In this insight, networks scientist Alberto Nieto explores the potentially game changing power of social network analysis to understand the formation of criminal groups, and how it is being used at UCL to obtain one of the most extensive insights into the criminal underworld of Bogotá, Colombia.
Law enforcement agencies are subject to increasing pressure to respond to complex crimes. Think about scams. Scammers scattered worldwide target people across several countries, with the United States and the United Kingdom being some of the hardest hit. Some of these scams require locals’ help to collect cash from victims and send it abroad. Once the police intervene in these cases, the money has travelled miles away through the financial system, leaving victims with few opportunities, if any, to recover their cash.
Collaboration between offenders, as this scenario shows, is essential for criminality1. Without the participation of multiple individuals, some crimes become harder or impossible to execute2. In the above example, without locals’ help, scammers would need to resort to electronic transactions to access victims’ money, increasing detection risks.
How can law enforcement agencies exploit the knowledge about the collaborative aspect of crime to prevent it in the first place? Understanding the environment in which crime occurs, in particular the related situational factors, can provide valuable insights to the police and other stakeholders for preventing crime.
In the instance of scams, the US and UK authorities might be interested in knowing how people located in different countries can find and trust associates in targeted countries. This process of accomplice selection has been subject to multiple theories (see van Mastrigt (2017)3 for a detailed description of the theories proposed so far). Some describe it as a rational decision reached through evaluating risks, benefits and the potential accomplice’s trustworthiness4. Other theories describe accomplice selection as a function of circumstance5. People meet other motivated offenders in their social contexts, and they co-offend when opportunities arise6. Unfortunately, these theories derive from observations that are limited in scientific validity. Many case studies used to derive them are based on accounts from teenagers participating in illicit activities. Hence, the question about how adult offenders select their associates for more sophisticated, global and complex crimes remains comparatively unanswered.
We need more insights to understand how adults choose accomplices, so that preventive interventions can be designed effectively. In this regard, a ‘networks approach’ to crime offers a promising alternative to study accomplice-selection processes. A network is a simplified representation of the interactions between entities comprising a system7; in this case, a network would involve individuals participating in illegal activities. By aggregating information about who commits a crime with whom, it is possible to extract co-offending networks that show how people connect to others through joint, criminal activities.
Such networks reveal unique patterns of interactions among offenders. For example, it is possible to identify individuals with numerous accomplices or uncover close-knit groups. By including information about when offenders create new co-offending relationships, it is also possible to understand how these networks grow and detect the individuals capable of attracting new accomplices.
Following this approach, an ongoing research project found more than 4,000 offenders connected through multiple shared criminal events in Bogotá (Colombia) between 2005-2018 (see figure below)8. The number of people in this network represents almost 41 per cent of the prison population as of 2018, according to figures provided by the National Institute for Penitentiaries and Prisons (INPEC). Such a sizeable network raises questions about how law enforcement agencies can stop this network from growing.
Crime researchers can step on network scientists’ shoulders and use the findings and tools developed in this field to gain a better understanding of co-offending networks. Networks science has abundant evidence on the mechanisms that predict how social networks grow (Barabási (2016)9 reviews some of the mechanisms that explain how networks evolve). For example, individuals are more likely to make new connections with the neighbours of their neighbours. Likewise, people tend to associate more with those sharing similar social characteristics (e.g., age, sex, and ethnicity) than with dissimilar others. Some tools predict new connections using features like, for example, the network-related distance between people.
Unfortunately, the field of studies of co-offending networks is small, with a limited number of reports about how adult offenders interact (Carrington (2002); Brantingham, Ester, Frank, Glässer, and Tayebi (2011); Bright, Whelan, and Morselli (2020)10 are few examples of studies that have looked into adult co-offenders). As more crime data become available, we will have more information about the properties of co-offending networks. More insights into offenders’ network-related characteristics that better predict new crimes will also become accessible. These insights, in turn, can help law enforcement agencies disrupt transnational fraud schemes, including the examples described above, and prevent other types of crime that rely on collaborative partnerships.
In short, a network approach to crime can help authorities change their policy to tackle complex crimes. Shifting from a reactive to a preventive approach allows the police to meet society’s demands.
References: [read more]
[1] Tremblay, P. (1993). Searching for suitable co-offenders. Routine activity and rational choice, 5, 17–36.
[2] Weerman, F. M. (2003). Co-offending as social exchange. explaining characteristics of co-offending. British journal of criminology, 43(2), 398–416.
[3] van Mastrigt, S. B. (2017). Co-offending and co-offender selection. The Oxford handbook of offender decision making, 6, 338.
[4] McCarthy, B., & Hagan, J. (2001). When crime pays: Capital, competence, and criminal success. Social forces, 79(3), 1035–1060.; McCarthy, B., Hagan, J., & Cohen, L. E. (1998). Uncertainty, cooperation, and crime: Understanding the decision to co-offend. Social forces , 77 (1), 155–184.
[5] Alarid, L. F., Burton Jr, V. S., & Hochstetler, A. L. (2009). Group and solo robberies: Do accomplices shape criminal form? Journal of Criminal Justice, 37(1), 1-9.; Reiss, A. J. (1988). Co-offending and criminal careers. Crime and Justice, 10, 117-170. doi: 10.1086/449145
[6] Reiss.; Sharp, C., Aldridge, J., & Medina, J. (2006). Delinquent youth groups and 5 offending behaviour: findings from the 2004 offending, crime and justice survey (Vol. 14) (No. 06). Home Office London.
[7] Newman, M. (2018). Networks. Oxford University Press.
[8] Nieto, A., Borrion, H., & Davies, T. (2021, June). Examining the importance of existing relationships for co-offending: a temporal network analysis in bogota, colombia (2005-2018) [in progress].
[9] Barabási, A.-L. (2016). Network science. Cambridge University Press.
[10] Brantingham, P. L., Ester, M., Frank, R., Glässer, U., & Tayebi, M. A. (2011). Co-offending network mining. In Counterterrorism and open source intelligence (pp. 73–102). Springer.; Bright, D., Whelan, C., & Morselli, C. (2020). Understanding the structure and composition of co-offending networks in Australia. Trends & Issues in Crime & Criminal Justice (597).; Carrington, P. J. (2002). Group crime in Canada. Canadian J. Criminology, 44, 277.[/read]
Insight by
