Crime as a Cascade Phenomenon

Cascades of Violence deploys data from across South Asia to conclude that war tends to cascade across space and time.

Professor John Braithwaite is a Distinguished Professor and Founder of RegNet (the Regulatory Institutions Network) at the Australian National University. He was awarded the BSC Outstanding Achievement Award in 2017 and is an Honorary member of the Society.

Braithwaite and D’Costa’s (2018) Cascades of Violence can be downloaded for free here

This post sketches why it could be analytically fertile to view crime as a cascade phenomenon. Once we see crime through the cascade lens, we can imagine how to more effectively cascade crime prevention. Like crime, crime prevention often cascades. Braithwaite and D’Costa show how peacemaking cascades nonviolence. Happily, there is strong evidence to support the conclusion that nonviolence is also a cascade phenomenon. Hence, seeing crime through the cascade lens opens up fertile ways of imagining a macrocriminology of crime control. Self-efficacy and collective efficacy are hypothesized as catalysts of crime prevention cascades in the macrocriminology that interests me.

Cascade phenomena are defined as those that spread to multiply instances of themselves, or to create contagions of related phenomena. Cascade explanations are staples across the physical and biological sciences: the cascading of particles in particle physics, cascading of particular particles called bacteria and viruses with infectious diseases, environmental cascades to climate change, cascading of liquids (lava, water) in the geological formation of planets. In the social sciences, cascade explanations have also been common in the writing of Rosenau, Schelling, Sunstein , Kuran, Sikkink and Gladwell, among others. With crime, we have long known that people are more likely to cheat on their taxes if they perceive a lot of cheating among others and that contagion effects are particularly likely with high profile crimes such as hijackings, assassinations, kidnappings, suicide bombing and spates of serial killing.

Non-criminologists have been more fascinated by cascade possibilities than criminologists. Mathematician Quetelet in 1835 was puzzled by the high statistical variance in crime across space and time. Economists often puzzle further that this variance is so huge compared to variables that are seen as candidates for explaining variation. This leads to the hypothesis that cascading on itself might provide a better explanation than exogenous changes in rational incentives driven by costs and benefits of crime. They point out that interactions among people could cascade to explain the variance. If one crack cocaine dealer interacts with five others to persuade them that becoming a dealer is smart, and each of them so persuades five others, and so on, then this dynamic can multiply huge space-time variance between a point in space-time where that process takes off and others where there has been no cascade.

Information cascades where people make decisions on the basis of their observations of other peoples’ actions seem particularly attractive for explaining why criminal behaviors like looting or rioting are normally near zero, but can multiply quickly once someone starts a stampede. Herding into illegal tax shelters is likewise an information cascade phenomenon according to my 2005 book, Markets in Vice, Markets in Virtue. Braithwaite and D’Costa note that more common kinds of crime also behave like wars in this regard, as they sought integrated explanation of crime-war clusters. They point out that the best explanation of whether your house will be burgled in the next six months in many countries can be whether it was burgled in the last six months; and likewise, the best explanation of whether your country will suffer a war this year may be whether it suffered another in the past three years. Whether the house next door was burgled or the country next door convulsed by war are also good predictors.

When Lawrence Sherman and other criminologists found that crime was concentrated at three per cent of the addresses of large cities and that policing strategies concentrated at those hot spots could substantially reduce crime at them, the natural reaction of criminologists was cynical. Our cynicism was directed at the hypothesis that criminals will respond by shifting their crime from old hot spots to nearby locales, or to create new hot spots.  Subsequent research did not bear out this displacement hypothesis.  Indeed, it showed not only that hot spot policing reduced crime at the hot spot, but it also had positive spillovers in reducing crime to lesser degrees in areas surrounding hot spots. Why did not criminologists then proceed with a sense of excitement at the surprise of having their expectations reversed? Why not explore and develop a converse theory that there may be cascade effects of crime prevention success?  Criminologists tend not to respond to overturned cynicism with excitement at the opportunity to build theory on new inductive insights, preferring to move on to cynicism about something else.

Reframing crime as a cascade phenomenon implies a shift from focus on individual offenders to building a new macrocriminology. Such a reframed macrocriminology is my current work-in-progress. Braithwaite and D’Costa’s study of cascades of violence across South Asia was a considerable empirical undertaking that could, perhaps, be submitted as a proof of concept, though no more than that. The conclusions of that book about war are undoubtedly more important than those about crime, particularly in showing what can be done with the insight that the best way of protecting ourselves from future wars is to stop getting into current ones. Yet a neglected reason for the importance of that policy work is that war and crime cascade into each other so profoundly.

My suspicion is that the cascade lens could illuminate a good framework for the kind of macrocriminological reframing that can make a fist of big patterns in the evolution of crime such as why western societies have dramatically less violent crime than they had centuries ago; why so many Latin American societies have so much more criminal violence than other regions; why East Asian societies have experienced dramatic reductions in violence for half a century or more; why in the same period the United States has had a higher crime rate than other Western societies. Mainstream criminology devotes remarkably little attention to such macro patterns compared to the attention mainstream economics devotes to why certain spaces and times have superior growth, or mainstream political science to why some spaces and times are less democratic, more authoritarian.

How could a framework like control theory be seen by many criminologists as one of the most empirically supported of all theories without confronting it with macro questions such as whether it really makes sense to say that the United States has so much more crime than Canada, Europe, Australia or Japan because Americans are less able to control their impulses? My proposal is that conceiving crime as a cascade phenomenon is one possibility for a better path to reconfiguration of criminological theory.


Professor John Braithwaite, School of Regulation and Global Governance (RegNet), Australian National University.



Copyright free image: from author.