Crime in the United States has been dropping steadily since the latter half of 2022, and one of the main challenges of explaining it is that the drop defies most simple explanations.
Entropy, teen pregnancy, drug abuse, unemployment et alia I would think would be in part the sequelae of lead poisoning, but lead poisoning is generally poorly instrumented---NHANES II and after used confirmed blood lead level and geometric means, which is mathematically incoherent.
The correct instrumentation would be milk tooth lead surveillance, preferably including a twin study component (monozygotic versus same sex fraternal twins), as uptake per unit intake of lead is almost certainly highly genetically variable, with higher uptake among most black ethnic groups, low uptake per unit intake among most east Asian ethnic groups.
Note that "Title X (the Residential Lead-Based Paint Hazard Reduction Act of 1992)," although passed in 1992, didn't really pick up steam until 1997 (i.e. 23 years before the murder boom, as one may expect on the basis of peak offending around age 20 and allowing a mean 2-3 year joint maternal bone lead storage and pregnancy period).
Early cleanup regulations (early versions of LEAD SAFE) required hand scraping with trigger bottle spray of water on area, which is not cost-effective, and was thus not generally done; power tools were used, which lead to increasing lead poisoning, until vacuums with HEPA filters attached to sanders became the norm, allowing lead poisoning to drop. Note that the NHANES blood lead protocol could not be designed better to miss such a spike in lead poisoning, per my argument in chapter four, section three.
Interesting. I wasn’t aware that lead poisoning surveillance was so statistically fraught. That being said, a trifecta of landmark federal laws and agency directives were implemented during the ‘70s and ‘80s that absolutely contributed to the 97% decline in blood lead levels of American children between 1969 and 2016: The Clean Air Act of 1970, the Consumer Product Safety Commission’s indoor lead paint ban of 1978, and the Safe Drinking Water amendment of 1986. Peer-reviewed studies suggest that this reduction in childhood lead exposure could be responsible for as much as half of the decline in violent crime that occurred in the US during the ‘90s:
When one speaks of lead contributing to x percentage of crime, or removing lead contributing y percentage of crime, there is an implicit model, in which
1. different causes of crime are added together,
2. each cause has a consistent variance in terms of crime causation, and
3. the lead term has only a scaling constant.
Without these constraints, it is mathematically meaningless to ascribe some percentage variation to an intervention in lead poisoning (i.e. mathematically naive), yet I reject such models, and I list two such considerations that force the issue:
Unemployed individuals are, for the same degree of lead poisoning, more likely to commit a crime, than are employed individuals.
When the whether is (within the variation of comfort) warmer, absent rain, there is a higher expectation of crime, popular lead poisoning being held equal.
From February to June this year, Chicago has had elevated murder rates, despite less lead poisoning, on account of a much warmer spring. The (IMAO) root cause of murder has been substantially reduced, yet there was more crime, but it would be mathematically naive in this context to reject lead as causal.
Well… as an old Bloomer cynic, I often ask common everyday folks in the grocery line and “ on the street” if they think is going down or up. 99% said it is going up. Every home in my lower middle class neighborhood has a Ring doorbell. The era of “Stranger Danger” exist deeply in our society. Please interview common everyday folks from all economic classes instead of the “ talking head” of academia
Why would a random stranger in a grocery line know anything about national long term crime trends? That's like asking me if soccer is becoming more or less popular in puerto rico.
Of course they think it’s going up. We see every crime online and live with breaking coverage these days. 30 years ago a kid could be kidnapped and you wouldn’t have wall to wall coverage for days. You wouldn’t even know about it unless it was in your time.
A super-fascinating study. We need more of these big-picture crime trends analysis. Thanks for bringing it to my attention. I'd welcome your thoughts about their choice to use UCR data instead of NCVS. Since the NCVS shows higher crime rates through the 70s and 80s, we might not see that gradual crime increase between 1970 and 1990 so the trends might not line up as well.
I'm very suspicious of their decision to not include age. I wouldn't be surprised if age was "too strong" and made their results less "sexy" and novel appearing. Age is a great driver or proxy for mobility, homeownership, incarceration and many of the other variables.
All the criminals have been hired by ICE. Guaranteed income, no chance of jail or even fines. What a deal.
I found this one fascinating too. Of course it makes everyone uncomfortable to consider that abortion can affect crime rates, but here you have it:
https://law.stanford.edu/publications/the-impact-of-legalized-abortion-on-crime-over-the-last-two-decades/
Where does the “lead toxicity” hypothesis (of which I am a strong believer) fit into this model of explanation?
Entropy, teen pregnancy, drug abuse, unemployment et alia I would think would be in part the sequelae of lead poisoning, but lead poisoning is generally poorly instrumented---NHANES II and after used confirmed blood lead level and geometric means, which is mathematically incoherent.
See chapter 4, in particular section 3.
https://we.tl/t-iBdu9iftqj
The correct instrumentation would be milk tooth lead surveillance, preferably including a twin study component (monozygotic versus same sex fraternal twins), as uptake per unit intake of lead is almost certainly highly genetically variable, with higher uptake among most black ethnic groups, low uptake per unit intake among most east Asian ethnic groups.
Note that "Title X (the Residential Lead-Based Paint Hazard Reduction Act of 1992)," although passed in 1992, didn't really pick up steam until 1997 (i.e. 23 years before the murder boom, as one may expect on the basis of peak offending around age 20 and allowing a mean 2-3 year joint maternal bone lead storage and pregnancy period).
Early cleanup regulations (early versions of LEAD SAFE) required hand scraping with trigger bottle spray of water on area, which is not cost-effective, and was thus not generally done; power tools were used, which lead to increasing lead poisoning, until vacuums with HEPA filters attached to sanders became the norm, allowing lead poisoning to drop. Note that the NHANES blood lead protocol could not be designed better to miss such a spike in lead poisoning, per my argument in chapter four, section three.
Interesting. I wasn’t aware that lead poisoning surveillance was so statistically fraught. That being said, a trifecta of landmark federal laws and agency directives were implemented during the ‘70s and ‘80s that absolutely contributed to the 97% decline in blood lead levels of American children between 1969 and 2016: The Clean Air Act of 1970, the Consumer Product Safety Commission’s indoor lead paint ban of 1978, and the Safe Drinking Water amendment of 1986. Peer-reviewed studies suggest that this reduction in childhood lead exposure could be responsible for as much as half of the decline in violent crime that occurred in the US during the ‘90s:
https://www.pnas.org/doi/10.1073/pnas.2118631119
https://www.degruyterbrill.com/document/doi/10.2202/1935-1682.1796/html
When one speaks of lead contributing to x percentage of crime, or removing lead contributing y percentage of crime, there is an implicit model, in which
1. different causes of crime are added together,
2. each cause has a consistent variance in terms of crime causation, and
3. the lead term has only a scaling constant.
Without these constraints, it is mathematically meaningless to ascribe some percentage variation to an intervention in lead poisoning (i.e. mathematically naive), yet I reject such models, and I list two such considerations that force the issue:
Unemployed individuals are, for the same degree of lead poisoning, more likely to commit a crime, than are employed individuals.
When the whether is (within the variation of comfort) warmer, absent rain, there is a higher expectation of crime, popular lead poisoning being held equal.
From February to June this year, Chicago has had elevated murder rates, despite less lead poisoning, on account of a much warmer spring. The (IMAO) root cause of murder has been substantially reduced, yet there was more crime, but it would be mathematically naive in this context to reject lead as causal.
Weather---auto error insert is at it again.
Interesting. But how does this tie in with the trends in murders of partners and family members?
Is there a trend or just something where a few stories are going viral?
Well… as an old Bloomer cynic, I often ask common everyday folks in the grocery line and “ on the street” if they think is going down or up. 99% said it is going up. Every home in my lower middle class neighborhood has a Ring doorbell. The era of “Stranger Danger” exist deeply in our society. Please interview common everyday folks from all economic classes instead of the “ talking head” of academia
Why would a random stranger in a grocery line know anything about national long term crime trends? That's like asking me if soccer is becoming more or less popular in puerto rico.
Of course they think it’s going up. We see every crime online and live with breaking coverage these days. 30 years ago a kid could be kidnapped and you wouldn’t have wall to wall coverage for days. You wouldn’t even know about it unless it was in your time.
I also speak to my long time and close friends for their views. Please go to your neighbors , friends and random stranger.
A super-fascinating study. We need more of these big-picture crime trends analysis. Thanks for bringing it to my attention. I'd welcome your thoughts about their choice to use UCR data instead of NCVS. Since the NCVS shows higher crime rates through the 70s and 80s, we might not see that gradual crime increase between 1970 and 1990 so the trends might not line up as well.
I'm very suspicious of their decision to not include age. I wouldn't be surprised if age was "too strong" and made their results less "sexy" and novel appearing. Age is a great driver or proxy for mobility, homeownership, incarceration and many of the other variables.
Perhaps somewhat consistent, see https://www.governing.com/urban/the-murder-rate-is-the-lowest-in-modern-history-heres-why. Tracks the murder rate with drug overdoses, alcohol deaths, and gun purchases.
Also, see https://assets.pubpub.org/ccb6ab371-fc70-4ef3-b2e5-d0d0580b4d37/p9e540d36-8304-4f88-8b5c-2e8ac2fa0ed9/ubce40ec2-e33e-44d2-9ff9-d6167c6c64f9/Controlling_Shootings_by_Focusing_on_Tail_Risks-01784105469249-71784106475538.pdf which does argue that concentrating attention on those most at risk of shooting or being shot has contributed.