Vanquishing Zombie Theories of the Crime Decline
Step 1. Accept that there are not two kinds of people in the world...
A couple of weeks ago I wrote about the Great American Crime Decline and the vast and varied speculation about the cause. I proposed there were several causes, perhaps as many as two dozen credible explanations. Many people disagreed, offering instead their own unified theory for the crime decline. Several responses took the form of, “I thought this was settled—it’s lead.” There was also a smattering of, “I thought this was settled, it’s abortion.” In less progressive circles, a different single explanation would likely have surfaced—mass incarceration.
It’s tempting to just dismiss these takes by pointing out that few complex social problems have a simple answer or to respond with an empirical argument. But on closer inspection, there is something more important going on. These answers are troubling for a larger reason that is perhaps not so obvious. In looking over the hundreds of serious academic papers and books explaining the crime decline it does seem to me that only these three explanations have connected with the general public. These three narrow causal mechanisms tend to define the boundaries of the popular narrative and they fit neatly into political boxes—there is a progressive answer (lead), a conservative answer (prison), and a counter-intuitive data science answer (abortion). But they share a single latent foundation. And that’s where the trouble can be found.
Each of these crime decline mechanisms presupposes that some people are criminal and some people are not, at least to an important extent. They presuppose that there is something innate and irredeemable about people who commit crimes. That criminality is binary—people are either criminals or good, law-abiding people. You can see this dichotomy in their preferred policy responses. What society really needs to do to reduce crime and violence is to keep people from being poisoned and becoming criminals (lead), or, to keep criminals from being born (abortion), or, to lock away the criminals (mass incarceration). Regardless of the politics that prefer one approach over another, these worldviews share a lack of nuance about criminal offending.
This binary view of criminal behavior closely follows what is known as the medico-psychological model of crime. This model of criminal behavior, borrowing from Cornish and Clarke (2008), essentially says the following, which is a paraphrase:
Criminal behavior is primarily the result of long-standing criminal predispositions and psychopathologies that cause individuals to offend. Research should be invested in programmes to prevent the development of criminality, which can be changed through appropriate treatments, and, these changes once made would then persist.
The lead theory says you are poisoned and develop criminal psychopathology, the abortion theory says you are born with it and the prison story goes even further, essentially arguing that regardless of how you come to your criminal psychopathology nothing can be done and criminals should be locked away since they can’t be fixed.
In thinking about what caused the crime decline—and how lessons from the crime decline can be used to implement a more effective crime policy today—it is absolutely critical to free ourselves from the medico-psychological model of crime. No one has done more to dispel this model than Cornish and Clarke who separately and together have spent a long career arguing that the causes of crime are multi-faceted and inter-related. They argue, and it is now widely accepted, that criminal offending is primarily a result of the current environment, far more than any “emotional inheritance.” Current mood, opportunity, and the example of others are far more predictive of criminal behavior than any underlying ‘criminal propensity’.
Somehow though, the medico-psychological, binary, you are a criminal or not a criminal model lives on in popular consciousness. It is a zombie theory. This, by the way, is an underappreciated reason why there is such widespread acceptance of the notion that rehabilitation doesn’t work. In 1974, a researcher—Martinson—famously noted that “nothing works” in treating people in correctional settings. And for two generations researchers have fought against this notion because there are loads of evidence-based programs now. But, this ignores the historical context of the nothing works statement since the programs Martison judged to be failures are not the same programs found to be effective today. ‘Nothing works’—because the kind of rehabilitation in favor during the 1960s and 1970s was dedicated to changing the predispositions and psychopathologies of offenders, which we now know are not real problems. The major problem—environments that privilege the choice to commit a crime—were intentionally not addressed in these treatments. So, of course, the treatments didn’t work.
So, the problem with the argument that the removal of lead in gasoline or mass incarceration or abortion caused the crime decline is that the argument is built upon a false dichotomy, of criminal or not criminal. As Cornish and Clarke note, a better theory,
[S]ees the desires, preferences and motives of offenders and potential offenders as similar to those of the rest of us – in continual interaction with immediate opportunities and constraints to produce, reinforce and sometimes reduce both criminal and non-criminal behaviours.
This Rational Choice theory sees people in the act of criminal offending as purposeful and rational. It is supported by oceans of data and is critical to understanding the crime decline. So too is the idea of Prospect Theory that says that purpose and rationality may lead to objectively poor decisions when you account for the very human tendencies of all of us to overvalue some risks and some rewards, particularly rare events, and to act without full information.
That doesn’t mean lead, abortion, and prison do not help to explain the crime decline, they do (well, lead and mass incarceration do). But they are sledgehammers on an anvil. They are the equivalent to looking at Singapore, which has almost no violent crime and ignoring the substantial limits on civil liberties that accompany that safety. It’s a massive trade-off and there is a much smarter path.
Categories of Crime Decline Mechanisms
Armed with the knowledge that environments matter, peer behaviors matter, and most people who commit crimes are acting rationally—as they understand it—most of the time, then it is relatively easy to group the 25 crime decline explanations into broad categories. The value in doing so is plain—once grouped, it is relatively easy to answer two critical questions:
Is there a policy lever that can directly affect the category? And,
Does that policy lever have anything to do with the crime policy?
The four bins I propose as crime decline categories are:
Routine Activities/Rational Choice
I have already described routine activities theory, it is the Cornish and Clarke theory I described above as the counter to the medico-psychological binary view of the world. Within rational choice is a very useful theory known as routine activities which describe the components of a criminal event, generally framed as a motivated offender, a suitable target, and a lack of guardianship. There is obviously a lot to say about this, but the central idea is that more guardianship or fewer motivated offenders or fewer suitable targets will reduce the likelihood of a crime. Big reductions in one of these components or a combination can lead to big reductions in crime. Each is more than worthy of its own discussion, but for now, to generalize:
In-home entertainment (reduces suitable targets)
Air conditioning (reduces suitable targets)
Less cash in circulation (reduces suitable targets)
Obesity/ Disability (reduces suitable targets/motivated offenders)
Crime Goes Online (reduces motivated offenders for street crime)
Abortion (reduces motivated offenders)
The answer to the two questions above the availability of policy levers and the role of crime policy is no and no. Short of a dictatorship, no policy lever would cause big changes in routine activities for a large number of people across the country. And, none of this has anything to do with criminal justice policy. These changes represent shifts in how we live day-to-day as we make choices in response to new goods and services.
Prevention is defined as the removal or reduction in risk conditions or an increase in assets. Here I’m applying a narrow understanding of prevention, which does not target a specific at-risk group, but rather reflects messages, policies, and practices targeted at everyone.
Declines in teen pregnancy (removal of a risk condition)
Decreased alcohol consumption in 16/17 year-olds (removal of a risk condition)
Removal of lead from gasoline (removal of a risk condition)
Security/Surveillance (increase in an asset)
More community-based organizations serving at-risk neighborhoods (increase in an asset)
Increased average educational attainment (increase in an asset)
Increased use of therapeutics in mental health (removes risk condition)
What is obvious from this list is that none of the population-level prevention efforts specifically targeted crime and violence. This is a critical point—that effective prevention targets general risk conditions, where that risk includes criminogenic factors but also includes risks for all sorts of other poor outcomes.
Large economic forces obviously have critical implications for local environments.
Price Stability (lower inflation, more complete information)
Economic Stability/Consumer Sentiment (less economic insecurity)
Immigration (neighborhood stabilization)
Gentrification (neighborhood investment)
Population-level socioeconomics (older, wealthier population)
Most of these are self-explanatory. The US is simply an older, wealthier place than it was in the 1980s. The odd explanation is that we seem to have a much better understanding of monetary policy than in the 1970s and 1980s. Yes, there are still sudden deep recessions, but the gnawing uncertainty signaled by volatility in inflation is no longer present. Again turning to Prospect Theory, there is some evidence that loss aversion makes people particularly susceptible to the risks of inflation, particularly when inflation affects consumer staples. It could be argued, therefore, that because the principle effect of sound monetary policy is to reduce anxiety, that these macroeconomic mechanisms are just an alternative means of population-level prevention. There is, of course, another more straightforward reason why less inflation means less crime—as inflation increases the cost of new goods, demand increases for used (or stolen) goods… But either way, none of these policy levers are undertaken with a goal of reducing crime and violence.
Criminal justice activities are slightly harder to categorize, but it is enlightening to think of these effects as being in a routine activities bucket (more guardianship or fewer motivated offenders) or a deterrence bucket (increased fear of committing a crime, increased risk of detection).
Mass incarceration (fewer motivated offenders)
Problem-oriented policing (increased risk of detection)
Increase in alternatives to incarceration/detention (prevention, fewer motivated offenders)
Better policing practice through technology (increased risk of detection)
Better policing strategy/operations (increased fear of committing crime)
Crime prevention through environmental design (prevention, more guardianship)
Some, like alternatives to incarceration with treatment, attempt to move effective prevention practices into the criminal justice system. Putting aside mass incarceration, the limitation of the criminal justice explanations is that they are mainly theoretical. Problem-oriented policing is hard to define and harder to measure so it is not clear how much policing falls into this category. Alternatives to incarceration, until relatively recently had such strict eligibility criteria that few people could access these programs. Similarly, it is not clear that improvements in technology lead to the identification of unknown suspects or whether they are more commonly used to induce pleas in known offenders. There is undoubtedly some effect from these changes in policy and practice, but the scale of the effect is simply not known.
The only explanation for the crime decline that defies categorization is the end of the crack epidemic. My read on this is that there was no great causal mechanism that caused the epidemic to end. Yes, mass incarceration was likely a precipitator of the end, but mostly the data suggest that the end was organic—there were few new cocaine initiates by the early 1990s, and without replacements for the people who went into recovery, or prison, crack use turned from an epidemic to a lower prevalence, but still too common, drug of abuse. Did young people just say no? Did they see too much carnage in the previous generation? Were they too afraid of prison? We don’t know. What is clear is that none of the categories above was the primary explanation for the crime drop.
Ranking Crime Decline Explanations
As a research project, the next exercise would be to assign a coefficient to each of these terms defining its relative contribution to the overall crime decline. Then, perhaps, to do a cost-benefit analysis and assign each category a benefit-to-cost ratio. Perhaps, something like that would be helpful.
I doubt it though. I think a much more important insight is that broad-based policies that serve the public interest through the prevention policies that are focused on the removal of risk conditions are the primary mechanisms for reducing crime and violence. Risk conditions, particularly those associated with poverty, are the primary mechanism for crime and violence. Removing these burdens and barriers is ultimately the most effective weapon against crime.
I leave you with a prediction. Perhaps in twenty years, as I enter my dotage, I will be motivated to say something about the great crime decline of the 2030s. While the short-term prospects for crime and violence in America look relatively grim, the longer-term prospects are better. The 2030s are much more likely to be safer than the 2020s if for no other reason than the fact that advances in technology continue to rapidly outstrip advances in privacy. Surely, these technologies will make crime easier to detect, though the cost will be high.
But regardless of the evolution of technology, one mechanism with the potential to create massive reductions in crime and violence has recently been launched, and at scale. There is every reason to believe that very high on the list of reasons for a crime decline in the 2030s will be the 2021 American Rescue Plan, which experts expect to cut childhood poverty by up to half. It is hard to imagine anything happening between now and then—at least anything that simultaneously preserves our civil liberties—that will be equally consequential.
 I know you watched Dexter and you are reading this and saying, “Yeah? What about Dexter?” I know it’s hard to let go…
 My favorite take on the ‘there are two kinds of people” worldview was the comedian who said, “there are two kinds of people in the world—the kind that thinks there are two kinds of people in the world, and the kind that don’t.”