When Rudy Giuliani ran for mayor of New York City in 1993, he campaigned on a platform of bringing down crime and making the city safe again. It was a comfortable position for a former federal prosecutor with a tough-guy image, but it was more than mere posturing. Since 1960, rape rates had nearly quadrupled, murder had quintupled, and robbery had grown fourteenfold. New Yorkers felt like they lived in a city under siege.
Throughout the campaign, Giuliani embraced a theory of crime fighting called "broken windows," popularized a decade earlier by James Q. Wilson and George L. Kelling in an influential article in The Atlantic.
"If a window in a building is broken and is left unrepaired," they observed, "all the rest of the windows will soon be broken." So too, tolerance of small crimes would create a vicious cycle ending with entire neighborhoods turning into war zones. But if you cracked down on small crimes, bigger crimes would drop as well.
Giuliani won the election, and he made good on his crime-fighting promises by selecting Boston police chief Bill Bratton as the NYPD's new commissioner. Bratton had made his reputation as head of the New York City Transit Police, where he aggressively applied broken-windows policing to turnstile jumpers and vagrants in subway stations. With Giuliani's eager support, he began applying the same lessons to the entire city, going after panhandlers, drunks, drug pushers, and the city's hated squeegee men. And more: He decentralized police operations and gave precinct commanders more control, keeping them accountable with a pioneering system called CompStat that tracked crime hot spots in real time.
The results were dramatic. In 1996, the New York Times reported that crime had plunged for the third straight year, the sharpest drop since the end of Prohibition. Since 1993, rape rates had dropped 17 percent, assault 27 percent, robbery 42 percent, and murder an astonishing 49 percent. Giuliani was on his way to becoming America's Mayor and Bratton was on the cover of Time. It was a remarkable public policy victory.
But even more remarkable is what happened next. Shortly after Bratton's star turn, political scientist John DiIulio warned that the echo of the baby boom would soon produce a demographic bulge of millions of young males that he famously dubbed "juvenile super-predators." Other criminologists nodded along. But even though the demographic bulge came right on schedule, crime continued to drop. And drop. And drop. By 2010, violent crime rates in New York City had plunged 75 percent from their peak in the early '90s.
All in all, it seemed to be a story with a happy ending, a triumph for Wilson and Kelling's theory and Giuliani and Bratton's practice. And yet, doubts remained. For one thing, violent crime actually peaked in New York City in 1990, four years before the Giuliani-Bratton era. By the time they took office, it had already dropped 12 percent.
The PB Effect
What happens when you expose a generation of kids to high lead levels? Crime and teen pregnancy data two decades later tell a startling story.
Second, and far more puzzling, it's not just New York that has seen a big drop in crime. In city after city, violent crime peaked in the early '90s and then began a steady and spectacular decline. Washington, DC, didn't have either Giuliani or Bratton, but its violent crime rate has dropped 58 percent since its peak. Dallas' has fallen 70 percent. Newark: 74 percent. Los Angeles: 78 percent.
There must be more going on here than just a change in policing tactics in one city. But what?
There are, it turns out, plenty of theories. When I started research for this story, I worked my way through a pair of thick criminology tomes. One chapter regaled me with the "exciting possibility" that it's mostly a matter of economics: Crime goes down when the economy is booming and goes up when it's in a slump. Unfortunately, the theory doesn't seem to hold water—for example, crime rates have continued to drop recently despite our prolonged downturn.
Another chapter suggested that crime drops in big cities were mostly a reflection of the crack epidemic of the '80s finally burning itself out. A trio of authors identified three major "drug eras" in New York City, the first dominated by heroin, which produced limited violence, and the second by crack, which generated spectacular levels of it. In the early '90s, these researchers proposed, the children of CrackGen switched to marijuana, choosing a less violent and more law-abiding lifestyle. As they did, crime rates in New York and other cities went down.
Another chapter told a story of demographics: As the number of young men increases, so does crime. Unfortunately for this theory, the number of young men increased during the '90s, but crime dropped anyway.
There were chapters in my tomes on the effect of prison expansion. On guns and gun control. On family. On race. On parole and probation. On the raw number of police officers. It seemed as if everyone had a pet theory. In 1999, economist Steven Levitt, later famous as the coauthor of Freakonomics, teamed up with John Donohue to suggest that crime dropped because of Roe v. Wade; legalized abortion, they argued, led to fewer unwanted babies, which meant fewer maladjusted and violent young men two decades later.
But there's a problem common to all of these theories: It's hard to tease out actual proof. Maybe the end of the crack epidemic contributed to a decline in inner-city crime, but then again, maybe it was really the effect of increased incarceration, more cops on the beat, broken-windows policing, and a rise in abortion rates 20 years earlier. After all, they all happened at the same time.
To address this problem, the field of econometrics gives researchers an enormous toolbox of sophisticated statistical techniques. But, notes statistician and conservative commentator Jim Manzi in his recent book Uncontrolled, econometrics consistently fails to explain most of the variation in crime rates. After reviewing 122 known field tests, Manzi found that only 20 percent demonstrated positive results for specific crime-fighting strategies, and none of those positive results were replicated in follow-up studies.
Did Lead Make You Dumber?
Even low levels have a significant effect.
So we're back to square one. More prisons might help control crime, more cops might help, and better policing might help. But the evidence is thin for any of these as the main cause. What are we missing?
Experts often suggest that crime resembles an epidemic. But what kind? Karl Smith, a professor of public economics and government at the University of North Carolina-Chapel Hill, has a good rule of thumb for categorizing epidemics: If it spreads along lines of communication, he says, the cause is information. Think Bieber Fever. If it travels along major transportation routes, the cause is microbial. Think influenza. If it spreads out like a fan, the cause is an insect. Think malaria. But if it's everywhere, all at once—as both the rise of crime in the '60s and '70s and the fall of crime in the '90s seemed to be—the cause is a molecule.
A molecule? That sounds crazy. What molecule could be responsible for a steep and sudden decline in violent crime?
Well, here's one possibility: Pb(CH2CH3)4.
In 1994, Rick Nevin was a consultant working for the US Department of Housing and Urban Development on the costs and benefits of removing lead paint from old houses. This has been a topic of intense study because of the growing body of research linking lead exposure in small children with a whole raft of complications later in life, including lower IQ, hyperactivity, behavioral problems, and learning disabilities.
But as Nevin was working on that assignment, his client suggested they might be missing something. A recent study had suggested a link between childhood lead exposure and juvenile delinquency later on. Maybe reducing lead exposure had an effect on violent crime too?
That tip took Nevin in a different direction. The biggest source of lead in the postwar era, it turns out, wasn't paint. It was leaded gasoline. And if you chart the rise and fall of atmospheric lead caused by the rise and fall of leaded gasoline consumption, you get a pretty simple upside-down U: Lead emissions from tailpipes rose steadily from the early '40s through the early '70s, nearly quadrupling over that period. Then, as unleaded gasoline began to replace leaded gasoline, emissions plummeted.
Gasoline lead may explain as much as 90 percent of the rise and fall of violent crime over the past half century.
Intriguingly, violent crime rates followed the same upside-down U pattern. The only thing different was the time period: Crime rates rose dramatically in the '60s through the '80s, and then began dropping steadily starting in the early '90s. The two curves looked eerily identical, but were offset by about 20 years.
So Nevin dove in further, digging up detailed data on lead emissions and crime rates to see if the similarity of the curves was as good as it seemed. It turned out to be even better: In a 2000 paper (PDF) he concluded that if you add a lag time of 23 years, lead emissions from automobiles explain 90 percent of the variation in violent crime in America. Toddlers who ingested high levels of lead in the '40s and '50s really were more likely to become violent criminals in the '60s, '70s, and '80s.
And with that we have our molecule: tetraethyl lead, the gasoline additive invented by General Motors in the 1920s to prevent knocking and pinging in high-performance engines. As auto sales boomed after World War II, and drivers in powerful new cars increasingly asked service station attendants to "fill 'er up with ethyl," they were unwittingly creating a crime wave two decades later.
It was an exciting conjecture, and it prompted an immediate wave of…nothing. Nevin's paper was almost completely ignored, and in one sense it's easy to see why—Nevin is an economist, not a criminologist, and his paper was published in Environmental Research, not a journal with a big readership in the criminology community. What's more, a single correlation between two curves isn't all that impressive, econometrically speaking. Sales of vinyl LPs rose in the postwar period too, and then declined in the '80s and '90s. Lots of things follow a pattern like that. So no matter how good the fit, if you only have a single correlation it might just be a coincidence. You need to do something more to establish causality.
As it turns out, however, a few hundred miles north someone was doing just that. In the late '90s, Jessica Wolpaw Reyes was a graduate student at Harvard casting around for a dissertation topic that eventually became a study she published in 2007 as a public health policy professor at Amherst. "I learned about lead because I was pregnant and living in old housing in Harvard Square," she told me, and after attending a talk where future Freakonomics star Levitt outlined his abortion/crime theory, she started thinking about lead and crime. Although the association seemed plausible, she wanted to find out whether increased lead exposure caused increases in crime. But how?
In states where consumption of leaded gasoline declined slowly, crime declined slowly. Where it declined quickly, crime declined quickly.
The answer, it turned out, involved "several months of cold calling" to find lead emissions data at the state level. During the '70s and '80s, the introduction of the catalytic converter, combined with increasingly stringent Environmental Protection Agency rules, steadily reduced the amount of leaded gasoline used in America, but Reyes discovered that this reduction wasn't uniform. In fact, use of leaded gasoline varied widely among states, and this gave Reyes the opening she needed. If childhood lead exposure really did produce criminal behavior in adults, you'd expect that in states where consumption of leaded gasoline declined slowly, crime would decline slowly too. Conversely, in states where it declined quickly, crime would decline quickly. And that's exactly what she found.
Meanwhile, Nevin had kept busy as well, and in 2007 he published a new paper looking at crime trends around the world (PDF). This way, he could make sure the close match he'd found between the lead curve and the crime curve wasn't just a coincidence. Sure, maybe the real culprit in the United States was something else happening at the exact same time, but what are the odds of that same something happening at several different times in several different countries?
Nevin collected lead data and crime data for Australia and found a close match. Ditto for Canada. And Great Britain and Finland and France and Italy and New Zealand and West Germany. Every time, the two curves fit each other astonishingly well. When I spoke to Nevin about this, I asked him if he had ever found a country that didn't fit the theory. "No," he replied. "Not one."
Just this year, Tulane University researcher Howard Mielke published a paper with demographer Sammy Zahran on the correlation of lead and crime at the city level. They studied six US cities that had both good crime data and good lead data going back to the '50s, and they found a good fit in every single one. In fact, Mielke has even studied lead concentrations at the neighborhood level in New Orleans and shared his maps with the local police. "When they overlay them with crime maps," he told me, "they realize they match up."
Location, Location, Location
In New Orleans, lead levels can vary dramatically from one neighborhood to the next—and the poorest neighborhoods tend to be the worst hit.
Maps by Karen Minot
Put all this together and you have an astonishing body of evidence. We now have studies at the international level, the national level, the state level, the city level, and even the individual level. Groups of children have been followed from the womb to adulthood, and higher childhood blood lead levels are consistently associated with higher adult arrest rates for violent crimes. All of these studies tell the same story: Gasoline lead is responsible for a good share of the rise and fall of violent crime over the past half century.
When differences of atmospheric lead density between big and small cities largely went away, so did the difference in murder rates.
Like many good theories, the gasoline lead hypothesis helps explain some things we might not have realized even needed explaining. For example, murder rates have always been higher in big cities than in towns and small cities. We're so used to this that it seems unsurprising, but Nevin points out that it might actually have a surprising explanation—because big cities have lots of cars in a small area, they also had high densities of atmospheric lead during the postwar era. But as lead levels in gasoline decreased, the differences between big and small cities largely went away. And guess what? The difference in murder rates went away too. Today, homicide rates are similar in cities of all sizes. It may be that violent crime isn't an inevitable consequence of being a big city after all.
The gasoline lead story has another virtue too: It's the only hypothesis that persuasively explains both the rise of crime in the '60s and '70s and its fall beginning in the '90s. Two other theories—the baby boom demographic bulge and the drug explosion of the '60s—at least have the potential to explain both, but neither one fully fits the known data. Only gasoline lead, with its dramatic rise and fall following World War II, can explain the equally dramatic rise and fall in violent crime.