The MS-13 Hierarchy Debate: An Exploratory Social Network Analysis of a Shot-caller Replacement by Regional Leadership
Abstract
This article provides an exploratory social network analysis (SNA) mixed methods case study of an MS-13 (alternatively, MS13) street gang clique (MS-13 regional component) in Southern California. It specifically uses longitudinal analysis looking at a before and after ego network comparison punctuated by the event of the replacement of a leadership figure within MS-13 known as a “shot-caller.” This leadership shift led to an organizational change and was initiated outside the clique by MS-13 regional Los Angeles leadership in a meeting event. The researchers built the networks from detailed court documents, including the 127-page Racketeering Influenced Corrupt Organization (RICO) indictment United States of America v. Balmore et al. (2017) filed in the Central District of California. The regional MS-13 leadership appeared to have effective intelligence on this sub-group of the overarching MS-13 network based on key metrics after the shift, including a reduction in the diameter (number of steps across the network) of the “after” ego networks. Our findings provide a small indicator that some components of MS-13 may have a more hierarchical networked structure than previously identified, calling for more research into this area.
Introduction
This article provides an empirical social network analysis (SNA) mixed methods case study of an MS-13 (alternatively, MS13) street gang clique.[1] It compares the network before and after a leadership change: the replacement of what is known in MS-13 as a “shot-caller.” This leadership shift led to an organizational change and was initiated by a council of MS-13 regional Los Angeles leadership and imposed on the Coronado clique from outside, although clique members were present at the meeting. It should be noted that clique refers to an MS-13 local group controlling a particular turf and is not here used in the context of a formal mathematical clique as is common in the SNA literature.[2] We built the networks by coding detailed court documents, including the 127-page Racketeering Influence Corrupt Organization (RICO) indictment United States of America v. Balmore et al. (2017) filed in the Central District of California.[3] The indictment included MS-13 co-offending events between 2007 and 2017.
Outside regional MS-13 leadership appeared to have effective intelligence on this clique. We conclude this because the shot-caller (S1) was replaced by an individual (S2) with high betweenness centrality, which is commonly understood as a brokerage metric and an indicator that S2 occupied a key position within the network. S1 was subject to significant disciplinary action after his replacement, which is consistent with the fact that shot-callers are the ones with authority over disciplinary action. Notably, our results indicate that S2, upon taking power, did not engage in extensive disciplinary action against gang members under his control. One interpretation is that he did not want to rock the boat, so to speak, as he established himself in the new position. By comparing the before and after networks, we found nuanced results. By some metrics, connectivity went down; by others, the network became more reachable with a smaller diameter (fewer steps to get across the network).
Our assessment is that there are multiple likely causes for these results. First and foremost, the after network appears to have become more efficient, using fewer connections yet reducing the distance across the network after the shot-caller change. This suggests that the outside regional leadership council of MS-13 had a keen understanding of the internal organizational structures of the network. The important organizational shift in the network was ultimately more efficient despite reduced connectivity as measured by fewer ties and lower density. This further suggests that MS-13 as a network may be more tightly connected than previously understood, given it has been described by many scholars as a loose network. It has been described by some scholars as a loose network with limited communication between spatially distributed, decentralized cliques and limited, albeit variable hierarchical control.[4] The structural variation among different parts of the MS-13 network is briefly discussed in the literature review below. The results of this study are a possible indicator of effective control of a broader network despite cell structures.
Second, these increases in efficiency may also have been caused by the reduction in the size of the network after the reorganization. Indeed, we provide results for two different ways to structure the network, and in the aftermath of the reorganization, there were three members who became “isolates” with no ties to others (See Appendix A).[5] We assess this had a minimal impact on the analysis regardless of whether isolates were included or excluded.
Third, we also discuss the role of missing data, an ever-present issue in the analysis of law enforcement evidence of dark networks, which inherently attempt to conceal their operations and may have played a role in our results.
This article makes an important empirical contribution to the literature on the loose or tight coupling of MS-13—a transnational street gang—structures.[6] While our data is not a comprehensive mapping of the entire MS-13 network—that data may indeed be impossible to gather—it does indicate that outside regional leadership council figures have effective intelligence on the network’s cliques and subgroups. This is a small but important indicator that the MS-13 network is potentially more tightly structured than previously believed. Of course, more research is needed on the degree of hierarchical control within MS-13 network structures.
The article proceeds as follows: (1) it begins with a review of the literature on social network analysis of dark networks and the qualitative and quantitative literature on MS-13 network structures; (2) it then describes the methodology used, specifically discussing social network analysis centrality metrics, gathering of data from court documents, network topography metrics, and community detection or subgroup analysis; (3) it next presents the results using descriptive statistics generated from software such as UCINET and Gephi; (4) it then discusses and analyzes those results utilizing visualizations of the before and after networks in addition to tables for network topography metrics; (5) it concludes by reviewing that our findings suggest a greater degree of hierarchical control in MS-13 at the regional level than previous academic literature has acknowledged, limitations of our analysis, and avenues for future research.
Literature Review
Social network analysis has exploded in the social sciences in the last five decades though it dates back to the 1930s and expanded in the 1960s under the tutelage of Harrison White whose students wrote seminal articles.[7] In the last two decades its explosion was largely due to the advent of the personal computer and widespread social network analysis software packages such as UCINET, GEPHI, R, ORA, etc.[8] There is a growing body of literature using social network analysis (SNA) for the study of criminal networks or dark networks, which include terror, insurgent, criminal, and resistance movements.[9] This article builds upon previous research using court documents to create a network from court records, i.e., charging documents on a large-scale MS-13 conspiracy case with a detailed 127-page indictment.[10]
Previous research has shown that the use of unstructured court and law enforcement data can be a fruitful source of network analysis data, though with significant processing by researchers.[11] Researchers reviewing the literature on SNA of drug and criminal networks have identified data issues such as missing data as hindrances to otherwise high-quality data.[12] Scholars such as Berlusconi (2013) have compared networks created by different types of court data, including arrest warrant data, wiretap, and sentencing documents. She found that while arrest warrant data had missing links that may lead to biased results, it was a “purposeful sample” of prosecutor data, and key players were included.[13] This is important here as the current study focuses on a highly detailed indictment to garner ties about key players and their roles as shot-callers, which is not necessarily determined by their SNA position within the network but is rather an operational leadership role as defined within the social constructs of the MS-13 gang and its US prison environment. It should be noted that the 13 in MS-13 stems from an alliance, which may be fraying, with the Mexican Mafia prison gang known as La Eme.[14] This analysis assesses the before and aftereffects of the removal and replacement of a shot-caller based on data and descriptions from within the legal documents produced by prosecutors. In this sense, it is within the vein of work by Bright et al. (2015) which combines individual actor roles with network metrics to find important actors.[15]
The analysis presented here also relies on subgroup network analysis of the two shot-callers in the period before and after the transition from one to the next. Subgroup analysis is highly useful for identifying cliques or cells within dark network structures. The Louvain algorithm in Gephi allows for the detection of communities that are denser than we would expect to appear randomly.[16] The shot-callers were consistently identified in different communities by this algorithm, as the results section will discuss.
The Debate: Is MS-13 Tightly or Loosely Structured?
Defining the structure and capacity of Mara Salvatrucha (MS-13) is a complex task. The gang is a lateral network with diverse components with varying degrees of connectivity over time and space. The gang initially emerged in Los Angeles and then spread to El Salvador and other parts of Central America (notably Honduras) and then to other regions in the United States.[17] Its structure, capacity, and leadership has been described in various ways with differing views or takes on its network configuration in various locations. Some question if it is a hierarchy or federation;[18] others emphasize its links to state corruption or describe its evolution in various locales.[19]
InSight Crime describes the MS-13 organization as having diverse leadership and organizational models across various components of the network. Some nodes are hierarchical, some transnational, some compete with other cliques, some control other cliques within a region, while others are organized within “programs” that control numerous cliques, sometimes from a distance.[20] Sullivan and Bunker have described geographic variations in MS-13 command and control (C2) structures,[21] while Sullivan and Logan have described the lateral leadership arrangements as “networks of influence.”[22]
MS-13, like many dark networks, has suffered fragmentation. For example, Sullivan and Bunker have documented a break away faction from MS-13 and subsequent internecine conflict. The MS-503 (Revolucionarios) has come into conflict with the main portion of MS-13 (led by the La Ranfla Histórica) leading to the March 2018 shooting of a MS-503 member in Mexico City.[23]
One core question involves assessing the MS-13 network’s transnational capacity (especially among its branches in various parts of the US and the relationship between these branches and those in El Salvador). A recent US indictment demonstrate US prosecutors believe they can prove that US based MS-13 ask permission of El Salvadoran La Ranfla leadership to kill in the United States see United States of America v. Trejo Chavarria et al. (2022). According to the Galveston, Texas indictment:
In or around September 2017, CARLOS ALEXI GARCIA-GONGORA, A/K/A “GARCIA,” A/K/A “LIL MALIGNO,” and others known and unknown to the Grand Jury, contacted MS-13 leaders in the United States and El Salvador and requested and received permission to murder Bryan Alfaro, a twenty-year-old male, because Bryan Alfaro was suspected of being a rival gang member.[24]
The US based MS-13 members would also send money back to leadership in El Salvador further demonstrating transnational organizational capacity.[25]
A second structural question is whether the various cliques within these various geographic hubs are more hierarchical than the literature implies. This question applies both within and among cliques and across the entire network and its components. As previously noted, Wolf, Correa-Cabrera et al., as well as Hershberg, Maguire, and Dudley, question the depth of transnational connectivity.[26] McGuiness emphasizes decentralized leadership over a mafia-style hierarchy. [27] Silva Ávalo described connectivity among the cliques in Central America, Northern Virginia, and Massachusetts.[28] While Nuño and Maguire suggest that MS-13 in Los Angeles may be more structured, they have doubts about the extent of transnational connectivity.[29] The recent gang crackdown by El Salvadoran president Bukele has led to significantly diminished territorial control by MS-13 in El Salvador, though at the cost of serious critique of human rights abuses.[30] This article seeks to help fill the lacuna in understanding the variations in MS-13 network organization and dynamics.
Structure and Gang Desistance in the MS-13 Literature
A key area where MS-13 and institutionalized gang structures have been carefully examined by scholars, is in the gang desistance literature. This literature led by scholars such as Cruz et al. (2023) helps us to answer important questions such as under what circumstances and pathways can individuals leave gangs such as MS-13?[31] Cruz et al. (2023) identify three pathways to leaving a gang: religious conversion, a secular pass or permission from leaders, and moving away. This is complicated if a gang controls territory like MS-13 in El Salvador. However, our subject here is set of cells operating in Southern California where is more difficult to suggest MS-13 has a level of control with territorial governance given effective US law enforcement institutions. Concepts such as “criminal insurgency,” “criminal enclaves,” “alternative governance,” or “criminal governance” may be more appropriate in El Salvador, or the global south given that these concepts convey some level of “plural” or shared authority with the state. [32] Indeed, Cruz et al. argue there are different gang structures in Guatemala, Honduras and El Salvador. They argue: “The level of coordination between the organizations and the extent of territorial control shapes governance. Organizational coordination determines the predictability of the gangs’ responses, whereas territorial reach defines the area in which those gang responses are in effect.”[33] In other work Cruz and Rosen (2024) find that gang exit via religious conversion helps to consolidate gang territorial control by assuring the gang that deserters wont join rivals.[34]
Key to our discussion is the internal discipline of gang members which has implications for internal organizational structure and the ability to exit gang life. Pyrooz and Decker (2011) point to “hostile” versus “non-hostile” exit from gangs and argue the modal response they found was nonviolent exit.[35] The capacity to punish members implies more hierarchical control and more internal discipline and cohesion. The case study presented here includes examples of internal discipline and our network analysis also suggests, in limited fashion, tighter control than the existing literature implies with its loose coupling phrasing. It should also be noted that MS-13 was recently added to the foreign terrorist organization list in the early days of the second Trump Administration in 2025. This will no doubt lead to increased targeting of MS-13 in the United States and could lead to more decentralized structures, as command and control ties are severed or a weakened.
Methods
Current Study
Inspired by previous research noted in the literature review, the current study focused on two specific research questions: (1) how do networks change after a shot-caller is removed from his position and replaced by a new shot-caller? And (2) what are the effects of the removal and replacement of shot-callers on their criminal involvement?
Data
The data used for this study includes a 127-page RICO indictment based on a multi-agency investigation of MS-13 cliques in Los Angeles. This indictment includes numerous sources of evidence collected through law enforcement, confidential informants, and wiretaps over a 9.5-year period, and the claims introduced in the 2016 indictment include 41 actors, 8 of whom were confidential informants. In total, 33 MS-13 members were indicted. During the investigation, the first shot-caller was removed from the position of shot-caller and was replaced with a new shot-caller on the same day. For this study, the focus will be aimed at the two mentioned shot-callers and changes in their network across the 9.5-year investigation, both before and after the removal and replacement of these key actors. This network includes two shot-callers and 12 actors (N=14).
Conceptualization and operationalization of ties and attributes
The ties among actors included in the dataset are based upon the methodology used by Jones and colleagues (2020), which conceptualized and operationalized social network ties for social network analysis examining RICO indictments (see Table 1). RICO indictments typically include overt acts that are used as evidence of operational connections/ties among actors that are engaged in criminal enterprise as co-offenders.[36] Operational ties were coded as bidirectional. For instance, if actor A told actor B to commit a crime on behalf of the criminal enterprise, the authors created a directional tie from A to B, and this tie would be included in the edge list. If actors A and B committed a crime together, the authors created a reciprocal tie between these two actors. Also, if actor A attacked actor B, a directional tie was drawn from A to B. The following Table 1 depicts the attributes (variables) for analysis.
Table 1. Conceptualization of attributes for analysis
Analytical strategies
We used UCINET to estimate the actor centrality. Three centrality measures were calculated: out-degree, in-degree, and betweenness centrality.[37] Out-degree measures the frequency of an actor’s outgoing ties. In-degree measures the frequency of an actor’s incoming ties. Betweenness measures the extent to which an actor lies on the shortest paths between other pairs of actors. It is thus seen as capturing the potential influence it has on the flow of information through a network.[38] We used Gephi for network visualizations.[39]
Results and Discussion
Qualitative Case Study Description
This case study is built largely from court documents and a 127-page indictment, USA v. Balmore et al. (2017). It should be noted that the grand jury indictment was dated 2016, but it was released in 2017 and indeed includes information and data documenting events from 2007 through 2017. We examine a subset of the data here focused on the replacement of a shot caller of the Coronado clique, referred to as S1. In roughly October 2014, at an MS-13 regional general meeting, including a leadership council, shot-callerS1 was replaced from outside the Coronado clique with a new shot-caller known here as S2. We use network data to create a before and a separate after network for comparison of the structural impact of this on the MS-13 clique. Below is a quotation from the indictment describing the events:
Overt Act No. 95: On or about October 26, 2014, during the MS-13 general meeting, defendants BALMORE, LEVERON, VIDES [S1], MELENDEZ, MELGAR, I. GARCIA, and ARDOIN [S2], Ramos, Perez, and other known MS-13, members, discussed which member would “take the keys” to MS-13, meaning be MS-13’s overall leader in Los Angeles. When nobody volunteered, defendants BALMORE, LEVERON, VIDES, MELENDEZ, MELGAR, I. GARCIA, ARDOIN, Ramos, Perez, and other known MS-13 members agreed that MS-13 would be run by the leaders of each clique. Overt Act No. 96: On or about October 26, 2014, during the MS-13 general meeting, defendant VIDES was removed as the shot-caller of the Coronado clique and defendant ARDOIN was reinstated as the Coronado clique’s shot-caller. Overt Act No. 97: On or about October 26, 2014, during the MS-13 general meeting, defendant BALMORE and Ramos chose defendant I. GARCIA, and two other unindicted co-conspirators, to administer discipline to defendant VIDES for violating MS-13’s code of conduct.[40]
In total, 12 actors had ties with S1 before he was removed as a shot-caller, and he had eight ties after he was removed. S2 had 12 ties before he became shot-caller and seven ties afterward. Table 2 presents the frequency of crime types that S1 and S2 were involved in before and after S1’s removal and replacement, as well as their before and after centrality metrics. There is not much change among most of the crime types after the removal and replacement of the shot-callers. The largest differences were found in the following crime types: meeting, victim of gang discipline, and gang disciplinary perpetrator. Both S1 and S2 were less often involved in gang meetings after their removal and replacement. There was an increase in S1’s victimization by gang disciplinary actions after he was removed as a shot-caller. There was a reduction in S2’s involvement as a perpetrator of gang disciplinary actions once he was replaced as the new shot-caller.
In terms of the individual centrality metrics, S1 had a higher out-degree than S2 before their removal and replacement of shot-callers, meaning that there were more ties going out from S1 than S2. However, after their removal and replacement, there were no ties going out from S1, suggesting that he lost all the privileges of giving orders to other members as a result of losing his rank within the gang. S1 had a lower in-degree than S2 before their removal and replacement, but S1 had a higher in-degree than S2 afterward. This suggests that S1 received more orders from other members or was punished more often by gang disciplinary perpetrators due to his loss in rank. S2 had higher values in betweenness than S1 both before and after their removal and replacement as shot-callers. The results suggest that S2 had more influence within the gang than S1 even before S1 was removed as the shot-caller, which was a sign that S1 had already started losing his influence within the gang even before his removal. The following Table 2 describes the crime types and individual centrality metrics for S1 and S2.
Table 2. Descriptives on crime types and individual centrality metrics that S1 and S2 were involved in before and after the removal and replacement as a shot-caller.
Figure 1 The Before Network (Gephi)
Figure 1 above depicts the network prior to the removal of S1 as a shot-caller. Figure 2 below shows the network after S1 was removed from his shot-caller position and replaced with S2. S1 and S2 are depicted as triangles. The layout algorithm Yifan Hu was used, which pushes peripheral nodes with fewer connections away from the center.[41] Node size varies in terms of betweenness centrality. Compared Figure 1 with Figure 2, we see a dramatic reduction in the betweenness centrality of S1, while S2’s betweenness remains high. Based on the size of the nodes, shot-caller S2 is clearly the dominant actor after he is appointed as shot-caller, although S1 and S2 share many of the same connections.
Figure 2. The After Network
Table 3 below provides the network topography metrics for both the before and after networks.
Based on the metrics presented in Table 3, it suggests that S1’s removal and his replacement by S2 may have impacted their ego networks. We can assess that this impact was likely disruptive because it reduced overall connectivity within the network. Despite the network going down in size from 13 to 10, we see that graph density also went down. We would expect that if connectivity were to remain the same and the number of nodes were to be reduced, that density would actually go up. This was not the case. Density went down despite the network going down in size. Further, we also see that the average degree (average number of ties each actor has) went from 2 to 1.4, and in yet another example of reduced connectivity within the network, the average weighted degree went from point 4.62 to 1.8. This suggests that in the after network, there were fewer repeated communications and actions taken between actors. These all suggest reduced communication and co-offending within the network. Furthermore, the average path length (i.e., the average number of steps between actors) declined from 1.84 to 1.36, suggesting it was easier to access other actors in the network after the leadership change.[42]
SNA and organizational theory often assume that more communication and density within a network leads to better operations and success for that network. Indeed, Karl Weick’s seminal work on sensemaking describes the lack of communication among the Mann Gulch disaster firefighters as one of the reasons for the loss of life.[43] Scholars, such as Morselli, have long discussed the need for communication in criminal and terrorist operations as they make tradeoffs between security and efficiency.[44] SNA, of course, provides us with no simple answers but rather informs our analysis. It could also be argued that the reduced connectivity of the network could be viewed as an attempt to more efficiently use a smaller number of ties/edges, and some of the metrics here lend themselves to that interpretation.
The metrics presented in Table 3 suggest that connectivity probably improved after the transition of shot-callers. The network diameter declined from 3 to 2, indicating it was easier to get across the network since the network was smaller. The average clustering coefficient—defined as the average coefficient of every individual in the network—also went down from .860 to .271, suggesting that there was less clustering overall in the “after” network. The next section will discuss some of the effects of leadership change, such as the reduced number of meetings, and will analyze what may have caused this.
There are several questions after reviewing this indictment through an SNA lens. First, why did S1 become a victim of gang discipline after his removal, while S2 received no gang discipline before or after he replaced S1? One possible explanation for this result is that shot-callers are in charge of discipline, and, therefore, once a shot-caller has been removed, he is more likely to be subject to discipline by gang authorities and his own gang clique. Further, the sample data suggests that the new shot-caller was in his position for a limited amount of time and, therefore, isn’t likely to be disciplined by his own clique or higher-ranking members in this nascent phase. Thus, this result could be an artifact of the time frame for which data was available.
Second, why did S2 discipline fewer people after he became the new shot-caller? One possible explanation is that the new shot-caller was precarious in his new position. Thus, he did not want to “rock the boat,” so to speak, and engender a potential rebellion against his new authority. It should be noted that in the after network at time 2, there was less clustering in the overall network. This suggests that after S2 was promoted to the shot-caller position and as one was effectively sidelined and disciplined, though he remained a part of the network, the network was now more unified and had fewer or lower levels of subgroup density. This suggests that outside leadership might have good intelligence on the internal structures of the network. The lack of meetings in the after network could also be a sign that the network was avoiding government surveillance and or that fewer meetings were required to run the network, which now had less clustering. Of course, the reduced amount of gang discipline and gang meetings can be artificial due to the nature of the current data as for the purpose of prosecution using evidence collected by the law enforcement.
Third, why was there a reduction in meeting attendance once S1 was removed as a shot-caller for both S1 and S2? One possible explanation might be that law enforcement is saving evidence for another case. Another possible explanation is that law enforcement had limited capacity to wiretap or access informants with relevant information.
Limitations
It should be noted that this data was collected by law enforcement and given that it is primarily arrest warrant or indictment data, it may indeed involve missing ties, which could have implications for the findings here. Additionally, the substantiated change of the before-and-after networks on the metrics may be due to the small sample size of the current ego networks. Nonetheless, as previous research by Berlusconi (2013) has shown, it is indeed a purposive sample, and thus, key players and their ties are likely to be included (though key players may be overemphasized in this data type). Although it does not capture all the ties of the key players, it still gives us some important insight into how networks change after the replacement of shot-callers. Any data set can have potential missing data, and it is important for scholars to discuss issues of validity as we do here and in the methodology section. Future research could expand the amount of data, which in this analysis is limited to the Los Angeles area.
Avenues for Future Research
Further research should be conducted to establish whether our findings are consistent across new cases related to MS-13 or even compare them with other gangs other than MS-13. We tentatively hypothesize that MS-13 is becoming more hierarchical in nature and that higher levels of leadership at the regional level are playing greater roles, as evidenced here, but also that global leadership could increasingly become important. This includes leadership from El Salvador and the attempts to establish more direct control by RANFLA councils. More recent case studies, which go beyond the 2007 to 2017 period covered in this indictment would be helpful. The effects of Bukele’s crackdown on MS-13 and the potential human rights violations in El Salvador could also impact network structures and is a fruitful area for future research. The recent US foreign terrorist organization designation by the second Trump Administration will also likely lead to increased pressure on the network, which may have unexpected impacts upon its structure.
It would also help to build knowledge on the possible increasing levels of hierarchy within the MS-13 street gang based on expansive datasets. Indeed, the private sector is engaging in this. Companies, such as Evidencity, have collaborated with academics and are building broader MS-13 and other organized crime networks to better understand their structures.[45]
Conclusions
This article has indicated that MS-13 shot-callers’ ego network may become more hierarchical and may be more tightly coupled than previous academic literature has been willing to acknowledge. The analysis of the removal and the replacement of an MS-13 clique shot-caller by an outside regional council might be one indicator of that. Interestingly, the new SNA structure of the shot-callers’ ego network was arguably more efficient despite lower density and average weighted ties; e.g., the network’s diameter declined, suggesting more reachability in the network, allowing actors to access resources and communications across the network more easily. In short, the network became more accessible amongst actors after the shot-caller transition imposed by the outside regional council (though with internal input given some members of the clique in question were council members). We must acknowledge this is a limited data set with potentially missing ties and more research is needed to determine whether this is a consistent finding for this network type or this is anecdotal.
Appendix A: Data Run with Isolates Included
Table 4. Network Topography Metrics with Isolates included
Figure 3. Before Network (Isolates Included)
Figure 4. After Network (Isolates Included)
Endnotes
[1] It should be noted that clique refers to an MS-13 local group controlling a particular turf and is not here used in the context of a formal mathematical clique as is common in the SNA literature. See Robert A. Hanneman and Mark Riddle, Introduction to Social Network Methods, 2005, http://faculty.ucr.edu/~hanneman/nettext/.
[2] Again, here clique refers to an MS-13 local group controlling a particular turf and is not used in the context of a formal mathematical clique as is common in the SNA literature. For formal definitions of clique see: Op. Cit., Hanneman and Riddle at Note 1.
[3] Indictment United States of America v. Balmore et al., 2017, CR No. 17CR00278, United States District Court for the Central District of California, 10 May 2017.
[4] See, for example, Sonja Wolf, “Mara Salvatrucha: The Most Dangerous Street Gang in the Americas?” Latin American Politics and Society. Vol. 54, no. 1. 2012: pp. 65–99, https://doi.org/10.1111/j.1548-2456.2012.00143.x; Guadalupe Correa-Cabrera, Mariely López-Santana, and Camilo Pardo-Herrera, “The Nature of MS-13 in the American Context: A Transnational Comparison.” International Journal of Comparative and Applied Criminal Justice. Vol. 47, no. 1. 2021: pp. 79–97, https://doi.org/10.1080/01924036.2021.1949739; and Eric Hershberg, Edward Maguire, and Steven Dudley, “Assessing the Transnational Criminal Capacity of MS-13 in the U.S. and El Salvador.” Washington, DC: National Criminal Justice Reference Service. January 2019 (Document 252538), https://www.ojp.gov/pdffiles1/nij/grants/252538.pdf.
[5] Op. cit., Hanneman and Riddle at Note 1.
[6] Ana Arana, “How the Street Gangs Took Central America,” The New York Times. 7 June 2005, http://www.nytimes.com/cfr/international/20050501faessay84310_arana.html;
Steven Dudley, MS-13: The Making of America’s Most Notorious Gang. Toronto: Hanover Square Press, 2020.
[7] Linton C. Freeman, The Development of Social Network Analysis: A Study in the Sociology of Science. Vancouver: Empirical, 2004; Mark S Granovetter, “The Strength of Weak Ties,” American Journal of Sociology. Vol. 78, no. 6. 1973: pp. 1360–80.
[8] Sean F. Everton, Disrupting Dark Networks. New York: Cambridge University Press, 2012; Stephen P. Borgatti, Martin G. Everett, and Linton C. Freeman, “Ucinet for Windows: Software for Social Network Analysis,” 2002; Stephen P. Borgatti, Martin G. Everett, Jeffrey C. Johnson, and Filip Agneessens, Analyzing Social Networks Using R. Los Angeles: Sage, 2022; Mathieu Bastian, Sebastien Heymann, and Mathieu Jacomy, “Gephi: An Open Source Software for Exploring and Manipulating Networks.” Vol. 3, 2009, pp. 361–62; Craig Rawlings, Jeffrey A. Smith, James Moody, and Daniel McFarland, Network Analysis: Integrating Social Network Theory, Method, and Application with R. Cambridge: Cambridge University Press, 2023.
[9] Gisela Bichler, Understanding Criminal Networks: A Research Guide. Berkeley: University of California Press, 2019; Op. cit., Everton at Note 8; Daniel Cunningham, Sean Everton, and Philip Murphy, Understanding Dark Networks: A Strategic Framework for the Use of Social Network Analysis. London: Rowman & Littlefield, 2016; Valdis E. Krebs, “Mapping Networks of Terrorist Cells,” Connections. Vol. 24, no. 3. 2002: pp. 43–52.
[10] Jun Wu, William Layne Dittmann, Nathan P. Jones, and John P. Sullivan, “A Social Network Analysis of an MS-13 Network: Structure, Leadership Roles, and the Use of Confidential Informants.” International Criminology, 9 March 2024, https://doi.org/10.1007/s43576-024-00113-9.
[11] Pilar Glaser, John P. Sullivan, and Nathan P. Jones, “A Social Network Analysis of the Guerreros Unidos Crime Syndicate,” Small Wars Journal. 5 January 2024, https://smallwarsjournal.com/index.php/jrnl/art/social-network-analysis-guerreros-unidos-crime-syndicate; Francisco Sollano Jr, “A Social Network Analysis of Genaro García Luna and His Alleged Ties to the Sinaloa Cartel,” Small Wars Journal. 11 January 2022, https://smallwarsjournal.com/jrnl/art/social-network-analysis-genaro-garcia-luna-and-his-alleged-ties-sinaloa-cartel; Nathan P. Jones, W. Layne Dittmann, Jun Wu, and Tyler Reese, “A Mixed Methods Social Network Analysis of a Cross-Border Drug Network: The Fernando Sanchez Organization (FSO).” Trends in Organized Crime. Vol. 23, no. 2. 2020: pp. 154–82, https://doi.org/10.1007/s12117-018-9352-9; Carlo Morselli, “Structuring Mr. Nice: Entrepreneurial Opportunities and Brokerage Positioning in the Cannabis Trade.” Crime, Law and Social Change. Vol. 35, no. 3. 2001: pp. 203–44.
[12] David Bright, Russell Brewer, and Carlo Morselli, “Using Social Network Analysis to Study Crime: Navigating the Challenges of Criminal Justice Records.” Social Networks. Vol. 66, 2021: pp. 50–64; Gisela Bichler, Aili Malm, and Tristen Cooper, “Drug Supply Networks: A Systematic Review of the Organizational Structure of Illicit Drug Trade.” Crime Science. Vol. 6, no. 1. 2017.
[13] Giulia Berlusconi, “Do All the Pieces Matter? Assessing the Reliability of Law Enforcement Data Sources for the Network Analysis of Wire Taps.” Global Crime. Vol. 14, no. 1. 2014: pp. 61–81, https://doi.org/10.1080/17440572.2012.746940.
[14] David Skarbek, “Governance and Prison Gangs.” American Political Science Review. Vol. 105, no. 4. 2011: pp. 1–15; David Skarbek, “Prison Gangs, Norms, and Organizations.” Journal of Economic Behavior & Organization. Vol. 82, no. 1. 2012: pp. 96–109; John P. Sullivan, “Maras Morphing: Revisiting Third Generation Gangs,” Global Crime. Vol. 7, no. 3–4. 2006: pp. 487–504; Isaac Poritzky, Nathan P. Jones, and John P. Sullivan, “Transnational Cartels and Prison/Jail Gangs: A Social Network Analysis of Mexican Mafia (Eme) and La Familia Michoacana Conspiracy Cases.” Small Wars Journal. 24 October 2022, https://smallwarsjournal.com/index.php/jrnl/art/transnational-cartels-and-prisonjail-gangs-social-network-analysis-mexican-mafia-eme-and.
[15] David Bright, Catherine Greenhill, Michael Reynolds, Alison Ritter, and Carlo Morselli, “The Use of Actor-Level Attributes and Centrality Measures to Identify Key Actors: A Case Study of an Australian Drug Trafficking Network.” Journal of Contemporary Criminal Justice. Vol. 31, no. 3. 2015: pp. 262–78.
[16] Vincent D. Blondel, Jean-Loup Guillaume, Renaud Lambiotte, and Etienne Lefebvre “Fast Unfolding of Communities in Large Networks.” Journal of Statistical Mechanics: Theory and Experiment. 2008, no. 10. 9 October 2008: P10008, https://doi.org/10.1088/1742-5468/2008/10/P10008.
[17] See Robert J. Bunker and John P. Sullivan, “Mara Salvatrucha (MS-13): A Law Enforcement Primer. FBI National Academy Associate, March/April 2018, https://www.academia.edu/36413311/Mara_Salvatrucha_MS_13_A_Law_Enforcement_Primer; John P. Sullivan, “Transnational gangs: The Impact of Third Generation Gangs in Central America.” Air & Space Power Journal (Spanish Edition). Second Trimester 2008, https://www.academia.edu/927368/Transnational_gangs_The_impact_of_third_generation_gangs_in_Central_America; John P. Sullivan and Robert J. Bunker, Eds., Strategic Notes on Third Generation Gangs. Bloomington: Xlibris, 2020.
[18] Steven Dudley and Héctor Silva Ávalos, “MS13: Hierarchy vs. Federation. InSight Crime. 14 February 2018, https://www.amazon.com/dp/1796095613/ref=sr_1_1?dchild=1&keywords=Strategic%20Notes%20on%20Third%20Generation%20Gangs&qid=1585934361&s=books&sr=1-1.
[19] On links to state corruption, see Douglas Farah and Marianne Richardson, “Gangs No Longer: Reassessing Transnational Armed Groups in the Western Hemisphere,” Strategic Perspectives. Vol. 38, INSS (Institute for National Strategic Studies), National Defense University. May 2022, https://inss.ndu.edu/Portals/68/Documents/stratperspective/inss/Strategic-Perspectives-38.pdf; Douglas Farah and Ksathryn Babineau, “The Evolution of MS13 in El Salvador and Honduras, PRISM. Vol.7, no.1 2017, pp. 58–73, https://www.jstor.org/stable/26470498.
[20] “MS13,” InSight Crime. 28 February 2024, https://insightcrime.org/el-salvador-organized-crime-news/mara-salvatrucha-ms-13-profile/. An example of “control from afar” is provided in Parker Asmann, “US Indictment Suggests MS13 Expanding Communications, Drug Trafficking.” InSight Crime. 12 January 2018, https://insightcrime.org/news/brief/indictment-ms13-us-operations/.
[21] Robert J. Bunker and John P. Sullivan, “Third Generation Gangs Strategic Note No. 13: Mara Salvatrucha (MS-13) Command and Control (C2) Geographic Variations.” Small Wars Journal. 29 January 2019, https://smallwarsjournal.com/jrnl/art/third-generation-gangs-strategic-note-no-13-mara-salvatrucha-ms-13-command-and-control-c2.
[22] John P. Sullivan and Samuel Logan, “MS-13 Leadership: Networks of Influence.” The CounterTerrorist. August/September 2010, https://www.academia.edu/1123626/MS_13_Leadership_Networks_of_Influence.
[23] John P. Sullivan and Robert J. Bunker, “Third Generation Gangs Strategic Note No. 11: MS-503—Mara Fragmentation and Murder,” Small Wars Journal, 10 April 2018, https://www.academia.edu/36402974/Third_Generation_Gangs_Strategic_Note_No_11_MS_503_Mara_Fragmentation_and_Murder.
[24] Indictment United States of America v. Trejo Chavarria et al. (2022), No. Case 3:22-cr-00018 (2022).
[25] Ibid.
[26] Op. Cit., Wolf; Correa-Cabrera, López-Santana, and Pardo-Herrera; Hershberg, Maguire, and Steven Dudley at Note 4.
[27] Michael McGuiness, “Prosecution, Deportation, and the MS-13: Recent Administrations’ Unwise Approach to Combatting Gang Violence. Virginia Journal of Criminal Law. Vol. 9, no. 1. 2020: pp. 1–32, https://heinonline.org/HOL/LandingPage?handle=hein.journals/virjcr9&div=3&id=&page=.
[28] Héctor Silva Ávalos, “MS13 Members Imprisoned in El Salvador Can Direct the Gang in the US.” InSight Crime. 5 December 2016, https://insightcrime.org/investigations/ms13-members-imprisoned-in-el-salvador-can-direct-the-gang-in-the-us/.
[29] Lidia E. Nuño and Edward R. Maguire, “The Nature and Structure of MS-13 in Los Angeles County.” Criminal Justice Review. Vol. 48, no.1, 2023: pp. 5–20, https://journals.sagepub.com/doi/pdf/10.1177/07340168211029990. They suggest structure in at p. 12, while questioning transnational links at p. 13. Nuño and Maguire’s work makes it clear that continued empirical research on the nature and structure of MS-13 is warranted.
[30] See Note 6 in José Miguel Cruz, Jonathan D. Rosen, and Yemile Mizrahi, “The Long Arm of the Gang: Disengagement under Gang Governance in Central America.” Criminology. Vol. 61, no. 4. November 2023, https://doi.org/10.1111/1745-9125.12352.
[31] Ibid.
[32] Benjamin Lessing, “Logics of Violence in Criminal War,” Journal of Conflict Resolution. Vol. 59, no. 8. 2015: pp. 1486–1516, https://doi.org/10.1177/0022002715587100; E.D. Arias, “The Dynamics of Criminal Governance: Networks and Social Order in Rio de Janeiro,” Journal of Latin American Studies. Vol. 38, no. 2. 2006: pp. 293–325; John P. Sullivan, “From Drug Wars to Criminal Insurgency: Mexican Cartels, Criminal Enclaves and Criminal Insurgency in Mexico and Central America Implications for Global Security.” Working Paper N09. Paris: Fondation Maison des Sciences de l’homme, April 2012, https://halshs.archives-ouvertes.fr/halshs-00694083/document; Anne L. Clunan and Harold A. Trinkunas, Eds. Ungoverned Spaces: Alternatives to State Authority in an Era of Softened Sovereignty. Stanford: Stanford University Press, 2010; Phil Williams, “Here Be Dragons: Dangerous Spaces and International Security,” in Ungoverned Spaces: Alternatives to State Authority in an Era of Softened Sovereignty, Anne L. Clunan and Harold A. Trinkunas, Eds., Stanford: Stanford University Press, 2010.
[33] Op. Cit., Cruz, Rosen, and Mizrahi, at Note 30 page 23.
[34] José Miguel Cruz and Jonathan D Rosen, “Leaving the Pervasive Barrio: Gang Disengagement under Criminal Governance,” Social Problems. Vol. 71, no. 1. 2024: pp. 254–70.
[35] David C. Pyrooz and Scott H. Decker, “Motives and Methods for Leaving the Gang: Understanding the Process of Gang Desistance,” Journal of Criminal Justice. Vol. 39, no. 5. 2011: pp. 417–25, https://doi.org/10.1016/j.jcrimjus.2011.07.001.
[36] Op. Cit., Jones, Dittmann, Wu, and Reese at Note 11.
[37] Op. Cit., Borgatti, Everett, and Freeman, “Ucinet for Windows: Software for Social Network Analysis” at Note 8.
[38] Linton C Freeman, “Centrality in Social Networks Conceptual Clarification,” Social Networks 1, no. 3. 1978: pp.215–39; Op. Cit., Everton at Note 8.
[39] Op. Cit., Bastian, Heymann, and Jacomy at Note 8.
[40] Op. Cit., Indictment United States of America v. Balmore (2017) at Note 3.
[41] Yifan Hu, “Efficient, High-Quality Force-Directed Graph Drawing,” The Mathematica Journal. Vol. 10, no. 1. 2006: pp. 37–71.
[42] For a discussion of Average path length and definitions see: Op. Cit. Everton at Note 8 and Op. Cit. Cunningham et. al. at Note 9.
[43] Karl E. Weick, “The Collapse of Sensemaking in Organizations: The Mann Gulch Disaster,” Administrative Science Quarterly. Vol. 38, no. 4. 1993: pp. 628–52, https://doi.org/10.2307/2393339.
[44] Carlo Morselli, Cynthia Giguère, and Katia Petit, “The Efficiency/Security Trade-off in Criminal Networks,” Social Networks. Vol. 29, no. 1. 2007: pp. 143–53, https://doi.org/10.1016/j.socnet.2006.05.001.
[45] “The MS-13 Illicit Network Index,” Evidencity, 2024, https://www.evidencity.com/ms-13.
