The Roots of Terrorism: Political Freedom and Other Determinants

The Roots of Terrorism: Political Freedom and Other Determinants

Drew Calcagno

Introduction

After 9/11, the West has placed great importance on ‘figuring out’ the roots of terrorism.  The American line of thinking tends towards the search for significant determinants, believing that addressing those determinants in some way might prevent terrorism or at least limit the growth of the tactic.  Since the influx of terrorism research began in earnest about a decade ago, there has been a debate between economists and political scientists on whether the roots are inherently economic or political.  International political economists and scholars concerned with spectra of violence find the debate to be even more nuanced, albeit muddled due to the ambiguity of defining and codifying terrorism data. 

While terrorism worldwide cannot be causally determined by any prescriptive set of variables, there are worthwhile strides to be made in the field of statistical analysis of terrorism.  Understanding which political-economic environments are most ripe for terrorism provides potentially predictive insight for those governments and institutions working diligently to stop it. 

Holistic views of terrorism demonstrate a tapestry of responses of ‘why’ terrorism occurs, and with such frequency.  Yet, the economics field has not yet shown a viable statistically-significant model to correlate any slew of economic variables with terrorism.  By contrast, the political science field has emerged with a working model.  Alberto Abadie shows that the level of political freedom correlates with the existence of terrorism.  Addressing the political freedom variable is no panacea for anti-terror initiatives, but it is one of the first of its kind in terms of significance, both statistically and holistically. 

This piece recreates Abadie’s model.  The recreation was necessary to establish an update on his decade-old study with new interpretations of the variables and new sources for panel data of 2003/2013.  This piece shows that Abadie’s initial assertion – transitional states with political freedom in flux see the most terrorism – remains true at a statistically significant level a decade after his study was created.  Nevertheless, while his model considers states with very low political freedom (dictatorial regimes) experience actually less terrorism than a transitional state, this piece finds that relationship to be less pronounced. Calcagno’s recreation predicts a less frequent, but still very present overall, existence of terrorism in states with very little political freedom.  This phenomenon may be due to social technology as a catalyst for terror in dictatorial regimes today where it did not exist as readily when Abadie created his study.  

Studies of Terrorism vs. Studies of Civil War

For the preponderance of the analysis, this piece will consider how political freedom interacts with terrorism, not the economic variables with which some economists erred to show as convincing determinants. While many terror-ridden countries happen to poor, the poverty was not the determinant.  One terrorism scholar, Krueger, suggests,  “victims of terrorist attacks…are much more likely to come from wealthy countries than poor countries.”[i]  Nevertheless, Abadie considers this assertion to be nearsighted; “Terrorist groups may decide to attack property or nationals of rich countries in order to gain international publicity.  As a result, transnational terrorism may predominantly affect rich countries.  The same is not necessarily true for domestic terrorism” (emphasis mine).[ii]

Krueger and Abadie agree that civil liberties and political freedom are very similar political variables.  “Political rights reflect the presence of democratic practices, such as open elections, while civil liberties reflect related rights, such as freedom of association and freedom of the press.”[iii] Krueger chooses to use civil liberties as the political proxy, whereas Abadie chooses political freedom.  Abadie discusses his thought process as follows;  “the reason is that endogeneity may be a more serious concern for the latter, if countries restrict civil liberties in response to terrorism.”[iv]  Nevertheless, Abadie goes on to explain that despite their high colinearity, the empirical regression results are the same with either proxy.[v]  To clarify, the Abadie model considers domestic terrorism as well as transnational, increasing the dataset nearly sevenfold.  In his study, the independent variable is the lack of political rights (political freedom).  The dependent variable is the risk of terrorist activity (terrorism).  In this piece’s recreation, the independent and dependent variables remain the same, although they are derived from different sources which will be expounded upon later.

In an attempt to show that economics could explain terrorism’s roots rather than a discussion of civil liberties or political freedom, one economist, Piazza, used the same transnational terrorism dataset from the State Department as did Krueger.  Including terrorism casualties as well as attack frequency, he found that economic factors (poverty, unemployment, and economic growth) had no significant effect on terrorism.[vi]

Krueger and Abadie’s models are canonical for the discussion of terrorism’s roots, but should not be confused with the literature on civil war’s roots.  While terrorism is the use of violence and intimidation in the pursuit of political aims, civil war is differentiated as a conflict involving at least 1000 battle-related deaths, typically between state-led forces and non-state led forces.  Collier and Hoeffler explain that economic variables explain the onset of civil war, showing that poorer countries are more likely to experience it than richer ones.[vii]  They cite that political variables – inequality, political rights, ethnic polarization, religious fractionalization – all prove to be insignificant.[viii]  However, critics point to a potential endogeneity problem in their calculations.[ix] Several take issue to the political convenience of Collier and Hoeffler’s oversimplification[x] as well as its lack of historical nuance or consideration of individual societies.[xi]

Fearon and Laitin explain that similar factors are to blame for civil war, citing poverty and slow growth, rough terrain, and large populations as a few of the variables.[xii]  “Research that uses similar kinds of statistical models to predict the incidence of civil war in a country has usually found that lower-income countries are more likely to experience a civil war.  Because of these contrasting results, it is important to distinguish between terrorism and civil war.”[xiii]  Rather than considering spectra of violence in civil war, this piece keys in on the spectra of violence in acts of terrorism. 

In Collier and Hoeffler’s defense, some scholars suggest that the best analysis of violence is economic, and when other social scientists’ studies make the most sense and “do a good job,” they are thinking economically.[xiv]  Cramer clarifies Hirshliefer’s economics by adding, “Actors [rationally] choose conflict where this is more profitable at the margin than exchange.  Others argue, similarly, that if the payoff to conflict outweighs the calculated risk then more time will be allocated to war.”[xv] Surely economics has its role in the rationality of violence, but perhaps it is not as comprehensive as Collier – or even Hirshliefer – suggest. 

In Ethnicity, Insurgency, and Civil War, Fearon and Laitin do not simplify the onset of civil war to merely economic causes, but they agree with Collier and Hoeffler that poverty plays a significant role in civil war violence.[xvi]  They consider instability as their lone political driver, but rely on structural determinants like rough terrain and a large population rather than the conventional suggestion that ethnic or religious differences drive intra-state conflict.[xvii]  Their study lends an example of how the introduction of political factors gets closer to a more culturally-aware explanation of violence. 

The idea of spatiality in development theory suggests that we ought to think about the microeconomic decision making of homo economicus, as well as holistic decision making – violence as a societal construct – as distinct facets of the topic at hand.[xviii]  In the debate of what is and what is not terrorism, spatiality – or the levels of the actors – should be considered. Contrary to the neo-classical economics that dominate the civil war debate, political economy presents a more comprehensive examination of terrorism’s determinants.  This piece will now transition to a discussion of terrorism and several recent studies that highlight this political economy approach using historically nuanced variables separate from economic factors. 

Terrorism and Political Economy’s Contribution

While Collier & Hoeffler and Fearon & Laitin look at civil war, Abadie and Krueger look at terrorism and incorporate more political economic and societal context. Abadie suggests that political-economic factors may even render purely economic variables insignificant.[xix] Yet, this amalgam of political and economic drivers show that descriptive variables can still be applicable and correlational, albeit more complex than Collier’s thesis of economic simplicity.

Abadie finds that economic drivers are insignificant and political ones are significant – specifically the measurement of political freedom.  He regresses economic variables (per capita GDP, GINI, HDI), political variables (lack of political freedom), ethnic variables (ethno-linguistic fractionalization scores), and geographical variables (size, elevation, and type of terrain) to find that in correlational as well as robust, causal models that only the lack of political freedom rang true as a significant descriptor of terrorism.  He finds that terrorism reaches its peak in a transitional state. The presence of political freedoms, typically found in democracies, are associated with relatively low levels of terrorism-related death and destruction.[xx]

The extreme lack of political freedom, typically found in dictatorships, provokes only marginally higher instances of terrorism. Transitional states, where the essential question – who gets political power and on what basis – is disputed, are the sites of the most frequent terrorism with the greatest death toll and effects of destruction.[xxi]Abadie’s study implies that terrorists act out of grievance for political freedom in contrast to Collier and Hoeffler’s opposite assertion: that the economic variables – the greed – cause violence.[xxii] Nevertheless, it is important to distinguish that these two theories do not entirely contradict.  First, both sides venture to say that some sort of correlational factor can describe violence.  Second, each study considers a different type of violence on the continuum: Collier and Hoeffler look at the onset of civil war while Abadie looks at the existence of terrorist attacks.[xxiii]

Modern international political economy suggests that lack of political freedom does indeed drive terrorism.  And yet, debate lies in which regime type is most affected.  Krueger goes on to dispute the convenient non-monotonic curve that Abadie finds.  Krueger reproduced the study with IV modeling and found a flatter line, meaning that transitional states do not necessarily exhibit more instances of terrorism on the basis of suppressed political freedom. However, the statistical significance of political freedom implies that “terrorism [declines] as steps are taken towards democracy.”[xxiv]  This piece’s recreation lay somewhere in between these two poles.  My dataset is different than Abadie’s and Krueger’s and still produces a curved relationship, albeit less pronounced than Abadie’s and closer to Krueger’s flat line, and only correlational.  Yet, like Krueger’s discussion of slopes for cost and benefit curves, the pronunciation of the political freedom curve in Abadie’s model, Krueger’s recreation, and mine, is less relevant than the success of finding an empirically supported variable – political freedom.  Nevertheless, with a statistically significant driver that encompasses more than just economics, we can say that terrorism does have a variable with explanatory power.  With success in the attempt to find statistical significance in a chaotic study of terrorism on the continuum of violence, the social sciences support the idea of violence as relational with other variables and is strongest in Abadie’s model.

In addition to his recreation of Abadie’s model, Krueger created another model that suggests terrorism is relational with other variables in a robust way.  Here, Krueger considers terrorism as a market of violence.  Using different axes of analysis than Abadie, he examines marginal cost and marginal benefit – relevant terms to homo economicus – as well as frequency of terrorism, to show that terrorist groups implicitly go through the same rational microeconomic laws that any other producer goes through.[xxv]  Following neo-classical thinking, he claims that these groups have an optimal amount of terrorism: where marginal cost and benefit meet. Likewise, the slopes of these marginal cost and benefit curves are governed by the law of diminishing returns and can be accurately represented with a distinct slope for each specific terrorist group.[xxvi]  The difficulty in creating an accurate model lies in the representation of mental calculus for benefits and costs.  Are deaths more beneficial than physical damage? By showing a terrorist market equilibrium, Krueger creates a framework that satisfies homo economicus and therefore remains in the vein of correlational interactions.

Terrorism and Political Freedom: The Abadie Model and its Recreation

Any effort to model something as chaotic and nuanced as terrorism ought to be met with skepticism.  However, if one wishes to attempt the task, Abadie provides the community with a critical work in statistical analysis of terrorism.  His regression and summary statistics are available here.[xxvii]  This piece expands upon these political, fractional, and geographical variables in my model, with updated data and a few updated variables, available upon request (see author contact information).  The statistics that are used in my model captures the essence of Abadie’s model, but with a several changes to the sources of the data and to the definitions of the variables.

The proxy for terrorism is open for interpretation.  Abadie used the World Market Research Center’s Global Terrorism Index (WMRC-GTI) to establish such a proxy for the dependent variable.  The WMRC is country-specific, including 186 nations (as were available at the dataset’s creation in 2003).  Qualitatively, the WMRC-GTI data is a predictive ranking.  It includes factors of scale, presence, motivation, and prevention of terrorism.  Quantitatively, it measures low to high exposure to terrorism from 1 to 100 (with 100 representing the countries with highest exposure).

The dataset for terrorism that Abadie used is no longer available and can no longer be accessed.  The organization that compiled the composite statistic for terrorism risk lost funding in 2004 and no longer issues that statistic.  Yet, this piece was able to create an alternative.  The Institute for Economics and Peace (IEP) has created a proxy variable for terrorism based on data of terrorist incidents since 1973, complied by the University of Maryland Global Terrorism Database.  The IEP suggests weights for a composite statistic that incorporates attack frequency, deaths, injuries, and property damage.[xxviii]  This piece applies these weights to the UMD data for 2013/14 for 162 nations, available upon request.

The political variable under consideration is a measure of the absence of political freedom.  The Freedom House Political Rights Index provides a score for political freedom.  Other political scientists have considered measuring civil liberties (see Krueger and Laitin’s work from 2003),[xxix] yet Abadie has suggested using political rights instead.  “The reason is that endogeneity may be a more serious concern for [a civil liberties measure], if countries restrict civil liberties in response to terrorism.  However, given that these two variables are highly collinear, [the empirical results] do not depend on which one is used in the regressions.”[xxx] The proxy for political freedom ranges numerically from 1 to 7 (with 7 representing the countries with least political freedom).

After running Ordinary Least Squares (OLS) regressions on the economic, fractional, and geographical variables, Abadie finds that there is no significant correlation between economic determinants and terrorism (while including the fractional and geographical data as additional explanatory variables in the equation as well).  The Abadie OLS results are available here.  In regression 6, Abadie shows that economic variables do not significantly explain terrorism, but a fifth of the variation in terrorism tends to be explained by the variation in political freedom, as well as three significant geographical and climate variables.   In regressions 7 and 8, using  and  as economic variables, they prove to be insignificant.  In this piece’s 2013 model, the focus is on modeling regressions 6, 7, and 8 with my selection of variables (with the new sources, available upon request).

Instrumental Variable (IV) estimation yielded the same result of an unacceptable level of significance for economic variables in the original Abadie model.  IV estimation is used for simultaneous effects correction, meaning that economic variables and terrorism may interact.  A country with poor economic development indicators may promote conditions for terror as well as terror promoting poor economic development indicators.  Thus, IV estimation should correct for this issue of simultaneous effects or reverse causation, shown here.

Nevertheless, the IV tests also yielded that economic variables are not significant descriptors of terrorism (see columns 6 and 7 of Abadie’s IV testing results here).  On the other hand, political freedom did have a significant effect on terrorism, albeit non-linear, during IV testing.  The following graph shows this result here.[xxxi]  

The New Model

This study considers an OLS regression model (summary statistics available upon request). My new model yields the non-monotonic result Abadie demonstrated, (also available upon request).  In an effort to show the non-monotonic relationship graphically, this piece utilized the data in the same format of a second-order polynomial trend line (as in the Abadie graph), as follows:

Figure 1: Calcagno’s Model of Terrorism and Lack of Political Freedom

Despite the negative coefficient on the squared political freedom term – meaning there is indeed a non-monotonic relationship – we do not see the neat bell curve that the Abadie model shows.  Nevertheless, given the significant changes in data sources and new type of terrorism data my new model continues to suggest a very similar result to Abadie’s 2003 model. 

Consider what this graph means:  economic variables have less explanatory power than do political variables.   GDP per capita was insignificant in the previous regressions, as well as the Abadie model, and one sees that other economic variables like Gini remain insignificant.  This fact is still important, however, because one can use these proxies to strip away economic development and isolate the political variable’s explanatory (or at least correlational) power. 

As discussed before, terrorism increases with a greater lack of political freedom, reaching its peak in a transitional state, around the value of 5.   More specifically, we see that a 1-unit decrease in political freedom (for example, elections are no longer free in a country like Kenya), then terrorism tends to increase by approximately 110% - so it tends to double the amount of terrorist attacks (weighted using the IEP methodology) – by the estimation of my model.  That estimation may be a bit presumptuous, yet the statistical significance is valid at the 1% level.   Scaled to a reasonable interpretation for cultural fractionalization, if fractionalization increases by .1 units (the range is from 0 to 1), which could loosely equate to the change in pluralist vs. Islamist media attention that has created a cultural schism in Nairobi, there is a correlational 10% increase in the likelihood of terrorist activity.  There have been terrorist threat increases in the region as a result of this media attention (particularly with al­-Shabaab) so the correlational increase is definitely probable (and statistically significant at the 10% level).  It is important to note that these relationships are correlational rather than causal as the Abadie model proved.  In essence, we have achieved the objective of showing that political freedom remains the strongest variable that correlates with terrorism now, as it was a decade ago when Abadie ran his regressions. While this interpretation shows a statistically significant correlation, it cannot conclusively prove causality with OLS techniques.  Improvements could come with an updated determination of the specific weights to assign to the terrorism data rather than accepting the IEP-chosen calculations. The IEP does not change its weights with different countries, nor does it consider any effects beyond death, injury, or physical damage.

Implications: A Qualitative Connection to the Leidig Model

This non-monotonic correlation function demonstrates that the lack of political freedom is not a purely linear descriptor of terrorism.  There is a non-linear bridge between regime types, and VADM Leidig, retired AFRICOM commander, colloquially called the trough, “hell.”[xxxii]  There are many instances of terrorism, confusion, and chaos.  His illustration is shown as follows:

Figure 2: The Hellish Transition Between Democracy and Dictatorship

This of a transition corroborates Abadie’s assertion that there is an “observed increase in terrorism for countries in transition from authoritarian regimes to democracies.”[xxxiii]  This piece considers three examples of terrorism peaking during the transitional stages of political freedom in Libya, Somalia, and Kenya.  I briefly recapture what Leidig described about Libyan transition, then dips into how his model and Abadie’s may be applied to the Horn of Africa, specifically in Somalia and Kenya of the past decade.  Applying the Abadie model to these three nations, this piece shows that political freedom is a significant explanatory variable for terrorism rather than grievance from economic conditions.

The following examples are applications of the model on the African continent.  This piece will demonstrate the practical application of the Abadie/Calcagno Models and their qualitative similarity to the Leidig Model for three country case studies. 

Case Study 1: Libya

Libyan transition away from the Gaddafi dictatorship illustrates the Leidig model of hellish conflict, including terrorism during regime change. Leidig described the unpleasant consequences of a “good war,” referring to the unplanned effects of Operation Odyssey Dawn in 2011.[xxxiv]  One of the most damning results was the lack of a stable government to take over the power vacuum.  In the meantime, Libyan militias took security concerns into their own hands.  Based on the Abadie model, Libya moved from a position of least political freedom (close to a value of 7, see Figure 1) to the transitional period (toward the value of 5, see Figure 1).  The infamous Benghazi diplomatic compound attack in 2012 was one of the most influential terrorist incidents of post-Gaddafi Libya, manifesting a high point of terrorism catalyzed by a forced transition away from authoritarianism.[xxxv]  Libyan militiamen, who would coalesce as the terrorist group Ansar al-Sharia in the coming years, raided the diplomatic compound and a CIA safehouse nearby.[xxxvi]  They killed the US Ambassador to Libya, a leading Foreign Service Officer, and two CIA security contractors, a major blow to American relations with the nation and the region.  Now years later, Libya is without much US or NATO support and has become a haven for the Islamic State.  The nation’s security concerns have prevented voter turn-out and the deteriorating political order left the international community with little other option than to pull out its personnel, most notably UN removal and embassy shut-downs back in July 2014.[xxxvii]

Case Study 2: Somalia

Somalia is a case of transitional state that exemplifies a peak of terrorism (Abadie’s peak) during its trough of political efficacy (Leidig’s hell).  Over the past decade, there have been times of utter disarray when the terrorist group al-Shabaab has filled the security vacuum with intimidating attacks around the region (both inside and outside of Somalia’s borders).  But once AMISOM began to have success in collaboration with the Somali Government in 2012, there was a new sense of political efficacy.  Under a new elected president (via the first elections since 1967),[xxxviii] Hassan Sheikh Mohamud, government forces and AMISOM – mainly the Kenyan Defense Force (KDF) contingent – took back Kismayo.  Kismayo is a prolific port, second in size only to Mogadishu.  Al-Shabaab had utilized the city as a stronghold for several years, creating impromptu taxes on charcoal and smuggling high volumes of it in and out of the country for terrorism financing.  The KDF, AMISOM, and Somali National Army (SNA) pushed the terrorist forces out of the port and therefore pushed the country towards greater political freedom.[xxxix]

Liberation from al-Shabaab’s local tyranny was a major step away from the Abadie model’s peak – a value of 5 (see Figure 1) and on the path to greater political efficacy elected by the people – towards a value of 4 (see Figure 1).  Likewise, the Leidig transitional state model would indicate that Somalia in late 2012 was rising up towards democracy, out of the hellish trough.[xl]  One success led to another, and after Kismayo was taken from the terrorists, other terrorist funding sources tumbled.  Piracy dropped by 70 per cent that year[xli] and al-Shabaab was stuck trying to find other means of making money – their sources were drying up left and right.  Nevertheless, rather than these economic factors, it was once again the political variable that changed terrorist activity.

The AMISOM coalition grew complacent in their late 2012 successes and eased their grip on Mogadishu, Somalia’s capital.  Despite anemic funds, al-Shabaab focused on disruption of political freedom, using gunmen attacks and bombings in the capital.  The capital’s main court was gunned down with 30 dead, the Turkish embassy was bombed, the presidential palace was attacked, and so was a UN Development Program compound – all in the political heart of the nation.[xlii]  When civilians’ political freedom was no longer secured by AMISOM and Somali forces in 2013, it did not matter how poor the country was or how poor al-Shabaab was, the terrorist incidents grew and dropped Somalia back into the Leidig model’s hell and the Abadie model’s peak of terrorism.  Despite attempts at political reform in 2016, al-Shabaab remains.  Somalia was not the only nation hurt by their own political disorder and al-Shabaab’s growth.  There were spillover effects in Kenya as well.

Case Study 3: Kenya

Political freedom hit the transitional state trough of the Leidig model after the controversial 2007 elections in Kenya.  In the absence of dictatorship or democracy, political violence broke out in 2007/8.[xliii]  Over 1000 died near the capital and primed the landscape for terrorism.[xliv]  The nation vacillated between potential for democracy and the transitional state for two years.  Despite an attempt for political order through a new constitution and coalition government in 2010, the terrorist attacks had begun.[xlv]  Late that year, a bus was bombed in Nairobi.  Several months later in the fall of 2011, Somali militants affiliated with al-Shabaab attacked the Kenyan coastline, raiding civilian resorts and a refugee camp. [xlvi]  Thus, Kenyan transition and resultant lack of political freedom led to domestic as well as transnational terrorism – there was no discrimination in the violence.  “The international hotels here are starting to empty; the restaurants and bars are bringing their tables in off the streets…The country waits to see what - beyond the two grenade attacks this week - Kenya's military offensive inside neighboring Somalia might yet provoke by way of retaliation.”[xlvii]  In 2012, while the International Criminal Court indicted Kenyan politicians for post-election violence,[xlviii] the chaos continued and the country remained in the Leidig model’s trough, resulting with an even higher peak of terrorism in the Abadie model.  Foreshadowing an even grander attack, al-Shabaab targeted a shopping center in Nairobi, resulting in 30 casualties in May 2012.  Hardly two months later, al-Shabaab killed fifteen people in two churches close to the Somali border.  Later that year, gunmen (presumed, but unconfirmed to be al-Shabaab) attacked AMISOM forces in the same town as the churches, Garrisa.[xlix]

Paralleling the Somali situation, AMISOM and the KDF had a stint of success at the end of 2012 into 2013 (surrounding the capture of Kismayo), but their hold on political order was short lived.  Just as attacks on Mogadishu grew under the new leadership of Mohammed Abdi Godane, al-Shabaab’s emir, they grew in Nairobi.  The largest attack of the group’s existence took place in September 2013 at the Westgate Mall, where al-Shabaab systematically bombed, shot, and raided their way to 60 deaths and scores more injuries.[l] The group’s leadership retaliated against Kenya for the Kismayo capture and saw this period of Kenyan transition as the perfect time to execute – and it was.  Representative of the lack of political order for a country in transition, there was a belief in Addis Ababa’s AU Headquarters that the Kenyan gendarme that responded to the Westgate attack were incompetent.  The forces hardly prevented further al-Shabaab damage, and possibly inflicted more harm on the mall’s structure and inhabitants with their haphazard weapon fire.

Conclusion

Rather than economic variables, political variables significantly correlate with terrorism.  Political freedom is a measure of the bridge between democratic and autocratic regimes.  Per this piece’s recreation of the Abadie model, there is a non-monotonic relationship (albeit not as pronounced as in the original Abadie model) between political freedom and terrorism.  In line with the Leidig, chaos and violence (including terrorism) are maximized in transitional states, particularly after removing a dictator from power.  This peak of terrorism for countries in the midst of regime change is shown in three African case studies: Libya, Somalia, and Kenya.  In Libya, when a Gaddafi was ousted, the chaos of transition promoted terrorist acts, culminating in an attack that killed an American ambassador and others by Ansar al-Sharia giving rise to safehavens for the Islamic State.  In Somalia, al-Shabaab has grown and succeeded in attacks whenever AMISOM and the Somali government did not ensure political freedom for its people (democracy side) or utterly remove it (autocracy side), nor keep a tight grip on major cities like Kismayo and Mogadishu.  In Kenya, political instability after the 2007 elections primed an environment for al-Shabaab and others to take advantage of chaos and attack foreigners as well as native Kenyans, culminating in an attack on a mall in Nairobi in 2013 and a school in Garrisa in 2015.  Terrorism is not a tactic of war that is easily defined or modelled, yet this piece presents one way forward – reconfirming Abadie’s notion that lack of political freedom significantly relates to the existence of terrorist activity.

Endnotes