« Weekend Bleg/Open Thread | Main | Benchmarks »

July 08, 2007

Comments

When I read that, I thought of something I remember, which is that Japan apparently has high levels of environmental lead, but very low levels within the population because of several strongly instilled cultural habits (bathing patterns, removal of shoes before entering an inside space)

Can this also be linked to the rise of the Republican Party?

Seriously, I'd love to know if this is for real. Fascinating stuff.

Puts an even more somber aspect on the Washington DC problem with lead in the public water supply.

Which started to be greatly reduced once it was brought out into the open. Will look for more recent follow-up...

It didn't sink in until I followed the link that the story is in the Washington Post, which makes it all the more bizarre that there's not even a nod to the DC water story, which was a huge featured series in 2004 in the Post.

I've thought for years that lead abatement would do so much more than hundreds of other things we spent money on, and this only reinforces that thought.

"Sixty-five to ninety percent or more of the substantial variation in violent crime in all these countries was explained by lead."


That seems impossibly high. Also this "x to y or more" formulation sounds a bit more like a salesman than a scientist, esp. since I guess "substantial" means he concentrated on the tails of the distribution.


Is there a table out there of factors by which one can estimate the benefit to the economy of spending a dollar on a project of type p? I would imagine that much of that $30B would end up making jobs thus reducing unemployment insurance etc. or otherwise directly reducing budgeted govt expenditures, even ignoring the undoubted benefit of lead reduction.

rilkefan: I'm guessing that they're just describing the results of a regression analysis, in which case it would be correlation, not necessarily causation, and there would be room for other confounding factors to play a role.

That said, they're saying that 65% of the variance in the crime rate, not 65% of crimes or something, is explained by variation in lead levels. That seems much less odd to me.

If true, or even partly true, raise a toast to Claire Patterson, who fought hard to get tetraethyl lead out of gasoline. And curse you, Thomas Midgley!

If true, or even partly true, raise a toast to Claire Patterson, who fought hard to get tetraethyl lead out of gasoline. And curse you, Thomas Midgley!

Looks to me from that article that he's arguing for causation, not correlation - that essentially nothing else matters.

"This is the reason you are seeing the crime rate soar in Mexico and Latin America, but [it] has fallen in the United States." [emphasis added]

That is, every society has a preset level of crime which no (so-far tested) policy can affect - except varying lead exposure.

That sounds crazy to me. I can well imagine reducing lead is an important policy - perhaps the single most important achievable policy among a variety of significant approaches - but 90% causation?

"one of the main drivers of differences in crime rates is environmental lead levels"

Right - this is what I would find unsurprising, but not what the Post interview indicates.

So whatever else there might be that affects crime rates, it isn't *changing* much. It may be different in different places but it hasn't changed. Lead levels have changed and the effects of those changes matches up.

It isn't real plausible that some other factor causes higher lead levels and also higher violent crime rates 20 years later -- though I can't rule that out.

Well, there was the Freakonomics argument about abortion. And there have been huge societal changes in this country - that's all irrelevant?

I haven't read the study yet -- it's very early in the morning for me.

From what Hilzoy reports, he argues that the correlations with the paint fit the data in 9 countries over a century. Other correlations fit sometimes and some places.

You can get spurious correlations, but then he has a theory that seems to fit too. That makes it more plausible than it would be if he coudn't find any JustSo story that fit.

Well, there was the Freakonomics argument about abortion.

That's an example of a hypothesis that can easily be tested by comparing different countries. This study seems to debunk the abortion connection pretty conclusively.

Looks to me from that article that he's arguing for causation, not correlation - that essentially nothing else matters.

I'd just voice a note of caution in attributing arguments that the researcher made with the way they are presented by the journalist.

Well, I'd also add that, IIRC, the Freakonomics argument had other countries showing a correlation as well, and said that abortion was only one of four different causes that caused crime to drop so dramatically in the 90's. I don't know about Britain and Europe, as quoted in the article, but it seems odd to me that such a big, obvious hole in the theory would never even have been mentioned in anything else discussing the idea before. That, combined with his apparently overzealous claims about lead's importance, would lead me to want more information before deciding anything. It is an interesting idea.

It is a real possibility, though, that the guy's quotes and ideas are just being mangled by the journalist. Happens far more often than you'd think in science reporting (I'm sure in all reporting, really, but I happen to know of more examples in science than anything else)

Hey buddy! Nice blog that you maintain here.. I just chanced upon your blog surfing the blogosphere. I was thinking.. you could try out some interesting widgets on your page and spice it up with more relevant information. E.g try out the new widget on http://www.widgetmate.com with your relevant keywords

I’m not sure what type of program you are envisioning so I can’t really say if I’d be for it or against it.

The government has already required that lead paint discussion be part of every real estate transaction since 1992. The buyer/leaser must be given an EPA information pamphlet, disclosure is required, buyer/leaser must be given the opportunity to test, etc.

So the government has been involved in this issue – at least to the point where buyers/leasers must be fully informed of the dangers and the risk inherent in the specific home they are considering. At that point I think that some personal responsibility is required.

As to a new government program of some sort, I’d generally be in favor of making testing and abatement services available free to those who can’t afford it themselves.

Two other points:

I seriously question that $32 billion figure. At around $100 for a professional test, you could barely cover all the testing needed, much less any serious abatement. And those are 1992 dollars.

Second – I’m a little wary due to what’s happened with the asbestos issue.

I am reminded of the eloquent suggestion that lead poisoning was an important contribution to the fall of the Roman Empire.

Lead was used extensively in Roman plumbing, including coupling those famous aqueducts that furnished water from mountain sources. Imagine the lead levels of the drinking water. They also "discovered" that if wine went sour it could be made sweeter (more flavorful--Yikes!) by storing it in vessels made of lead!

(We, being more modern, prefer to inhale toxins -- tobacco, grass or a cocktail of industrial emissions -- or ingest them in the form of refined prescription drugs.)

I am quite sceptic about such monocausal reasoning, especially concerning lead. While there is no doubt (apart from lead lobbyists hired by EPA under Bush) that lead is bad for health, to blame "all" on it looks to me about as reasonable as the old claim that the downfall of the Roman empire was caused primarily by the fashionable lead glass decanters used by the nobility. Additionally chronic lead poisoning causes fatigue and apathy, not the usual traits of violent criminals.
Lead (and mercury and arsenic etc.) may be a contributing factor but I would consider it very difficult, if not completely impossible, to isolate its effects from those of other massive changes in the natural and cultural environment.
Anyway, I don't think that studies like these are actually necessary to justify getting noxious heavy metals and other unhealthy compounds out of the environment. They may even prove harmful because if they turn out to be bogus, they will be used as argument against doing anything (like the above mentioned lobbyist).

I'd just voice a note of caution in attributing arguments that the researcher made with the way they are presented by the journalist.

Yes. Or several notes, a chord maybe.

And it's not just DC:

Lead Levels in Water Misrepresented Across U.S. Utilities Manipulate or Withhold Test Results to Ward Off Regulators

By Carol D. Leonnig, Jo Becker and David Nakamura
Washington Post Staff Writers
Tuesday, October 5, 2004; Page A01

Cities across the country are manipulating the results of tests used to detect lead in water, violating federal law and putting millions of Americans at risk of drinking more of the contaminant than their suppliers are reporting.

@OCSteve: A lot of the problem with lead paint is in rental housing, where, where tenants have the lead exposure and the (absentee) landlords have the 'personal responsibility'.

OCSteve: also, I'm not sure why it ought to be the responsibility of property owners, most of whom did not contribute to the problem in the first place, and, as Nell points out, will not always be the ones to benefit from it. If they had knowingly poisoned their properties, I'd be in favor of holding them accountable; as it is, it's not so clear to me why they should be left holding the bag.

By "a serious program" I mean one that funds lead removal and/or abatement for either anyone or anyone who can't afford it, using a definition of "can't afford it" that's not too strict. A lot of property owners in inner cities are not themselves wealthy, and putting roadblocks in their way often just results in property sitting abandoned, further harming neighborhoods, other people's property values, etc.

If we would all benefit from this, then imho we should step up to the plate. Think of it this way: it's a lot less than the cost of the misbegotten surge, which somehow we found it in our hearts to pay for.

There is some merit in OCSteve's wariness about scope and costs.

The scope is vast--essentially every structure that was standing before about 1970 or so has lead paint in it or on it. The cost could be immense--you can't just scrape it off with a putty knife. Instead, you first have to *encase the whole thing in a huge vacuum chamber with HEPA filters attached*. Otherwise, you're actually doing more harm than good by putting lead dust in the air.

So imagine every structure having this sort of treatment, tents of plastic around it, special vacuum-trucks pulled up along-side, armies of guys in moon-suits and respirators, the whole nine yards. Every building and structure in the country that is over forty years old.

I just don't see this happening for 32 billion.

Would it cost as much as Bush's War? Well, since estimates on that start at around a trillion and go up, clearly it would not. If Bush had not lied us into war, we could now do lots of things we cannot do.

But the money is gone now.

I'm not saying I oppose the proposal--I'd love to see lead levels in kids reduced. Maybe there are other ways to monitor and treat the kids directly? Alas, chelation has never worked yet.

But notice that there is a fine line between an old-style, responsible conservative reaction (can we afford this? are we underestimating costs? are there unintended consequences? could we get the same results some other way?), and a new-style, irresponsible right-wing reaction, the Norquist Government Derangement Syndrome reaction of "all government programs are bad! find any excuse to oppose them! feign concerns about competence when what really drives you is ideology!"

If it's the first, then we have a technical disagreement, and we can work to get better data, do better modeling and prediction, design the system as well as possible, and keep costs proportionate to benefits, all the while agreeing about shared goals.

If it's the second, we're just looking at deranged government-haters, and they would oppose even the most efficient plan if it made government look good.

Does the article discount independent correlations between lead paint and poverty, and poverty and violent crime? That seems like the obvious connection to me.

dougie: I would think that since it's trying to explain the variability in crime rates, it would: generally, poverty levels and levels of lead do not move in lockstep, since lead levels involve things like banning leaded gasoline, and, in one case cited by the article, a program to help people put in safer replacement windows that people would have a harder time falling out of, which had the unintended side benefit of removing the lead paint that was on the replaced windows.

If poverty levels remained substantially the same, and lead levels rose or fell substantially, then looking at the crime levels would let you see which of the two was doing more work.

This is a stunner. Not that lead causes crime, but that it's so substantial. With something as complex as intercountry crime rates, if lead really has a 65% r-square the causation is unquestionable - you just can't get that level of correlation with spurious variables. In general causation could run the other way, or there could be a shared cause, but in this case the time lags and the biology finger lead as the cause, not the effect or a signal. I wish the article showed the actual data - it's a travesty this is behind Elsevier's paywall. Here's the journal link, for those with library access: http://www.sciencedirect.com/science/journal/00139351

The cost could be immense--you can't just scrape it off with a putty knife. Instead, you first have to *encase the whole thing in a huge vacuum chamber with HEPA filters attached*. Otherwise, you're actually doing more harm than good by putting lead dust in the air.

Wouldn't that be worth careful study?

Suppose we could remove 90% of the lead at reasonable cost. Would that be worth doing?

Suppose we could do without the HEPA filters and we did make a moderate amount of lead dust in the air. Would that be too horrible to consider? As it is, we're talking about a chronic problem that causes subtle effects. A short-term pulse of lead dust might have undetectable effects. It's worth some study.

It gets in our way to think that we can't make any improvement unless we make it perfect and do it perfectly.

I'd be happy to email the article to anyone who wants to go over the data. (hbok at mac dot com.)

Nell: A lot of the problem with lead paint is in rental housing, where, where tenants have the lead exposure and the (absentee) landlords have the 'personal responsibility'.

Hilzoy: I'm not sure why it ought to be the responsibility of property owners, most of whom did not contribute to the problem in the first place

I think that a line can be drawn there. My concern is ‘slumlords’ getting a potentially very expensive abatement on the taxpayer’s dime under at least some circumstances. If a landlord purchased an apartment building post 1992 then we know that they are aware of the hazard. If they did nothing to address the issue in the years since then they have been knowingly exposing their tenants to the risk. I’d support a variety of options depending on the circumstances. Possibly something like this: If the rental units were purchased prior to 92, then a grant for abatement (based on need – the tenants are primarily low income etc.); if purchased after 92, then something not quite a grant – maybe a low cost loan. I’d also support changes in state laws that say if you want to rent a unit you have to prove it is lead free or that it has been properly abated. Certainly any landlord who wants to work with HUD could be forced to take care of it.

At the individual homeowner level, anyone can request and receive a lead test kit free of charge (no need to show need). These “swab kits” run $20 or less. If the do it yourself test is positive, then they can apply for a more detailed professional test and risk assessment (free based on need) and depending on those results a grant for abatement (again free based on need).

As a final note – I don’t even care about this particular study to justify any of that: the risk is bad enough on its own merits. Lead poisoning is awful and no child should be exposed to the danger. I’m all for addressing it at a national level, but I’m wary of it turning out the same way asbestos concerns did.

In any case the removal of the lead is only the first step. It must be made sure that the replacement is not actually as bad. Often metal is replaced with organic compounds that can have serious bad side effects. PVC tubes (that was for a time the standard replacement for lead tubes over here) can e.g. exude plasticisers and residual vinyl chloride (both quite unhealthy), similar effects are known with paint etc.
Again, I doubt monocausality on principle*, especially with something as complex as societal development(s).

*if there is not lots of corroborating evidence.

This is a stunner. Not that lead causes crime, but that it's so substantial.

Remember that he's looking at changes. Any other causes that didn't change much wouldn't show up at all. But he's implying that the various things that did change -- abortion, police methods, etc -- probably didn't have a lot of effect. But there are lots of things that might affect impulsivity -- like other heavy metals -- that wouldn't show up because we haven't first sprayed the whole country with them and then quit.

In 2001, sociologist Paul B. Stretesky and criminologist Michael Lynch showed that U.S. counties with high lead levels had four times the murder rate of counties with low lead levels, after controlling for multiple environmental and socioeconomic factors.

After controlling for multiple factors. This looks like a solid effect, but Nevin's claim is that it accounts for much of the variance. He hasn't directly claimed how much of the incidence it accounts for.


So, say this gets accepted. How will it get spun? Here's one idea -- mexico and points south still have a lot of lead. So it's predictable that immigrants will be more impulsive and do more violent crime.

According to this (see the second page) it costs about $10,000 to delead an apartment, so $32 billion does seem low.

Still, the benefits are potentially massive.There are a number of studies on crime costs. One summary is here(PDF)

A 1999 study puts the annual cost at $1.7 trillion.

Even with smaller numbers, generally in the range of $500 billion to $1 trillion, it's important to remember that these are annual costs being compared to one-time deleading costs. A 5% reduction in a $500 billion cost generates $25 billion a year. Even at a very high 10% discount rate this justifies a $250 billion expenditure.

And of course these numbers don't include the costs of mental retardation or other problems.

Some of the other costs that could potentially be saved:

1) Special education costs.
2) Psychological treatment costs, which considering that many of the people affected may come from low-inconme areas, would normally be borne by the states and federal governments.
3) Incarceration savings, plus costs of trials, etc.

It is easy to look at what the potential cost of the clean-up might be, but this is frequently done without taking into consideration the overall savings which can be, as hilzoy pointed out, far above the iniial expenditures.

OCSteve, I don't think your solution is that far away from anything that would be acceptable to us darn lefties. In fact, I don't really have any problem with what you propose at all.

I am sure Halliburton could receive a no-bid contract to handle it.

OCSteve: which problem with asbestos are you worried about? (Just curious.)

"Even with smaller numbers, generally in the range of $500 billion to $1 trillion, it's important to remember that these are annual costs being compared to one-time deleading costs."

This is an incredibly important point. Lead's effects continue across decades. A one time expenditure of even the $300-$700 billion range, spread out over a 5-10 year implementation period, would be very cost effective.

My only caution would be that you would have to be careful to make it a very targeted implementation system. I don't see any reason for whatever agency we use to have much of a budget 10 years from the beginning of the project. Deal with the problem and shut it down.

lj: "I'd just voice a note of caution in attributing arguments that the researcher made with the way they are presented by the journalist."

Well, that's why I quoted the researcher - "X is the reason for Y". Of course people get inaccurate trying to explain technical results to laypeople and reporters, but that's some serious sloppiness if so. Anyway, will try to peek at the article.

This is an interesting (but older) companion article. It appears that Nevin originally began this work under contract from HUD some years ago. There is a lot more meat here than in the new WaPo article.

The costs continue to pile up. In a study in the July 2002 issue of Environmental Health Perspectives, a team led by Landrigan estimates the total annual health care costs of lead poisoning: a staggering $43.4 billion a year, or about 2.2% of total U.S. health care costs. The costs include direct treatment costs and also such collateral damage as special education and reduced lifetime learning capacity. "It's a national scandal," Landrigan says.

HUD says “lead paint hazards lurk in about 25 million” homes. If we accept the $10,000 remediation figure then call it a ballpark $250 billion one time cost. That would be more than recouped in about one generation. And that is just for those directly affected and does not consider any costs associated with crime.

Looking at the article in detail, I think the 65% claim is overstated, although I'd still support aggressive measures. The problem is that a great deal of the changes in both lead exposure and crime are long-term secular increases from 1940 to various timepoints between 1970 and 1990. I would want to look at the r-square for the *residuals* of lead and crime rates, after detrending for time effects.

The "acid test" is what crime rates do as lead drops. For most countries, the lead levels are still pretty high, and really only the US has yet seen crime rates from a low-lead cohort past the peak. Crime rates are way down but not as much as they "should" be. Canada has seen a fall-off too but not to a distinctly low-lead state. With many countries (Britain, Italy) you see crime rates level off just where you expect them to, but are not yet supposed to drop.

I'm still convinced there is a substantial effect, especially in view of basic biology and some of the cited effects of higher lead levels in arrested accused criminals. I just think it'll probably be more like 30% than 65%+. As mentioned above, even a small effect justifies aggressive lead removal because crime is so expensive. That said, OCSteve is very right to be careful about *how* we fix things. Strong reactions to critical problems are a great time for corruption, political pork, etc.

I don't know what problems OCSteve has with asbestos mitigation, but my concern was and is overreaction. Asbestos is bad, but it's a fairly limited problem apart from asbestos workers. It's basically dangerous only if inhaled as an aerosol, and consequently apart from asbestos workers and people who live near mines or factories that work with it asbestos-related illness is very rare. It's fairly well behaved and doesn't form aerosols in meaningful quantities unless you grind it up, so the requirements for encasing asbestos work areas with sealed enclosures is over the top. Also, unless there's construction work, it stays where it is, so there's no need to *remove* asbestos. Nonetheless I've seen a community house tented and invaded by workers in bunny suits to remove 2 panels of asbestos in a closet. It's pretty crazy.

Lead paint is much easier to get off (sanding is obviously the worst, but it chips too). Lead is also dangerous if eaten, while asbestos is pretty innocuous if a 2-year swallows a bit.

Hilzoy :which problem with asbestos are you worried about

It seemed to me that asbestos became one huge boondoggle for politically connected contractors. It was important that it be dealt with, but the way it evolved over a couple of decades makes me wary of how this could become a similar situation.

Also – what Curt said.

Or several notes, a chord maybe.

A chord of caution. I like that.

Hi,
I read the article.This lead exposure leads to permanent damage of the brain and are linked to higher numbers of arrests, particularly for violent crime.

Hi,
Nice blog.The "acid test" is what crime rates do as lead drops. For most countries, the lead levels are still pretty high, and really only the US has yet seen crime rates from a low-lead cohort past the peak.

Good informative post.Lead paint is much easier to get off. Lead is also dangerous if eaten, while asbestos is pretty innocuous if a 2-year swallows a bit...

The comments to this entry are closed.

Blog powered by Typepad