that g + s is the total amount of development programs, and it is therefore assumed that they increase people's marginal productivity of time at work In other words, these development programs can be thought of as productive social expenditures, in the spirit of Barro's (1990) productive government expend tuie

The representative agent j's utility function is given by

c, +a(s)w(fc,) (2)

where u(b,) is increasing and concave in b, and a(s) is mci easing in s

y Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

Public Choice (2007) 131 177-195 183

This requires further explanation First, u(bt) is the utility derived from fostering the tenonst group's goals Second, the tenonsts' charitable investments have two goals The first, because the group is altruistic, is to improve the productivity of people in the economy The second is to advertise the group's values and objectives, in order to gain people's support This is reflected in the function afs), the value attached to participation in the agent's preferences,

which is increasing in the tenonsts' investment in chanties Formally, the key feature in (2) is that the marginal utility of bt is increasing in s, that is, people's preferences are such that chanties (advertising) act as a complement to their individual contribution to tenonsm This aspect of the present paper is inspired by Becker and Murphy (1993) who model advertising as a good consumed by agents This way of considering advertising is particularly relevant in the present paper, since people benefit from (consume) the services provided by chanties,

while at the same time being exposed to the terrorist group's values For example, a school might be named after a martyr, or strong ideological bias can be introduced in teaching some subjects More generally, the chanties' staff can promote the cause that the terrorist group claims to fight for, like the right of a people to have its own land, or freedom from occupation by foreign military forces

Note that, while operating local chanties can be a very efficient conduit for promoting the ideals of violent extremist groups, the idea of an advertising effect of charity is not limited to the activities of such violent groups Many development NGOs invest in advocacy and not only in service provision For example, environmental NGOs not only finance programs to save animals and forests, but also aim to educate people around the issues involved

Azam (2005b) assumes that education changes people's "world view", in the sense that it changes the value they attach to tenonst activity This is in line with his previous paper on suicide bombing (Azam, 2005a) where tenonst attacks aim to increase future generations' welfare, and the author argues that more educated people are likely to be more sensitive to the fate of the next generation These papers underline the importance of people's perception of the tenonsts' cause In the present model, people are exposed to the group's advertising when they come to affiliated chanties, and this is assumed to affect positively their valuation for the cause

Furthermore, note that Equation (2) implicitly assumes that people only derive utility from their own contribution to tenonsm and not from the resulting impact of the group, oj(K)(B(T) + i/sb), which implies that their decision is not directly impacted by the level of government counter tenor K This is equivalent to assuming a warm glow type of altruism However, it will be shown that K affects the equilibrium level of support, via its effect on the tenonst's investment in chanties

The timing of events is as follows First, given the (exogenous) government policies K and g, the tenonst group chooses its direct investment in attacks, T, and its investment in chanties, s

Then, taking s as given, agent i maximizes her utility given by (2), subject to her budget and time constraints

(g + s)l = ct (3)

bt+l<L (4)

The budget constraint assumes that development programs are provided for free As far as the chanties operated by tenonst groups are concerned, previous literature has argued that they are not equally available to all members of the population More specifically, services are sometimes provided conditional on some level of involvement in the group's

ti Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

184 Public Choice (2007) 131 177-195

ideology and activities (Berman, 2003, Berman & Laitin, 2003) However, even if such requirements do exist, extremist groups also provide their charitable services without requiring contributions m return

Using the constraints (3) and (4), the agent's optimization problem is solved by choosing bt so as to maximize the following

(L-bt)(g + s) + a(s)u{bt) (5)

subject to the constraints bt < L and bt > 0 The model is solved by backwards induction, starting with the agent's utility maximization

for a given level of terrorist chanties For simplicity, all agents in the population of potential terrorist sympathizers are identical, and the size of the population is normalized to one, so that the contribution of representative agent i is also the total (i e bt = b) The focus, then, is on the main goal of the present analysis, namely, to show how different types of tereonst groups arise

3 1 Popular support for the terrorists

The following proposition establishes the representative agent's voluntary supply of time in favor of the terrorist group

Proposition 1. (i) The agent supplies a positive amount of time in favor of the group's cause if

(8 + s)<a(s)u'(Q) (6)

(u) She becomes a full time supporter if

(g + s)<a(s)u'(L) (7)

Proof: This is obtained by solving the agent's utility maximization problem, using the Kuhn and Tucker theorem and the complementary slackness conditions D

Proposition 1 points out the fundamental role of government's development policy as a determinant of people's voluntary support to terrorism The role of poor economic opportu nities in people's decisions to join terrorists or rebellions has been emphasized in previous literature In a model where people choose between terrorism and an economic activity,

Bueno de Mesquita (2005) shows that economic downturns increase mobilization in favor of the terrorists Finally, in the civil war literature, Azam and Mesnard (2003) argue that potential rebels are more likely to go to war the lower the probability of benefiting from future government spending Since the present analysis assumes that potential supporters are rational, they compare the utility gain from participation with the opportunity cost, namely, the forgone income due to time spent away from their economic activity The latter cost is increasing in g, since development programs such as health, education or improved local infrastructure, improve people's marginal productivity at work (and hence their income), thereby making it more costly to spend time contributing to terrorism

The next proposition characterizes the interior solution to the agent's utility maximization problem, that is, where 0 < bl < L

A Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

Public Choice (2007) 131:177-195 185

Fig. 1 Agent's voluntary supply of time

Increase in s

g+s

Proposition 2. At an interior solution, the volunteer supply of terrorism by agent i is given by be such that:

(g + s) = a(s)u'(be) (8)

The right hand side is the marginal utility of participation. It is increasing in s, since the tenorists' charities allow them to advertise their cause, thereby raising the value potential supporters attribute to contributing to the cause. The marginal cost of participation is re flected on the left hand side by the marginal productivity of agent i, that is, the marginal increase in income due to an extra unit of time spent in her economic activity. As al ready mentioned above, government development programs increase this opportunity cost of tenorism.

Interestingly, the tenorist's charitable investments, s, increase both sides of Equation (8). This gives rise to a fundamental trade off for the terrorist group that needs volunteers for its attacks: on the one hand, more charity services make it more efficient at convincing people that its fight aims to improve their fate, which raises a(s), the value agents attach to their participation in tenorism; on the other hand, a higher value of s makes people's economic activity more productive, which increases the opportunity cost of volunteering for the group. Figure 1 represents the interior solution defined by (8).

The diagram makes it clear that an increase in s has an ambiguous effect on the voluntary supply of time be by the representative agent. On the one hand, it makes the a(s) u'(b) curve shift upward, which is the advertising effect; but on the other hand, it raises the line g + s,

which is the opportunity cost effect. Depending on the relative size of these two shifts, be may either increase or decrease.

This ambiguous effect is not limited to terrorism and can be applied to many insurgent and revolutionary movements. As mentioned in the introduction, Fenero (2004) models a political cooperative that, in addition to engaging in revolution, produces a commercial good. The author points out that the commercial sector may represent a "temptation away from revolution" if it becomes profitable enough.

Formally, the impact of s on be is obtained by totally differentiating (8) with respect to s:

be'(s) [1 -a'(s)u\be)]

[a(s)u"(fr')] (9)

ti Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

186 Public Choice (2007) 131 177-195

Therefore, the agent's voluntary contribution will rise with the terrorist's investment in de velopment programs if a'is)u\bc) > 1 In words, this is the case if the advertising effect, that is, the increase in the marginal utility of bc following an increase in s, dominates the opportunity cost effect, that is, the fact that an extra unit of s increases the marginal return to time in the economic activity by one Note that if ais) were assumed linear, this would correspond to a constant marginal advertising effect of charity expenditures However, the concavity of ais) would assume a decreasing marginal advertising effect of s In this case, the sign of (9) would be more likely to be positive for low values of s On the other hand, the advertising effect is more likely to dominate for high values of s ais) is assumed to be convex

Next, the terrorist group's investment decision is analyzed

3 2 The teironsts' investment in chanties

In this section, for simplicity, the cost functions of the group's terrorist and charitable activities

are assumed to be linear and the cost interactions between activities are ignored, that is CAiT,s) = T and CN(s, T) - s In addition, in this section as well as in the rest of the paper, I focus on the case where bc"(s) < 0, that is, the marginal impact of chanties s on popular support for the tenonst group is decreasing Note that this marginal impact, be\s), can be either positive or negative depending on whether the advertising effect dominates the opportunity cost effect

The terrorist group chooses its investments s and T in order to maximize its utility, taking into account the representative agent's reaction function beis) The following proposition states conditions under which the terrorist's investments in chanties and attacks are positive

Proposition 3. (i) The group invests a positive amount in charities if

6Tv'(g) + coiK)x/fbe,iO) - 1 > 0 (10)

(u) It invests a positive amount in attacks if

oj(K)B\0)- 1 >0 (11)

Proof: By solving the utility maximization problem of the tenonst group, the discussion of the complementary slackness conditions leads to (10) and (11) D

Since oj'iK) < 0, condition (11) defines a threshold Kmm for K, the level of govern ment crackdowns, above which the group does not invest anything directly in attacks In addition, since v"i) < 0, condition (10) characterizes a threshold level of government de velopment programs, gmm, below which the terrorists become active in the charity/NGO sector

Using the previous result, Proposition 4 summarizes how various types of terrorist groups arise, depending on government policies K and g 4

4 In his model of a two-product political cooperative, Ferrero (2004) studies the group's optimal allocation of effort between revolution and reform He argues that the relative size of the group's two wings depends on trust and credibility constraints vis- -vis its customers and workers, as well as on a liquidity constraint on its operations

ti Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

Public Choice (2007) 131 177-195 187

Fig. 2 Different types of 'terrorists

K

Proposition 4. (i) If K < Kmm and g > gmm, the group is a "pure " terrorist organization, in the sense that they invest in attacks but not in charities

(u) If K < Kmin and g < gmin, the group is a "hybrid" organization that invests both in attacks and charities

(in) If K > Kmin and g < gmm, the group is directly active only in the charity sector, and attacks result solely from the voluntary supply of time by its supporters

(iv) If K > Kmm and g > gmin, the group is inactive

Figure 2 presents this typology of gioups in a more visual way Note that gmin is a function of K, but may be increasing or decreasing This depends on the sign of bc'(0), which in turns depends on whether the advertising effect of the tenonst chanties dominates their oppoitunity cost effect, as shown in the previous section For example, if the former dominates,s then, (10) and (11) give rise to the case depicted on Figure 2 Some simple comparative statics are woith examining First, an inciease in 9T, the valu

ation of the group for public goods (i e its "altruism"), shifts the gmm curve upward, thereby enlarging the range of parameter values for which the terrorists are active in the NGO sectoi This is perfectly intuitive and similar to many models of voluntary contributions to public goods

Clearly, case (iv) may represent the goal of government policy m this simplified frame work It suggests that an appropriate balance of social programs and counterterronsm is necessary to achieve this situation The role of social programs is not only to improve eco nomic conditions, but also to reduce the opportunity for tenonsts to advertise their goals through chanties Given the recent empirical evidence that poor people are not moie likely to become tenonsts (Krueger & Maleckova, 2003), the prevention of such advertising is probably where government's social expenditure can play a role

Furthermore, one may argue that case (m) can be another goal of government policy Indeed, a pure charity poses no security threat and even contributes to public good provision Note that if both K and g are costly foi the government, investing K = Kmin and g = 0 is a cost-minimizing strategy to eliminate direct attacks T, and results on the group being a puie charity However, this does not eliminate the advertising effect of charity and the government may dislike this pure charity because of its political values In particular, if b' > 0, the charity's activity s leads to an increase in popular support for the group's political values, which might translate into the voluntary supply of terrorist acts by the population Finally, even if this does not result in such extreme behavior, it may still have a impact on future

Pure charity

5 One could graph the case where the opportunity cost effect dominates, and then gmin would be a decreasing function of K However, this would not change the intuition for the following result The relevance of the distinction between the two cases will become clear shortly

ti Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

188 Public Choice (2007) 131 177-195

voting choices, another influence that the government may not like (though this is outside the scope of the present papei)

Several well known extremist organizations fit among the first three categones Fnst, a group such as the Basque sepaiatist organization ETA belongs to type (i) Founded in 1959 by a group of student activists, ETA's goal is the independence of the Basque country from France and Spain The group is well known for its assassinations and bombing in both countnes Although it does have an official political blanch, it does not mn any noticeable charitable wing, and the group's base of supporters is believed to be small (no moie than several hundred) 6 The Corsican Aimata Corsa, as well as the Front de Libeiation Nationale de la Corse, aie among other examples that fit in case (i) Such gioups, opeiatmg in developed countries with relatively good access to public goods, were indeed unlikely to have much opportunity to seek suppoit by developing chanties Of course, their political wings ti y to promote then nationalist ideals, but they are far less convincing in such contexts than organizations of type (ii), which benefit from a stiong social base, thanks to their active involvement in chantable woik

Three examples, which have already been intioduced in Section 2, fit paiticularly well in the "hybrid" second categoiy The first is the Palestinian Hamas, whose ecent electoial victory shows how strong a popular base it has established over the years, notably thanks to its very efficient netwoik of local NGOs throughout the West Bank The second is the Lebanese Hezbollah, whose network of Islamic NGOs is well known by its supporters as well as by its enemies As reported in its profile by the Institute for Countei Tenorism, "the puipose of the aid was to gain the support of the oiganization's activities" 7 Recent studies (Ghandour, 2002, Fawwaz, 2004) show that it has maintained a high level of popular support Finally, the Liberation Tigers of Tamil Eelam also have connections with the most important Tamil NGOs Although they are separate bodies, The Libeiation Tigers' web page provides a link to that of the Tamil Rehabilitation Organization, and relayed the appeal for donations aftei the tsunami in 2004

These three prosperous hybrid organizations can take advantage of the pool economic opportunities available to a laige part of their espective local populations Of course, in the same countries, terrorist groups claiming to represent the same ideals are only involved in violent activities, and therefore belong to type (i) The above formal analysis suggests that such gioups are characterized by a low valuation foi public goods (0t in the model) so that, even in the presence of poor economic conditions, they are less likely to entei the local NGO sector to gam populai support For example, this may explain the diffeience between the two main tenonst groups in the Palestinian territories, the Hamas and the Islamic Jihad The formel, as seen above, has a strong commitment in social welfaie programs, while the lattei seems to have focused on violence Thus, the second group may be characterized by a veiy low level of the exogenous paiameter 0T Fuithermore, Ghandour (2002) aigu s that the

Algenan Islamic Salvation Front lost a laige pait of its populai base when it decieased its involvement in chanties and specialized in violence According to this example, 0T could be thought of as the outcome of changes in the gioup's stiategy designed by its core membeis

Type (in) is probably the most surpnsing Although the label "tenonst" has been used in the papei so far, it is impoitant to point out that this categoiy coriesponds to non violent activist gioups who are then clearly not tenonsts This leflects the case of gioups which, despite then sharing some of the goals of violent oigamzations, have decided to mobilize suppoit at the

6 The Institute for Counter Terrorism, http //www ict org ii/

7 Ibid

ti Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

Public Choice (2007) 131 177-195 189

grassroots level without promoting violence The Egyptian Muslim Biotherhood is a good example of such a strategy Founded in 1928, it remains the most popular opposition group in Egypt, and is committed to a "non violent, reformist appioach to slamism" 8 Despite having been banned several times,9 it operates many chanties and has a strong support base among the poor and most industries' trade unions (Ghandour, 2002) But this non violent approach to the promotion of radical religious ideas, combined with its widespread presence in Egyptian society, have made it subject to many legal restrictions on its activities In the

words of the present formal model, it faces a high level of K, while operating in segments of the population that face poor economic conditions These characteristics, as well as a high

willingness to contnbute to public goods, help to explain why it falls in the "puie chanty" category

Finally, it is interesting to comment on the inteiior solution, that is, on the investments in attacks and chanties made by a hybrid organization In the special case of linear cost functions used in the piesent section, the first order conditions are given by

0Tv'(g + s) + irba(s)co(K) = 1 (12)

oj(K)B\T) = 1 (13)

The effect of policy parameters K and g on the group's charitable investments s, can be investigated by totally differentiating (12) with espect to K and g, espectively Then, by rearranging both sides of the esulting equations, one can obtain the following

ds [\lfbu(s)a)'(K)} - = ---y-L-J-Z - (14) dK [-6Tv"(g + s)-ifb "(s)co(K)] v

ds = l-0Tv"(g + s)] dg [6Tv"(g + s) + iffr"(s)o(K)]

Then, since v"(g + s) < 0, co'(K) < 0 and ba'(s) < 0, the sign of (14) depends on that of bc'(s), while (15) is always negative

In the case where the adveitising effect of chanties dominates the opportunity cost effect (i e when b(/(s) > 0), for a given level of g, (14) shows that an increase in government crackdowns K leads to a decrease in s, which in turn decreases the voluntary supply of time to tenorism

If the opportunity cost effect of chanties dominates, more government crackdowns make the tenonsts increase their investment in chanties, but since in this case, bc'(s) < 0, this also leads to a decrease in the voluntary supply of tenorism In addition, egardless of the sign of bif(s), since B"() < 0, an increase in government

crackdowns reduces direct attacks F

Fuitheimoie, since (15) has a negative sign, an inciease in g i educes the gioup's chai liable investments s, which in tuin leads to a deciease in populai suppoit if bc '(s) > 0, or an increase in popular suppoit if bc'(s) < 0 Therefoie, the negative effect of crackdowns on the overall impact of the group can be reinforced by an accompanying increase in g if bc'(s) > 0, or a decrease in g if bc'(s) < 0 Note that, in practice, it is difficult foi the government to know the sign of bc/(s), that is, to know whether the advertising effect of chanties s dominates

8 Article in Wikipedia, available at http //en wikipedia org/wiki/Muslim_Brotherhood 9 Ibid

A Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

190 Public Choice (2007) 131 177-195

their opportunity cost effect If the government lacks this infoimation, the manipulation of g may undermine counter terrorism efforts Foi example, when bc\s) > 0, the efficiency of an increase m crackdowns K would be reduced if, at the same time, the government provided less social programs g

4 The interaction between terrorists and local charities

4 1 NGO independent of tenonsts

So fai, the analysis has consideied that the adveitismg effect comes from the tenonst group's investment in chanties Howevei, in many cases, local NGOs are likely to shaie the values of violent organizations without necessarily being related to them Foi example, the Palestinian NGO netwoik, an umbrella oigamzation compnsing 92 local NGO members, states the following as pait of its mission statement "conti lbute to the national esistance to end occupation" and "advocate for the lights of the Palestinian people locally, legionally, and globally" 1()

This suggests that local tenonst gioups, even when they do not invest in chanties them selves, benefit from some level of nee adveiusing foi then cause, which, following the pievious section, can mci ease populai suppoit Meanwhile, tenonst opeiations are likely to impact local (non tenonst) chanties

This section examines the intei action between a pui ely charitable oigamzation and a tenonst gioup, when both actois share the same political values but make then decisions non coopeiatively Assume that a local NGO, indexed by N, chooses its chantable investment s to maximize the following

0Tvig + s)-CNis,T) (16)

Following the arguments presented at the beginning of Section 3, the NGO is impacted by tenonst activities thiough its cost function Then, the existence of tenonst activity mci eases the cost of charity operations, so that C^T(s, T) > 0 Indeed, as aigued above, chanties whose ideology resembles that of violent groups are moie likely to face investigations, which may esult in restrictions on their activities and theil fund raising ability It is impoitant to point

out that all local chanties are likely to face such restnctions, whether or not they are related to tenonst gioups, since the link between tenonsts and their chanties is very often hard to establish Furthermoie, recall that CNis, T) is assumed increasing and convex in s

Then, assume that a local tenonst gioup, indexed by A, invests 7 to maximize the following

coiK)(BiT) + irb(s)) - CAiT, s) (17)

Note that (1), the utility of the teironst-chanty group descnbed in Section 3, is simply the sum of (16) and (17) That is, the group analyzed in the pievious section can be interpieted as an integrated organization comprising both an armed and a chantable wing Now, while they shaie the same political aspirations for their people, the tenonst group and the NGO aie considered sepaiate, independent actors

1 http //www pngo net/pngo htm

ti Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

Public Choice (2007) 131:177-195 191

Again, as argued at Section 3, the advertising effect induced by the NGO's activity facil itates the organization of attacks, so that CAS(F, s) < 0. In addition, recall that CA(F, s) is increasing and convex in T.

Finally, assume that both actors make their investment decisions simultaneously. Then the Nash equilibrium levels of s and F is analyzed.

The NGO's best response to any level of F is given by the first order condition:

eTv'(g + s) = C?(s,F)

The terrorist's best response to any level of s is given by:

co(K)B'(F) = CA(F,s)

(l8)

(19)

Lemma 1. Fhe NGO's best response is decreasing in F, and the terrorist's best response is increasing in s.

Proof: Total differentiation of (l8) with respect to F yields:

(C?T(s, F)) s'N(T) ?

(6Tv"(g + s)-C?s(s,F))

Since v(.) is concave, C^s(s, F) > 0, and C^T(s, F) > 0, (20) has a negative sign. Similarly, totally differentiating (19) with respect to s yields:

K(s) (CA(F,s)) (o(K)B"(F)-CAT(F,s))

Since B(.) is concave, CAT(F, s) > 0, and CAS(F, s) < 0, (19) has a positive sign.

(20)

(21)

D

Note that Lemma 1 also holds in the special case of constant marginal costs investigated at Section 3.2, i.e. with C% = 0 and C$T = 0. The Nash equilibrium is represented on Figure 3. The next section investigates potential effects of international aid received by the local

NGO on the Nash equilibrium levels of tenorism and charity.

Fig. 3 Nash equilibrium levels of NGO and terrorist activity

ti Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

192 Public Choice (2007) 131 177-195

4 2 Potential effects of international aid

For the sake of simplicity, international aid is modelled as a downward shift of the NGO's cost function Consequently, when the NGO receives aid, its best response function, depicted on Figure 3, shifts upward

International aid becomes a hot topic when discussed in contexts where the interactions between tenonst groups and local NGOs are hard to assess For example, since the end of 2002, the United States Agency for International Development (US AID) "requires Palestinian NGOs receiving US aid to sign a pledge that they do not 'provide material support or resources to any individual or entity that advocates, plans, sponsors, engages in, or has engaged in tenonst activity' "11 Members of the Palestinian NGO network have refused such a procedure, arguing that this is too broad a definition of support to terrorism

The present model assumes that the local NGO's activity implies some degree of adver tising for goals shared by the tenonst group, and this may or may not translate to increasing voluntary involvement in tenonst activities, independently of the NGO's will Therefore, the advertising effect is assumed to be driven by the convergence of both actors' political views, without any intention of the NGO to make people become terrorists

As explained above, here, international aid to the local NGO is studied as a comparative statics exercise using Figure 3 When the NGO receives aid, its best response curve shifts upward so that, at any level of T, the NGO increases its investment in charity All things being equal, this raises the equihbnum values of both s and T This undesirable positive effect on attacks T can be mitigated by an increase in military assistance in order to lower coiK), thereby shifting the tenonsts' best response function downward

Furthermore, using the results of the previous section, an increase in s may increase or decrease popular support (/?) for tenonsm If the advertising effect dominates (i e if be\s) > 0), then b increases, but it decreases if the opportunity cost dominates Then, note that aid leads to an increase in the overall impact of the group (i e direct attacks T plus popular support bis)) unambiguously only if be\s) > 0 In the case where bc'is) < 0, the increase in T may be outweighed by the decrease in b

Distributing aid directly to the government in return for increased social expenditures g would increase the opportunity cost of supporting tenonsm However, since local NGOs often have more efficient ways to reach the population, they cannot be ignored by aid disbursements

Finally, one may argue that aid can be used for "counter-advertising" purposes The example of American aid through the World Food Program mentionned in Section 2 certainly fits with this principle In the present model, a simple way to capture such effects is to make the value of one's contribution to tenonsm a decreasing function of the aid received, that is, uib, a) with duib, a)/da < 0 Such aid would make participation less attractive The question of whether this kind of policy is likely to work lies outside the scope of this paper, since a proper treatment of the issue would have to deal with other factors, such as the role political neutrality plays in international NGOs' activities

4 3 Back to an integrated tenonst-chanty organization

Sections 4 1 and 4 2 have investigated the case where the terrorist group and the NGO are independent bodies, but that affect each other's activity through their respective cost functions

The key assumptions were that tenonst activity increases the cost of NGO operations, while

1 ' Ibid

ti Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

Public Choice (2007) 131 177-195 193

charity work by the NGO has an advertising effect that may benefit the terrorist group by increasing its support base but also by reducing its costs The idea behind the latter effect is that the influence of an NGO which shares the values and goals that terrorists claim to fight for creates a more favorable environment for the group to organize their attacks

The present section goes back to the case of an integrated terrorist-chanty organization, as in Section 3, in order to compare the investments F and s of the integrated structure with those of the independent actors described in Sections 4 1 and 4 2 Therefore, the following is a kind of welfare analysis which aims to derive the levels of charity and terrorism that

maximize the joint surplus of the terrorist and the NGO, and to compare them with their Nash equilibrium levels given by (l8) and (19)

The integrated tenonst-chanty chooses s and F to maximize its objective function, given by Equation (1) In the following analysis, it will be useful to keep in mind the following properties of the functions entering the organization's objective, stated at the beginning of Section 3 First, recall that CN(s, F) is increasing and convex in s, and CA(F, s) is increasing and convex in F Moreover, Cj(s, F) > 0, CA(F, s) < 0, which is in line with the intuition developed previously, namely, the fact that tenonst attacks increase the cost of NGO activities

and charity work reduces the cost of terrorist operations In addition, CjT(s, F) > 0 and CA(F, s) > 0 further ensure the convexity of the cost functions Finally, recall that v( ) and B( ) are increasing and concave, that b'(s) may be either positive or negative depending on whether the advertising effect dominates the opportunity cost effect, and that b"( ) < 0

Then, focusing on the interior solution, the first order conditions for s and F axe given by

QTv'(g + s) = C?(s, F) + CA(F, s) - co(K)x/fb\s) (22)

oj(K)B'(F) = C?(s, F) + CA(F, s) (23)

Comparing (22) with (l8), since CA(F,s) < 0, the compansion between s in the integrated and disintegrated cases depends on the sign of b'(s) If the advertising effect dominates, b'(s) > 0, and then, more charity is provided by the integrated organization than by the NGO alone Indeed, the fact that b is increasing in s is a positive externality of NGO activity on the tenonst objective A separate NGO entity only provides charity based on its altruistic

motivation, but does not internalize its impact on popular support for the cause However, if the advertising effect is dominated by the opportunity cost effect, more charity

leads to a decrease in active popular support b (i e b'(s) < 0) This negative effect of charity may or may not offset the fact that this activity also lowers the cost of any given level of attacks, that is, the fact that CA(F, s) < 0 In this ambiguous case, the integrated organization provides more charity compared to the Nash equilibrium if the following condition is verified

CA(F, s) - co(K)irb'(s) < 0 (24)

Indeed, if it holds, the benefit in terms of a reduction in the cost of terrorist attacks outweighs the decrease in popular support This is more likely if \// is small, that is, if the group cares little about popular support

What about the level of terrorist attacks F^ Comparing (23) and ( 19), given that Cj (s,F) > 0 and B is concave, the integrated group always performs fewer attacks This is because terrorist attacks increase the cost of NGO activities, and this adverse effect is internalized

when both components (charity and terrorist) are integrated Thus, the above analysis suggests that when local NGOs clearly state their political values,

the perpetration of tenonst acts by independent but like-minded violent groups may affect

ti Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

194 Public Choice (2007) 131 177-195

their activities negatively The model has not considered the role of contributions by politically neutral local NGOs in such contexts, and this is an interesting avenue for future research

5 Conclusion

Violent groups, such as terrorist and rebel organizations, sometimes invest significant re sources in social work, notably in the form of chanties and NGOs For example, the Pales tinian Hamas is reported to devote more than 95 percent of its budget to social welfare programs Recent papers have investigated the case of extremist groups that contribute to lo cal public goods Berman (2003) and Berman and Laitin (2005) model these organizations as clubs which provide social welfare only to their members Their analysis helps to understand how these groups manage to enforce apparently inationnal levels of sacrifice among their members The present paper is a useful complement to this emerging literature By modelling a terrorist group's chanties as an advertising device, the analysis explains how different types of organizations arise in equilibrium, depending on exogenous levels of government policies Some may specialize in tenonsm, others may be of a "hybrid" type, mixing tenonsm and charity Some do not even engage in tenonsm, and focus entirely on public good provision Then, it is acknowledged that in many cases, local NGOs share the political values of violent organizations without necessarily being related to them This allows a fully tenonst group to benefit from some free advertising for its cause It is assumed that terrorist activity increases the cost of NGO operations, while charity work by the NGO has an advertising effect that

may benefit the terrorist group by increasing its support base but also by reducing its costs The Nash equilibrium levels of charity and tenonsm chosen by the separate NGO and

tenonst entities are compared with those invested by an integrated tenonst-chanty organi zation Since the integrated terrorist-chanty internalizes the additional benefits of charity in terms of popular support and reduced cost of attacks, it may have more NGO activity than an independent NGO

Furthermore, the Nash equilibrium level of tenonst attacks by an independent terrorist group is always higher than that of an integrated tenonst-chanty because the integrated organization internalizes the negative impact of terrorist activity on NGO operations This result suggests that recent requirements imposed by the US and the EU, that local NGOs commit to be unrelated to terrorist groups, might result in more tenonsm rather than less

Acknowledgements I thank Jean Paul Azam and Cynthia Howson for helpful comments and discussions I am especially grateful to two anonymous referees whose extensive comments have helped to improve the paper significantly

References

Azam, J -P (2006) On thugs and heroes Why warlords victimize their own civilians7 Economics and Gov ernance 7(1), 53-73

Azam, J -P (2005a) Suicide bombing as inter-generational investment Public Choice, 122(1), 177-198 Azam, J -P (2005b) How to curb "high quality" terrorism7 A case for boosting aid, working paper, University

of Toulouse, mimeo Azam, J -P, & Mesnard, A (2002) Civil war and the social contract Public Choice, 105(3), 455-475 Barro, R (1990) Government spending in a simple model of endogenous growth Journal of Political Economy,

98(5), S103-S126 Becker, G S ,& Murphy, K M (1993) A simple model of advertising Quaterly Journal of Economics, 108(4),

941-964

ti Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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

Public Choice (2007) 131 177-195 195

Berman, E (2003) "Hamas, Taliban and the Jewish underground An economist's view of radical religious militias", NBER WP 10004

Berman, E , & Laitin, D D (2005) Hard targets Theory and evidence on suicide attacks NBER WP 11740 Besley, T, & Gathak, M (2001) Government versus private ownership of public goods The Quarterly Journal

of Economics, 116(4), 1343-1372 Bueno de Mesquita, E (2005) The quality of terror American Journal of Political Science, 49(3), 515-530 Bueno de Mesquita, E , & Dickinson, E S (2005) The propaganda of the deed Terrorism, counterterronsm

and mobilization, working paper, Department of Political Science, Washington University Fana, J R & Ace, M D G (2005) Terror support and recruitment American Economic Association Annual

Meeting Papers, available at http //www aeaweb org/annual_mtg_papers/2005papers html Fawwaz, M (2004) Action et id ologie dans les services ONG Islamiques dans la Banlieue Sud de Beyrouth

In S BenNe ssa, N Abd al-Fattah, S Hanafi & M Carlos (Eds), ONG et Gomernance dans le Monde Arabe, Cedej (Le Caire) et Karthala (Pans)

Ferrero, M (2004) Revolution or reform7 Socialism's dilemma as a rational choice problem Homo Oeco nomicus 27(2), 251-282

Ghandour, A R (2002) Jihad Humanitaire Enqu te sur les ONG Islamiques Flammarion (Pans) Glazer, A , & Konrad, K A (1996) A signaling explanation for charity American Economic Review 86(4)

1019-1028 Guevara, E C (2001) Textes Militaires La d couverte (Pans) Hopkins, R , & Scott CD (1999) The economics of non-governmental organizations The Development

Economics Discussion Paper Series The Suntory Center, London School of Economics Jack, W (2001) Public policy towards non governmental organizations in developing countries, woiking

paper, Department of Economics, Georgetown University Krueger, A B , & Maleckova, J (2003) Education, poverty, political violence, and terrorism Is there a causal

connection7 Journal of Economic Pet spec ta es 17(4), 119-144 Levitt, M A (2003) Hamas blood money Mixing good works and terror is no formula for peace, Available

at http //www ict org ii

Levitt, M A (2002) The Network of Terrorist Financing, available at http //www ict org ii Rose-Ackerman, S (1998) Altruism, nonprofits, and economic theory Journal of Economic Literatme 34,

701-728

Springer

This content downloaded from �������������128.111.180.8 on Wed, 15 Dec 2021 23:40:33 UTC�������������

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