Dogs and Cats Living Together? Explaining the Crime-Terror Nexus

ABSTRACT Why do some terrorist groups cooperate with criminal organizations? This behavior is puzzling because there are reputational reasons for each of these kinds of groups to avoid the other, yet such cooperation seems to be increasingly common. The growing literature on the “crime-terror-nexus” examines terrorist-criminal cooperation, but questions remain. We discuss relevant research, and present hypotheses. Analysis of nearly 400 terrorist organizations using a newly-coded measure of inter-organizational cooperation suggests that certain types of terrorist groups are more likely to work with organized crime: those that are involved in the drug trade, cooperate with other terrorist groups, are older, or are more lethal. These relationships are robust. We also find some evidence that the crime-terror nexus is more likely for groups that have state sponsors, have larger memberships, operate in more capable countries, or operate in less-democratic countries. There is also some evidence that ethnically-motivated terrorist groups are less likely to cooperate with organized crime. Interestingly, there is little or no support for the idea that the nexus is related to territorial control or religious ideology. These findings go against some previous research and suggest steps for future analysis.


Introduction
On March 17, 1992, an explosives-filled van detonated in front of the Israeli Embassy in Buenos Aires, killing twenty-nine people.A Hezbollah faction apparently used explosives purchased from Syrian arms traffickers in the Tri-Border Area for the attack. 1 In the early 2000s, the Neapolitan Camorra supplied Basque Homeland and Freedom (ETA) with arms and false documents. 2More recently, the Calabrian 'Ndrangheta purchased weapons from Eastern Europe and transported them to Islamic-State-controlled territory in Libya. 3These examples represent only a small sample of documented cooperation between terrorist organizations and organized crime groups.
Crime-terror cooperation alarms governments around the globe.In 2019, the UN Security Council unanimously adopted Resolution 2482, expressing concern about connections between terrorism and organized crime, noting the benefits for terrorists in particular. 4Europol's 2021 terrorism situation report described cases of crime-terror cooperation throughout Europe, involving a wide variety of groups. 5hy do terrorists cooperate with organized crime groups, and what factors lead to this cooperation?As the examples above illustrate, these connections occur in diverse environments and among groups with distinct motivations.This cooperation is puzzling because criminals might want to avoid the government attention frequently brought upon terrorist organizations, while many terroristsseeking to represent a religious or ideological cause-do not want their own reputation sullied by links to common or predatory criminals. 6A substantial literature (discussed below) seeks to understand the so-called crime-terror nexus.There are two primary aspects of the nexus: (1) terrorist groups engaging in crime, and (2) links between terrorist organizations and criminal organizations. 7Our paper focuses on the latter.While a great deal has been written about terrorists' involvement in crime, much remains to be known about inter-organizational cooperation among terrorists and criminals.How common is such cooperation?What kinds of terrorist groups are likely to cooperate with criminal groups?
We address these questions to better understand the increasing threat of the crime-terror cooperation.Current literature on the topic often does not go beyond individual cases or examine it crossnationally.Additionally, much of the work focuses on the drug trade only, while we consider involvement in crime more broadly.This paper introduces new data and presents one of the first quantitative analyses of why some terrorists cooperate with organized crime.Understanding which factors lead to such cooperation complements qualitative work on the crime-terror nexus, and helps to identify new methods for counterterrorism practitioners to counter those links.
The next section defines key concepts and offers an overview of the existing literature on cooperation among terrorist groups and the crime-terror nexus.The third section introduces a theoretical model of how terrorist groups make the decision to cooperate and presents related hypotheses.The fourth section describes the data employed and outlines a new variable on cooperation of terrorist groups with organized crime groups.The fifth part shows the findings of the analysis, while the final section discusses the results and important implications.

Understanding terrorists, criminals, and cooperation
It is important to define terrorism, terrorist organizations, organized crime and cooperation.There are a variety of definitions of terrorism. 8This article employs the common definition of terrorism as premeditated threat or use of violence by substate actors to obtain a political or social goal through the intimidation of a wider audience beyond the immediate victims. 9The main elements of this definition are (1) violence, (2) political or social goals, and (3) the intimidation of the wider audience.The second element in particular is important for distinguishing this kind of violence from less-political violence, crime, such as assault during a robbery or the murder of a business rival.
We follow others in defining terrorist organizations as subnational political groups that use terrorism. 10The "subnational" aspect suggests they are not state actors, even though some receive support from states.Terrorist groups are fundamentally political organizations, motivated by some goal such as imposing their ideology or religion on a government or achieving political rights for a group.Organizations must use terrorism (as defined above) to be terrorist groups.Some scholars specify that groups must primarily use terrorism as opposed to other tactics to have the label, but we are agnostic about particular levels of terrorism required. 11ow do terrorist groups and organized crime differ?The difference between terrorists and criminals lies in their motivation. 12Terrorists use crime to obtain resources to pursue their political or religious goals.In comparison, criminals only use violence to secure and maintain their operations with the goal of profit-maximisation over politics. 13Hoffman also observes a difference in the motivation of terrorists and criminals; terrorists are altruists and believe they serve a greater good, while criminals serve no cause and desire only to increase their personal material wealth. 14rganized crime groups seek to maintain the status quo or an environment beneficial to their illegal and profit-making activities. 15Violence is employed only to maintain this environment.Many organized crime groups have little resemblance to the popular perceptions of highly structured organizations such as the Mafia.Increasingly, organized crime groups are often networks of actors, only loosely affiliated rather than hierarchical organized, as a result of factors such as responses to law enforcement pressure. 16The crimes committed by such groups can range from money laundering, human trafficking, drug smuggling and counterfeiting of goods.Consequently, an organized crime group is defined as a structured group of individuals dedicated to committing serious crimes to gain profit or material benefit.This is a paraphrased version of the UN definition, although other valuable definitions exist. 17The dividing lines between terrorist and criminal groups are sometimes blurry, but generally a group can be put in one category or the other due to its primary motivations and activities. 18ooperation can be defined as sustained working across organizational boundaries for positive ends for those involved. 19Cooperation requires meaningful interaction between both groups, and not only ideological affinity or rhetorical support. 20Terrorist organizations cooperating with others, such as other terrorist groups, is a relatively rare event. 21Terrorists are clandestine actors who aim to keep their operations a secret.Nevertheless, terrorists do occasionally cooperate with other groups, whether of the same type (terrorists), or others, such as criminals.Why do individuals or groups cooperate?Almost anything can be achieved by cooperation because one is not limited to its own resources and expertise. 22Cooperation therefore opens doors that would otherwise be locked. 23cholars of terrorist organizations and related phenomena (rebel groups, militant groups) increasingly study cooperation among these types of groups. 24Some work looks at consequences of such associations, 25 while other research studies the sources of cooperation. 26A diverse set of factors seems to explain cooperation, such as shared ideology and power asymmetries.
A larger and less quantitative literature analyses cooperation between terrorists and criminals.Many articles employ theoretical models and case studies to explain why terrorists do or do not cooperate with organized crime. 27The literature can be divided into two camps.The first argues that terrorists and criminals can cooperate without substantial complications. 28Makarenko introduces the concept of a crime-terror continuum.This describes a dimension with terrorist groups on one side and organized crime on the other end to explain the different kind of linkages and transformation within crime-terror cooperation.This model was later refined to consist of two planes rather than one dimension. 29he other camp focuses more on issues within cooperation of terrorists and criminals and argues the different motivations hinder actors' ability to find common ground and establish a relationship. 30n one of the earliest works on the subject, Dishman argued that terrorists are unlikely to cooperate with criminals, and instead create own "in-house" capabilities. 31Hutchinson and O'Malley make a similar argument; small terrorist groups can self-generate their funds and equipment, whilst larger and more organized groups create "in-house" capabilities. 32The authors argue that the resulting lack of need, as well as "ideological distinctiveness," makes cooperation between terrorists and criminals unlikely. 33ew articles employ quantitative methodology to study crime-terror cooperation.An important exception is Perliger and Palmieri's work on instances of cooperation between terrorists and criminal groups. 34This is a dyadic (pairs of groups) analysis, looking at what shared attributes of the two types of groups are related to their degree of cooperation.For example, they find that terrorist groups and criminal groups have higher levels of cooperation when they share a religion or come from the same geographic region.This type of analysis is helpful, but is also distinct from our own, which seeks to show which terrorist group attributes are associated with cooperation with organized crime.
Other work looks at overlapping themes, such as terrorist group involvement in crime.For example, Asal et al. examined factors leading to terrorist group involvement in drug trafficking. 35hey find that terrorists with ethnic motivations are especially likely to be involved in the drug trade.Semmelbeck and Besaw study 183 terrorist organizations and find group size and religious ideology associated with crime involvement. 36These studies are valuable, but they are also different from ours because they look at crime involvement by terrorists, and not necessarily cooperation with organized crime groups.

Rational terrorists, efficiency-security trade-off, and cooperation
Our argument starts with the assumption that terrorist organizations are rational actors.This is consistent with a long line of research, 37 although aspects of it have been debated. 38The idea of rationality does not imply legitimacy or omniscience, but simply that organizations take actions that they think will most efficiently bring about their desired outcomes.
A more specific framework for understanding terrorist behavior is the efficiency-security trade-off.Generally, this is "the interplay between the need to act collectively and the need to individually assure trust and secrecy within these sensitive collaborative settings." 39The main idea is that higher efficiency behavior by terrorist groups is associated with lower security.Another way to state this is that when groups have lower security levels, they seek out efficiency.Most research using this framework focuses on the organizational management between leaders and operatives within terrorist or criminal groups, and on their organizational structure. 40For example, Shapiro describes the relationship between leaders of a terrorist group and their operatives as a principal-agent relationship with a high degree of uncertainty and problems of trust. 41hile scholars usually use the efficiency-security trade-off to explain decisions within a group or network, we apply it to decision-making process leading to inter-organizational cooperation.If the efficiency-security trade-off can be used to explain internal decision-making, then it can also be applied to the choice to cooperate with other groups.Intragroup relations and processes are important for the survival of a group and the success of operations.Problems with intragroup processes are solved internally and with terrorist groups trying to find a balance between the security and the efficiency of the group. 42Therefore, intergroup relations must be established through the efficiencysecurity trade-off, as external factors play an important role in the survival of the group.In summary, intergroup relations are shaped not only by internal factors, which explain internal decision-making and the survival of groups, but also by external factors, which have an impact on terrorist groups.The following theoretical framework builds upon the work of Morselli and others, and it aims to explain the processes which lead to crime-terror cooperation.Due to the quantitative approach in this paper, the efficiency-security trade-off is more suitable to examine cooperation with organized crime than the continuum of the crime-terror nexus, which has more blurry boundaries between the different stages.
Every terrorist organization needs money.While some operations are not necessarily expensive, groups also need money for training camps, weapons, bribes or salaries. 43Annual budgets can reach up to hundreds of millions of dollars for the wealthiest terrorist organizations, but for all terrorists, funding is essential. 44After acquiring enough revenue to guarantee a survival of the group, other needs, such as training follow.
How can a terrorist group fulfil their needs?Desouza and Hensgen state "no organization or business is an island, nor can any operate effectively in isolation." 45Due to their clandestine character, terrorist groups are often unable to obtain resources legally and have only limited access to the regular market.Especially after the Cold War and the decline of state sponsorship, terrorist groups have been forced to look for other sources of support. 46Cooperation between criminals and terrorists has therefore become an important factor in the funding of terrorist groups and fulfilling the needs of the organizations, such as seeking expert knowledge or operational support. 47ooperating with other groups can have several benefits for the terrorist organization.It can give access to specialised knowledge (e.g.smuggling), to specialised services (e.g.forged identity documents), operational support (e.g.human trafficking) or financial support (e.g.collusion in trade of illicit products). 48These alliances can be straightforward exchanges of goods for money.For example, the Camorra supplied ETA with arms and documents. 49Cooperation can often be more complex.For example, the Islamic Movement of Uzbekistan supported Central Asian crime groups transporting drugs from Afghanistan to the Caucasus. 50Here, the group is more involved in the organized crime structures and runs smuggle routes on their own behalf.However, not only the terrorists profit from criminals.As Shelley states, focussing on organized crime, "cooperation with terrorists may have significant benefits for organized criminals by destabilising the political structure, undermining law enforcement and limiting the possibilities for international cooperation." 51Terrorist activity may also shift the focus of law enforcement away from organized crime, which allows criminals to work under less pressure.
Applying the efficiency-security trade-off to these instances of crime-terror cooperation highlights the decision-making process behind cooperation.The first consideration is the "security situation" of the group from security services and law enforcement.Because of the clandestine nature of terrorist groups, the security situation is vital.The following example provides a more detailed illustration of the concept of the security situation: Group A is a new terrorist group with ten members and has not been detected by security services.Group B is a more mature group and has fifty members.The pressure from security services is higher on group B than on A, due to the higher number of attacks and the higher member count.The security threat for group A is therefore low.The security situation is influenced by many factors.These can be divided into two main categories: organizational factors such as the age of the group and environmental factors such as the quality of local law enforcement services. 52Each factor influences the security situation positively or negatively, which results in the first half of the efficiency-security trade-off.
The other half of the trade-off is the efficiency of the group.Every group has an interest in working in the most efficient way to support their cause.Due to the clandestine character of terrorist groups, the group cannot work as efficiently as a legal recognised business.To increase group efficiency, terrorists may cooperate with other groups.Krebs examined the terrorist network which was responsible for the 9/11 attacks and found that the inclusion of a wider network has a positive impact on the effectiveness of a group. 53Individuals are part of the broader cell, but the associated group can also provide supporting elements within the networks.Overall, the security situation of a terrorist group influences its decision-making process and affects its likelihood of cooperating with an organized crime group.When the security situation of a terrorist group is safe, the group is less likely to cooperate with an organized crime group.
Global anti-terror efforts against the funding of terrorist groups has made crime-terror cooperation less risky than cooperation with other terrorist groups. 54The intensification of counter-terror policies and measures against funding of terrorism means the flow of money within and between terrorist groups is now more easily impeded. 55Indeed, the increased pressure might be responsible for more terrorist groups linking up with organized crime.Consistent with this, some research suggests that crackdowns on donors explains why terrorists have increasingly turned to criminal fundraising like kidnapping. 56nother reason why insecure groups are more likely to engage in crime-terror cooperation, beside the pressure of law enforcement and state activity, is the outsourcing of non-core tasks.Here, a terrorist group acts like a legitimate company.Non-core tasks, like money laundering or extortion, can carry enormous risk for the terrorist group due to a lack of skills or experience. 57Besides an increased risk of detection due to a lack of experience with crime, development of "in-house" capabilities increases the risk of detection in general.Those groups are then not just under the attention of counter-terror law enforcement, but also of "traditional" law enforcement.
Picarelli and Shelley note narcotics trafficking is the riskiest crime field for terrorist groups to engage in. 58Makarenko and Mesquita also argue that terrorist groups do not establish "in-house" capabilities out of fear of the organized crime groups, which control the illicit market and want to maintain market stability and predictability. 59To fulfil the needs of the group and simultaneously reduce the risk, terrorist groups cooperate with experts in these fields, namely organized crime groups.As the old cliché goes, "If you want the job done right, hire a professional."Terrorist groups with a better security situation are less likely to cooperate with organized crime because they are under as much pressure as other groups, while a group with an insecure security situation must focus all their resources on the survival of the group.Hence, terrorist groups with a relatively safe security situation can concentrate their activities on gathering new knowledge and other activities, for example studying the operational nuances of suicide bombing, which can be achieved easier alone or with another terrorist group than in cooperation with organized crime. 60Next, we explore specific factors that influence groups to work with organized crime.

Organizational factors
Perhaps the most obvious explanation for terrorists cooperating with criminals is when they both are involved in the drug trade.Many groups using terrorism also produce or traffic drugs to support their violence, or simply to enrich themselves.Examples include the FARC's involvement in the cocaine industry in Colombia (not only taxing, but trafficking), and the PKK and the Taliban trafficking heroin and other drugs in their respective countries. 61nce these militant groups are involved in an illicit business such as the drug trade, it seems only natural that they would partner with other illicit firms to help produce, transport, sell wholesale, or sell at the retail level.While the drug business is known for competition, there is also a great deal of collaboration among "firms" with complementary skills.Some criminal groups specialize in trafficking, for example, while others specialize in retail sales. 62Sometimes groups might need to work together because one has the monopoly on a particular geographic territory, and another might make an agreement to share it or pass through safely.These arrangements would potentially benefit both the (drug-involved) terrorists and the criminal groups as they seek to find compatible business partners.
We do not expect a perfect overlap between drug trafficking and terrorist cooperation with criminal groups.Some terrorists in the drug business might do so alone.And many terrorists cooperating with criminal groups cooperate on issues unrelated to drugs.Drugs are of course only one type of activity related to organized crime.Additionally, the proposed relationship is debatable.It could be that terrorist groups involved in crime such as illegal drugs do not need criminal groups, and therefore they might avoid working with criminals.However, for the reasons outlined above, militant groups involved in the drug trade seem especially likely to also work with criminal organizations.H1: Terrorist groups involved in the drug trade are more likely to cooperate with an organized crime group than terrorists not involved in the drug trade.
Terrorist cooperation with other terrorist groups is also likely to have an impact on potential collaboration with organized crime.Many terrorist groups cooperate with other terrorist groups.For example, groups carry out joint attacks, train together, and share weapons. 63Terrorist cooperation seems to be beneficial for involved groups as it is associated with increased lethality, 64 group longevity, 65 and tactical innovation. 66Inter-group cooperation is also likely to be related to working with organized crime.Inter-terrorist cooperation could lead to organized crime cooperation through several mechanisms.First, terrorists that are used to working with other terrorists are simply already accustomed to external relations, so building ties with criminal groups could be a logical next step.Such terrorist groups might have already delegated tasks to a peer group, such as logistics, so it might seem reasonable to also delegate some fundraising or money laundering to a criminal organization.
Second, regarding the efficiency-security trade-off, extant inter-terrorist cooperation suggests groups have already calculated they are willing to give up some operational security for potential benefits.They might make the same estimation regarding working with criminal groups, and find it similarly beneficial.Overall, cooperating with one type of group should increase the likelihood of cooperation with another type.Furthermore, the more terrorists cooperate with other terrorists the more likely they should be to cooperate with organized crime.

H2:
The more cooperative relationships a terrorist group has with other terrorist groups, the more likely it is to cooperate with an organized crime group.
Age and experience are other important factors for clandestine groups.Hoffman compares the survival of a terrorist group can be compared to Darwin's theory of evolution. 67New and old groups must compete with others for public audience and the rare resources.Old groups have already demonstrated their capacity to survive for a long period and adapt their activities and strategies to law enforcement and security efforts made by the state.Therefore, funding and supply chains are changed or established in a secure way.In comparison, young groups do not have the experiences with law enforcement services and need to establish sources of funding and supplies.Young groups tend to struggle not only with the daily survival, but also have the need to build up the organization and consequently require more resources.Ideally, they might cooperate with organized crime, but since they are young they might not yet have the links or enough of an established reputation to gain the trust of criminal groups.
Older terrorist organizations, on the other hand, are likely to have members with individual connections to organized crime who can broker agreements.Furthermore, the group will have an established reputation that can reduce uncertainty for criminal groups that might potentially work with it.In general, older terrorist groups already adapted to their environment can therefore concentrate on other activities which might be too difficult or risky for younger groups.For example, branching out through cooperation with other organizations.At this point, when the survival of the group is secured and the daily struggle is manageable, the risk of cooperation between criminals and terrorists is no longer outweighed by the profits.
H3: An older terrorist group is more likely to cooperate with an organized crime group than a younger terrorist group.
Terrorist groups also differ substantially in their activity levels.Well-known groups like ISIS or Al Qaeda have caused many casualties.Apart from very active and lethal groups, many groups have caused no fatalities. 68The activity level of a terrorist group and its lethality also influences cooperation of terrorists and criminals.An active group that wants to maintain their level of activity and commit more attacks needs more resources than a group which is less active.Several studies have shown that cooperation among terrorist organizations is associated with group lethality. 69Increased capabilities for operations, in form of manpower, also mean an increased need for resources.Therefore, groups which are already more active than other groups can only maintain their level of activity with cooperation.Consequently, the decision to cooperate is influenced by the resource demand of the terrorist group.
In general, higher activity requires operational support, which can also be provided by organized crime networks, due to their skills such as forging documents and human smuggling. 70Furthermore, increased activity increases the pressure from law enforcement and other security services.Cooperation with another terrorist group would therefore carry a higher risk, due to this pressure.Crime-terror cooperation carries less risk for the terrorist group and offers more material and operational support to the group.Summing up, a higher activity of terrorist groups leads to a higher demand for resources to maintain this level of activity.This higher demand for resources can be fulfilled by cooperation with organized crime, which also carries less risk, in a narrow perspective, for the terrorist group than when they commit crimes themselves or cooperate with other terrorist groups.
H4: A more active terrorist group is more likely to cooperate with an organized crime group than a less active group.
Another organizational factor that might be relevant is state sponsorship.Scholars have long argued that the decline in state sponsorship at the end of the Cold War forced terrorist organizations to turn to crime. 71This suggests a substitution effect between state support and criminal involvement.However, state sponsorship did not end in 1989.Certain states continue to fund organizations that use terrorism. 72Consistent with the idea of a substitution effect, it seems likely that terrorist groups receiving funds from states have less need to turn to organized crime for collaboration.Meanwhile, groups without such support should be more likely to work with criminals.The substitution logic relates to the security-efficiency argument because organizations with sponsors might have fewer security concerns than organizations with such a funding source.This would further imply a lower likelihood of reaching out to organized crime for more efficient operations.
Beyond pure substitution, it has been noted that state sponsors sometimes try to restrain the behavior of their proxies.For example, state sponsorship can lead to less civilian victimization. 73imilarly, states might discourage groups from cooperating with organized crime.Overall, these various possible causal mechanisms suggest the following hypothesis: H5: A terrorist group with a state sponsor is less likely to cooperate with an organized crime group than a group without a sponsor.
A final organizational condition that is likely to affect crime involvement is territorial control.Some terrorist organizations control vast swaths of land, while others are more underground and operate from their own homes in urban areas or dispersed across the land.Scholars find territorial control associated with a host of important outcomes, from civilian victimization 74 to groups employing female fighters. 75Holding territory has been linked to group involvement in crime in a number of ways.Groups with territory can use it to grow and produce drugs, and also have a population they can exploit through kidnapping and extortion. 76nce terrorist groups are already involved in crime themselves, it is probable that they will then work with organized crime groups.This seems especially likely for groups involved in trafficking of any products (or people), since they often need criminal groups to handle logistics or retail sales.Furthermore, the territory held by military groups might overlap with or border territory of criminal organizations, so some coordination is necessary.Overall, holding territory should be associated with cooperation with organized crime.

H6:
A terrorist group that holds territory is more likely to cooperate with an organized crime group than a group without territory.

The broader environment: state capacity
Beyond organizational factors, characteristics of the broader environment are likely to play a role in the decision of terrorist groups to work with organized crime.We focus on state capacity in particular.Terrorist organizations are based in a variety of different types of states.Some try to survive in highlycapable countries with advanced security institutions while other groups operate more comfortably in relatively weak states.These differences have implications for the efficiency-security trade-off.Terrorists in the most capable states are highly insecure, as government counterterrorism forces could be monitoring their phone calls, infiltrating their organization, or watching their group members meet.The theorized trade-off suggests that groups in these situations are more prone to seek out advantages however they can.One likely effort to mitigate the challenges of operating in this kind of environment would be to make partners with organized crime.As discussed above, the potential benefits of such cooperation are worth the possible risks when situations are dire.
The notion that cooperation between terrorists and criminals would be relatively common in more-capable states is consistent with foundational work on the crime-terror nexus.Makarenko's important research outlines a continuum of possible terrorist-criminal actions, from cooperation to "convergence" into more threatening hybrid organizations. 77Makarenko and Mesquita argue that the latter situation is more likely in weaker states. 78Potential convergence among terrorists and criminals is argued to have occurred in countries like the Philippines. 79In comparison, terrorist groups in highly capable states face strong law enforcement, better monitored borders, and cooperation with other states via organizations like Europol and therefore are in a more complex operating environment. 80As a result, relationships between organized crime and terrorist organizations are different in such states.The environment is less secure, and terrorists need to be more adaptive and operate more clandestinely.Crime-terror cooperation can be more useful and more secure than committing substantial crimes on their own, as groups do in a weaker states. 81In sum, this suggests the following hypothesis.

H7:
A terrorist group based in a more capable state is more likely to cooperate with an organized crime group than a group based in a less capable state.

Research design
To test the hypothesis, we combine the Bad Allied and Dangerous 1 (BAAD1) terrorist group dataset by Asal and Rethemeyer with a new variable coding cooperation with organized crime groups. 82The unit of analysis is the terrorist group.Models use Asal and Rethemeyer's organizational data and country-level variables of each group's home base state to explore correlations between these attributes and connections to organized crime groups.

Data and operationalization
The BAAD1 dataset consists of 395 terrorist organizations from throughout the world and provides information ranging from the organizational age to the motivation of the group.The cross-sectional dataset includes data from 1998 to 2005 and is based on the Memorial Institute for the Prevention of Terrorism's (MIPT) Terrorism Knowledge Base.This is a widely-used list of terrorist organizations, and the article introducing it has been cited more than 500 times as of summer 2022.There are other datasets of terrorist groups, but BAAD1 has the advantage of including groups from both the 1990s era of ethnonationalist and leftist violence as well as the more recent Salafi-jihadist-dominated wave of terrorism. 83he dependent variable is cooperation between a terrorist group and an organized crime group. 84It is a dichotomous variable coded from a variety of sources, including academic journals and books, government reports, and newspaper articles on LexisNexis.The variable is coded 1 if there is evidence of cooperation with organized crime from 1998 to 2005 and 0 if not.There are fifty terrorist organizations coded as cooperating with organized crime in our data (forty-eight in the sample we have all variables for), about 12 percent of all terrorist groups.Some examples include al Qaeda in the Islamic Maghreb providing route security for crime groups in the Sahel and Basque Homeland and Freedom (ETA) working with the Neapolitan Camorra.
Most explanatory variables that are organizational attributes come from Asal and Rethemeyer 2008 unless otherwise noted.To test H1, Drug involvement is a dichotomous indicator coded 1 if there is evidence that the group was involved in illegal drug business. 85Almost 9 percent of our groups are coded for such involvement.Terrorist allies is used to test H2.It is a count of the number of other terrorist organizations that the group is cooperating with. 86For H3, models include Group age, which indicates the organizational age of a terrorist organizational-the number of years since their founding until the end of 2005.It ranges from one to eighty-seven years.To test H4, models include Fatalities, the total fatalities associated with the group between 1998 and 2005.Due to the extreme range of the variable, we use a natural logarithm.State sponsored is included to test H5.It is a dichotomous variable indicating material support from a country government.To test H6, we include Holds territory, a dichotomous variable coded 1 if the terrorist group controls territory, particularly movement into and out of it, within a country.
Regarding H7, about state capacity, BAAD1 codes a "home base" country for each organization, the country in which it primarily operates.Models include several state-level variables, all of which are coded for their value in 1998 to reduce the possibility of reverse causality.To test this hypothesis, State income p.c. (log) is the per capita gross domestic product for the group's home base country in 2010 constant dollars.The source is the World Bank Development indicators, via the Quality of Government project. 87It is a natural logarithm.
The analysis also includes several control variables.Regarding organizational control variables, we include Group size, which is an ordinal variable indicating the membership size of the terrorist organization.It varies from 0 ("0-100 & low confidence"), 1 ("100-1,000"), 2 ("1,000-10,000") to 3 ("10,000 or more").It is possible that larger groups should be more likely to cooperate with organized crime, as they need more resources to sustain their membership, and could be more visible for criminal groups looking for partners.We also include ideological motivations of the group, which is often mentioned when it comes to activities related to organized crime.There is no academic consensus whether ideology affects crime-terror cooperation, especially involvement into drug related crime.Some scholars suggest that religiously-motivated groups should be less likely to engage in crime due to legitimacy concerns, but there is mixed support for this empirically.To take into consideration the possibility, we include binary variables indicating if the group has one of the three types of motivation: religious, or ethnonationalist, or leftist.As for state-level control variables, models include State regime type, the liberal democracy measure from the Varieties of Democracy (V-Dem) project, also via Quality of Government. 88It ranges from 0 to 1, where higher levels indicate more liberal democracy.

Model
Models are logistic regression due to the dichotomous dependent variable.If ordinary least squares is used, results are similar.Standard errors are robustly clustered at the country level to take into consideration the non-independence among groups that exist in the same state.The use of a variance inflation factor test suggests that multicollinearity among explanatory variables is not a problem.Pairwise correlations suggest that there could be an issue with Group size: it is correlated with Fatalities (log) at .52, and with a few other variables at greater than .4.To make sure this does not affect results, we report a model with Group size excluded.

Results
We first report baseline models in Table 1, testing for bivariate correlations between the hypothesized variables and the outcome variable.Most of the models show initial support for the hypothesis, with statistically significant and positively-signed coefficients as expected on Drug involvement, Terrorist alliances, Group age, Fatalities (log), and Holds Territory.This is only tentative support, however, as the models do not control for other factors and therefore might suffer from omitted variable bias.Furthermore, it is notable that two variables do not show hypothesized results: State sponsored is significant and positively signed, while the hypothesis expected a negative relationship.The coefficient on State income p.c. (log) is statistically insignificant, suggesting a lack of a relationship.
Table 2 shows the main results.Model 8 only includes the control variables.Model 9 includes only the variables representing hypothesized relationships.Model 10 is the full model, our primary model.Two other models test the robustness of the results.Model 11 excludes Drug involvement in case it is seen as possibly too correlated with the dependent variable.Model 12 excludes Group size in particular because of possible multicollinearity.Across all three models which include it, Drug involvement is statistically significant and positively signed.Terrorist groups involved in the illicit drug business are more likely than other groups to cooperate with organized crime, consistent with H1.Terrorist allies is significant and positively signed in all models in which it appears.Terrorist organizations with more terrorist allies are more likely to cooperate with criminals.This provides support for H2.Group age is also positive and statistically significant, suggesting older terrorist groups are more likely to cooperate with organized crime.H3 is supported.Fatalities (log) is also statistically significant and positively signed.Terrorist organizations that kill larger numbers of people are more likely to work with organized crime, consistent with H4.
The coefficient associated with State sponsored is statistically significant in three of the four models.However, it is in the opposite direction of what was expected in H5-it is positively signed, suggesting groups with state sponsors are more likely to cooperate with criminal groups.This is surprising given the common assertion that terrorist involvement with organized crime is a substitute for state sponsorship-and this is why connections to organized crime have flourished since state sponsorship became less prevalent after the end of the Cold War. 89It could be that groups getting involved in crime themselves is a substitute, but that alliances with criminals follow a different logic.More research on the precise connection between state sponsorship and terrorist connections to organized crime would be fruitful.But it is interesting that state sponsorship and cooperating with criminals seems to go hand-in-hand in our sample, going somewhat against the conventional wisdom.
The coefficient associated with Holds territory is statistically insignificant in all but one of the models.There is little support for H6.This is an interesting non-finding given the importance of territorial control for activities like drug production.However, it could be that some groups holding territory are seeking legitimacy in the eyes of the public, and therefore they try to distance themselves from organized crime groups.The Taliban, for example, at least made an effort to ban opium production for several years. 90Overall, however, territorial control is apparently orthogonal to cooperating with organized crime, as some groups that hold territory do it, while others do not.Finally, State income p.c. (log) is also statistically significant and positive in three of the models, suggesting that terrorist organizations in more capable or wealthier states are more likely to cooperate with criminal groups.This lends some support to H7.There is less support for this hypothesis than others, however, as the variable does not have a statistically significant coefficient in models without control variables.Regarding control variables, Group size is statistically significant and positively related to cooperating with criminals in two of the three models in which it is included.There is some evidence for the idea that larger groups are more likely to work with criminal groups.None of the group motivation variables are robustly statistically significant.The crime-terror nexus seems to be more related to resources than ideology.However, the coefficient on Ethnic motivation is statistically significant and positively signed in two of four models.There is partial support for the notion that ethnically motivated groups are less likely to work with criminals.This could be due to the fact that these types of groups have an easily identifiable potential base of support, and therefore have less need for crime.Alternately, it could be that these groups are especially likely to be legitimacy-seeking, as they are more often quasi-state actors.Finally State democracy is statistically significant and negatively signed in two models.Thus, there is some evidence that in more democratic countries, cooperation with criminals is less likely.Some research suggests that illicit group survival is challenging in lessdemocratic states, 91 and therefore terrorists and criminals might be more likely to reach out to each other in these situations.

Discussion and conclusion
This paper asks which factors lead some terrorist groups to collaborate with organized crime groups while others do not.Drawing on rational-choice assumptions and the efficiency-security trade-off of terrorist group strategy, we presented a number of possible explanations.Quantitative analyses using a newly-coded variable measuring terrorist-criminal cooperation found support for much of the argument.The following factors are robustly associated with such cooperation: involvement in crime, connections with other terrorist groups, group age, and group lethality.There is some evidence for an association between criminal cooperation and these attributes: state sponsorship, state capacity, and group membership size.The state sponsorship finding is somewhat surprising since the literature generally posits that the crime-terror nexus was bolstered by a decline in state sponsorship.There is also some evidence of a negative relationship involving ethnic motivation or state democracy.
These results offer important contributions to the study of terrorism and its intersection with organized crime.The paper makes a theoretical contribution by considering how the efficiencysecurity trade-off can apply to inter-organizational relationships.It is also one of the first quantitative analyses of cooperate between terrorist and criminals, and introduces new data on such cooperation.
Furthermore, findings of this study contribute to research and suggest the following implications for counterterrorism.The increasing role of cooperation with organized crime confronts states with new challenges.Terrorist cooperate strategically with organized crime to fulfil their needs; law enforcement can target these ties.This suggests a need for increased cooperation among agents working on counterterrorism and those combating organized crime.If a state is not able to disarm a terrorist group, it could also focus on "drying out" a terrorist group by cutting their sources for funding and other resources.However, the data gathered and findings also emphasize that some groups prefer to engage in crime alone, rather than collaborate with other groups.Only about 12 percent of terrorist groups had meaningful cooperation with criminal organizations.Thus, law enforcement needs to keep an eye on links between terrorists and criminals, but it also should not forget about terrorists engaging in crime on their own.Further research on cooperation between terrorists and organized crime can go in a number of directions.One could be to study the effect of the organizational age on cooperation with longitudinal data to reduce questions about reverse causality and have more confidence about specific relationships.Another direction could be the relationship between cooperation, "in-house" crime capabilities and the home-base state of a terrorist group.Do only the capabilities of the security forces affect cooperation or does support by the civilian population also have an impact?How is cooperation initiated and how does it evolve over time?A limitation of this study was that it did not use data that changed over time, but important dynamics are likely.
This study focused on terrorist organizations and the factors that might make them more or less likely to cooperate with criminals.A different article could gather data on criminal groups to see why (presumably) only some of them cooperate with terrorist organizations.Additionally, why do terrorist groups and criminals sometimes fight each other as opposed to cooperate?Finally, it would be interesting to see if the structure of the organization-hierarchical or more horizontal-might affect its propensity to cooperate, as some scholars have suggested. 92More fine-grained analysis could also look at the roles individuals such as brokers play in inter-organizational cooperation. 93Future research should gather more nuanced data to continue work on this important topic.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Notes on contributors
Brian Phillips is a Reader in the Department of Government at the University of Essex and an affiliated professor at CIDE in Mexico.His research and teaching interests include terrorism, civil conflict, and crime.
Alexander Schiele is a PhD student at the University of Essex and a lecturer at the Berlin School of Economics and Law.His research interests include policing terrorism, social network analysis and the crime-terror nexus.