July 12, 2007

Lead Poisoning and the Great 1960s Freakout:

Lead Poisoning and the Great 1960s Freakout: Rick Nevin, an economist not affiliated with a university, has gotten some publicity attributing the rise and fall of crime rates and the rise of illegitimacy to lead poisoning, since lead is known to reduce IQ and impulse control in children exposed to it. This is a theory with rather far-reaching implications, such as offering an explanation of the Great Freakout that began in America around the time of the Kennedy Assassination and the Beatles appearing on the Ed Sullivan Show and tapered off with surprising rapidity in the mid-1990s. The Washington Post claims:

"This study shows a very strong association between preschool blood lead and subsequent crime rate trends over several decades in the USA, Britain, Canada, France, Australia, Finland, Italy, West Germany, and New Zealand. The relationship is....consistent with neurobehavioral damage in the first year of life and the peak age of offending for index crime, burglary, and violent crime....Regression analysis of average 1985-1994 murder rates across USA cities suggests that murder could be especially associated with more severe cases of childhood lead poisoning."

This is a very interesting theory.

On the other hand, Japan, which had lots of lead-spewing cars and extremely dense population, and thus should have been a prime victim of lead poisoning, never experienced a Great Freakout. So, that's one strike against the theory.

In general, I'm skeptical of there being a "Single Bullet" explanation for the the 1960s, but it's worth looking at this in some detail, more than I can muster here, but I'll take a first look at it.

More generally, as Steven D. Levitt pointed out on his Freakonomics blog:

"Still, although both Post reporter Shankar Vendantam and the cited economist, Rick Nevin (whom I’d never heard of), appear quite convinced by the time-series data, I am not. When you have a variable like crime that goes up for a long time then goes down for a long time, it is easy to find other variables that share that pattern and appear to have a causal impact, even though the relationship is completely spurious."

Yet, Levitt goes on:

"About seven years ago, Michael Greenstone and I tried to look into this same issue using airborne lead measures at the local level, as well as other approaches. We ultimately gave up without finding anything. That largely soured me on the lead/crime link."

"Recently, however, Jessica Wolpaw Reyes at Amherst has put together what appears to me to be the most persuasive evidence to date in favor of a relationship between lead and crime. Rather than looking at a national time-series, she tries to exploit differences in the rates at which lead was removed from gasoline across states. I haven’t read her paper with the care that a referee would at an academic journal; but, at least superficially, what she is doing looks pretty sensible. She finds that lead has big effects (and, for what it’s worth, she also confirms that, when controlling for lead, the link between abortion and crime is as strong or stronger as in our initial study, which did not control for lead.)"

It's a long paper (70 pages) and a lot of good work went into it. After a quick read, though, I would venture to say that it has a couple of flaws. The first is that while it finds strong relationships among states over time between lead consumed in gasoline and overall published violent crime rates, it doesn't find much of a relationship between lead and murder. Dr. Wolpaw Reyes rationalizes:

"The weak murder results could also stem from the rarity of murders (rendering identification more difficult), a weaker effect of lead on murder than on other violent behavior, or a different functional form for this relationship (such as an increasing marginal effect). Given that murder is the most violent of violent crimes, it is not unreasonable to hypothesize that only substantial exposure to lead will produce this extremely aggressive and violent behavior, while more moderate exposure will have more moderate effects."

Uh ... Nah, I'm not at all persuaded that lead would influence people to be more impulsively violent but not more lethally impulsively violent. I don't see that as plausible.

The problem for how much confidence to place on this paper is that murder is the most accurately measured crime. Other crimes vary dramatically in likelihood of being reported to the police across time and place, but attention must be paid to a dead body with a hole in it.

Second, the big increase in the murder rate was from roughly 1965-1975. She assumes a 22 year lag, from postnatal exposure to lead, so that was from births in 1943 to 1955, but she says that lead poisoning due to leaded gasoline didn't peak until 1970, so how come the murder rate didn't keep going up? It bounced around a lot after that, but the Great American Freakout was basically from the mid-60s to the mid-70s. .
Here's her graph of "Gasoline Lead Exposure 1960 to 1990" (p. 66), which shows that lead from gasoline peaked in 1970. She writes: "Gasoline lead exposure rose until 1970 and then fell."

Third, when looking at the crime fall in the 1990s, this study appears to have the same flaw that dragged down Levitt's abortion-cut-crime theory -- a failure to look carefully at crime rates by age cohort, combined with an intoxication with analyzing complex state-level data that leads to a failure to do simple national-level reality checks. (She was clearly influenced by Levitt, so this shortcoming is not surprising.
) Wolpaw Reyes simply assumes a 22 year lag between lead poisoning around the time of birth and the violent crime rate. But, we can easily look at more detailed data for different age cohorts, which shows that the crime decline of the 1990s began among older individuals, not among the younger people supposedly benefiting from lower lead or higher abortion.

If you assume a 22 year lag to violent crime, then this graph looks great because the murder rate started to fall after about 1991.

But, that's the same mistake Levitt made way back in 1999: he forgot to look at the crime rates among narrower age cohorts. For the 17-and-under crowd, the two worst years were 1993-94. In other words, they were born when lead pollution had already fallen by almost half (just as th
ey were born when legal abortion was close to its peak, which is a problem for Levitt's theory).

Here's the relevant homicide rate graph showing that for the 14-17 year-old cohort, the murder rate moved in exactly the wrong direction for the first 6-7 years of the great lead decline.

Similarly, here's the non-lethal serious violent crime rate for 12-17 year-olds as reported by the government's massive annual survey of crime victims. It too shows the crime rate for the relevant cohort moving in the wrong direction.

Now, this hardly disproves the lead-crime theory, it just subjects it to an obvious reality check, which demonstrates a lack of care over looking at the precise age of criminals. Now, it could well be that the crack wars overwhelmed the lead effect at the national scale. Or, it could be that lead has an impact, but it's strongest at an older age, like 6 or 7. Or, her graph showing a lead peak in 1970 is inaccurate (other sources suggest a slightly later peak).On a side note, Levitt also writes:

"Roger Masters, a professor at Dartmouth, has also been doing interesting research on this subject, although I am also not very familiar with his work."

Besides discussing lead and crime in Massachusetts' towns, Masters writes:

"2. Communities using either fluosilicic acid (H2SiF6) or sodium silicofluoride (NaSiF6) have significantly higher rates of crime than those using sodium fluoride or delivering unfluoridated water (with the exception of towns with naturally fluoridated water).

"3. The use of fluosilicic acid (H2SiF6) to fluoridate public water supplies significantly increases the amounts of lead in the water (whereas the use of sodium silicofluoride (NaSiF6) or sodium fluoride (NaF) does not."

In other words, if Dr. Masters is not mistaken, Col. Jack D. Ripper in "Dr. Strangelove" was right: fluoridation is poisoning our precious bodily fluids.

I returned to this topic in a 2013 Taki's Magazine article: "Did Heavy Metal Poisoning Cause the Sixties?" In it, I offer suggestions on how to test this question.

A simple suggestion for further research is that America's Environmental Protection Agency keeps a list of SuperFund CleanUp sites polluted by lead. Typically, these are locations with lead smelters, mines, factories, or dumps. It would be a fair amount of work, but hardly impossible to compare crime rates in small towns with names like Smelterville, Leadville, Galena, and so forth to similar towns with no point sources of lead pollution.

I've read newspaper accounts of the lawsuit against the smelter firm in a small town in Missouri. The plaintiffs argued that lead pollution was damaging the children of the town -- yet, the plaintiffs didn't seem to emphasize high crime rates. Instead, the poisoned children were described as unenergetic, sluggish, and so forth. 

Anyway, there is a lot of information available on the web from these kind of lawsuits, so I'd recommend anybody interested in the topic just read up on it.
(Wandering even farther afield, Jack D. Ripper was played, brilliantly, by Sterling Hayden, the stepfather of American Conservative editor Scott McConnell.)


Anonymous said...

There's too much concern here with changes in crime rates, and not enough with the actual levels themselves.

Anonymous said...

It isn't lead. It's idealism. Goofy unworkable ideals - for instance, "flat-earthism" on race - bobbed up in the wake of our WWII victory, and we had the wherewithal to put them into practice. One trillion on social programs, anyone?

I would trace the crime wave of the 1960s-1990s to the Civil Right Act of 1964 and other "Great Society" legislation; more deeply, the underlying enthusiastic willful ignorance (aka idealism). It's reflected like the sun in the pieces of a shattered lightbulb, all over the culture: more lenient criminal law philosophies, Boazian anthro come to fruition, the whole set of attitudes labeled "liberalism" label. There is a physical, racial explanation of these poisonous ideas, but I find K. MacDonald a little farfetched (though not QUITE as farfetched as Jack D. Ripper's theories).

Anonymous said...

Does Wolpaw Reyes only look at the lead levels at birth? What happens if you try to estimate the total amount of lead
people were exposed to while growing up (and also as fetuses)?

Anonymous said...

Note to self: When sniffing gasoline,make sure its unleaded!

Anonymous said...

Steve --

Two things stand out.

1. "Freak outs" are not unique to say 1965-75. For example 18th Century London was even worse, with Gin consumption rampant, most people drunk out of their minds, due to medicating themselves to stand deplorable conditions. The solution was the Victorian promotion of "virtue" which entailed fixing the worst of the conditions and pushing beer and pubs in "moderation" to the gin joints.

2. Your comments on the lack of a great social gulf in Heinlein's America between Cops and Doctors.

Taking these two into consideration, perhaps the explanation for the freak out was:

A. Social gulf developed as the elite grew very rich, moved to exclusive neighborhoods. Effect most pronounced with the end of Jim Crow segregation and Black elites moving out from say South Central to Baldwin Hills. But still in effect in say Culver City where elites moved to say, Santa Monica or Bel Air.

B. As a result of elites no longer being invested personally in owning property, public safety and Victorian-style "public morality" was let go as elites didn't care: they didn't live in those areas.

C. As in the case of Karen Toshima in LA, social decay pushed outward from poor/working-class areas to threaten elites property values in West LA or the East-side. Hence Mayor Rudy in NYC or various anti-gang/crime initiatives in LA. Which stabilized the crime rates.

I'm sure there are other factors involved. But this might be a partial explanation.

Anonymous said...

Another question: did/does lead poisoning help drive the "demographic transition?"

Some historians have speculated that lead poisoning (from lead-alloy tableware, lead water pipes, leaded wine, etc.) explains both the low fertility and the apparent goofiness of upper-class Ancient Romans.

My day job really doesn't permit me to do any research on this right now, but from at least the Roman era through the Renaissance upper- (chivalric-) class folk were famously violent and impetuous--and they tended to eat off metal tableware alloyed with lead, wear and handle objects decorated with leaded pigments, etc.

There is still debate around the proposition that the medieval rich were more fecund, in part because it's hard to disentangle all the influences on multi-generational fertility. (For example, the Black Death in the 1300's killed a smaller proportion of rich than poor--so afterwards the rich made up a larger fraction of the population, but without necessarily producing more babies.)

From the early industrial period until quite recently the story of economic progress was also, for many people, the story of increased lead exposure.

Consider the example which attracted Mr. Sailer--leaded gasoline fumes, which from 1930-ish to as late as 2000 was effectively a proxy for automobility, which in turn was a proxy for industrialization.

Perhaps the demographic transition would have been less dramatic if lead poisoning had not degraded fertility (either directly, or by making potential mates more quarrelsome).

Of course, we're back to the correlation/ causation problem here...

Anonymous said...

I think this is why people try to provide anecdotal evidence when confronted with charts and graphs. Low IQ isn't descriptive enough. Do people suffering from lead poisoning end up with pervasive developmental disorder or do they resemble someone who has frontal lobe damage from being in a car wreck?

Anyone who has spent time around the mentally retarded (and they are very nice people to be around, generally very pleasant) knows that even if on the off chance they built up a murderous rage, they're often too fragile or uncoordinated to cause much harm before being stopped.

I think this is a case of garbage in/garbage out. Yes the researchers have beautiful coefficients and are referring to to prior statistical analyses that were considered valid but they're calculating abstractions of abstractions.

Math brains!

Anonymous said...

Japan has a phenomenal abortion rate, for several reasons, one being the unavailability of birth control besides condoms for many years, and the lack of Christian anti-abortion howling.

What have the effects of that been on the Japanese?

Anonymous said...

I thought most childhood lead poisoning was caused by eating old paint chips that had flaked off old buildings. Never heard of kids huffing gasoline.

Tom T. said...

Reyes notes, "This means that two children who are otherwise identical but whose lead levels differ by 15 μg/dL (approximately the decline in lead levels between 1976 and 1990) would exhibit an average IQ difference of 7.5 points."

Has such an increase in IQ been noted over that timespan?

Tom T. said...

Sorry; followed a link here and didn't see how old this post was. Disregard me.