25 ms·

What Caused the 2020 Homicide Spike?

34 points by feross 4y ago

cyanydeez 4y ago
- jprd 4y ago
  I wish I could give this all the points. I don't have the energy to deal with that site?/blog?/tumblr?/paidrantpage? again, but yeah. Reporting. The rest is stupid obvious for anyone not living in a gilded cage.
- ALittleLight 4y ago
  Are you saying that reporting, the activity of journalists, caused the murder spike or are you saying that you are reporting this post to hackernews moderators? Either way I think you should explain a bit.
andrewflnr 4y ago
Why is it verboten for left-wingers to believe this? To me, police pulling back from black communities in retaliation for the protests fits into their narrative. But then again I don't really understand what either side is on about, these days.
- addicted 4y ago
  Unfortunately the article has a lot of issues detailed by several commenters below. I’m personally convinced that at least a significant part of the rise in crime definitely was related to police pullbacks (both for anecdotal and more data driven reasons…but I need to clean up that data still). This article, however, does not make that case even close to convincingly.
- xbar 4y ago
  I believe there was a spike. Check out the flatline for Asian murders per 100,000. I am surprised, given the amount of Asian hate that is being reported.
  - alsaaro 4y ago
    Curious there was a massive spike in reported hate crimes against blacks (45%), to 2,871 (2020) from 1,972 (2019). Strangely, Alexander writes: "there was no corresponding rise in hate crimes the way I would expect if this was whites targeting blacks, so I think the perpetrators were most likely also black." Alexander is a smart guy, and he almost certainly found the same data I did, so I suspect his use of the word "expect" is doing a lot of work here.
    - derbOac 4y ago
      Theres a lot of that unfortunately. For example, when discussing foreign trends he's similarly sloppy, acting like "I searched Google a bit and this is the first I found". But rates went up in Brazil in 2020 — this had substantial news coverage at the time — and it doesnt take long to find this: https://foreignpolicy.com/2022/05/20/homicide-murder-violence-united-states-latin-america-caribbean/ https://foreignpolicy.com/2022/05/20/homicide-murder-violenc... I also don't think it's reasonable to expect psychosocial effects of lockdown issues to just pop up immediately in March 2020 like a light switch. I'm not really convinced the effect wasn't there but I'm also not convinced it was, and think Alexander was being a bit too armchair about the whole thing in a way that made me wonder what else he was missing.
      - EdwardDiego 4y ago
        Yeah, the expectation of an immediate effect from lockdown struck me as rather odd.
  - vanattab 4y ago
    Has anyone here witnessed asian hate in response to covid. I have seen much reported about it but have really ever witnessed any myself.
    - mushbino 4y ago
      I lived in a Chinese neighborhood at the time and there weren't any hate crimes. I feel like news agencies find a hot topic latch on and blow it out of proportion. Fear sells like nothing else.
alsaaro 4y ago
My pet theory is that the 2020 homicide spike may have been caused primarily by Covid unemployment checks, for which almost all were eligible -- at least at their base line level --, eligibility even extended to people who did not qualify for unemployment because they were not employed. Borrowing costs for business owners through PPP, and the Federal Reserve buying assets; juiced consumption so high we even ran out of spare change. All goods were inflated; this includes drug markets. If you use a proxy for drug use as fatal drug overdoes, drug overdoses increased thirty percent between 2019 to 2020 (71K to 91K), an unprecedented jump. I suspect that over the short term drug market supply is inelastic, and the people that sell drugs simply kill each other for market share, increasing the murder rate. This also may explain the precipitous drop in crime in the mid-1990s; the federal government largely stopped mailing people welfare checks, and people didn't have excess money for vice, subsequently fueling the drug trade and its associated gang conflict. https://nida.nih.gov/research-topics/trends-statistics/overdose-death-rates https://nida.nih.gov/research-topics/trends-statistics/overd...
- snypher 4y ago
  I'm going to disregard your link between welfare recipients and drug use for now; is there anything to show thay this spike in homicide is drug related?
  - alsaaro 4y ago
    You are correct to question whether there is a connection between welfare and drug use, perhaps there isn't. But a broader point is that for most people your income is due to someone's spending, so a reduction in income (welfare and unemployment in this case) may reduce spending on all things, necessities and vices alike. As for the connection between homicides and drugs, this is a general assumption, perhaps there is no connection.
srean 4y ago
The book Freakonomics has a chapter that analyses the sharp fall of homicide that followed a sharp rise. It makes for an interesting read. Its topical given the recent overturning of Roe Vs Wade. The authors present many potential explanations that were put forward at that time to explain the phenomena. These included, more cops on the ground, stringent sentences, broken-window policy and so on. In the author's assessment all these hypotheses turned out to be wrong. According to the author's analysis the likely cause was the ratification of the Roe-Wade judgement. The causal chain is as follows, legalized abortion resulted in the reduction of the number of people who were likely to get very little parental support and social support while growing up. Arguably, this population is more predisposed towards crime and its relative reduction led to a fall that aligned with the timeline -- around the time the newborns would have hit the late teens. Attributing an irrefutable cause is hard.
- drewrv 4y ago
  Crime data is messy and can take decades of debate without ever coming to any conclusions. Beyond the abortion argument, I’ve heard convincing claims that the cause of a decline in crime was due to phasing out leaded gasoline. For 2020, I blame Tiger King for glamorizing a life of crime. I’m prepared to put together some infographics about it and the theory will be as convincing as any other.
  - srean 4y ago
    Indeed. I am familiar with the leaded gas hypothesis. Inferring causality from observations without randomizing the intervention is hard to neigh impossible. In Levitt's defense, he uses natural experiments, where natural variation does the work of random assignment of a subject to 'treatment group' and 'control group' and does the causal analysis. This is over and above accounting for the known hypotheses that purport to explain the phenomenon. Short of RCT it doesnt get much better than that.

34 points by feross 4y ago

jprd 4y ago
I wish I could give this all the points. I don't have the energy to deal with that site?/blog?/tumblr?/paidrantpage? again, but yeah. Reporting. The rest is stupid obvious for anyone not living in a gilded cage.
ALittleLight 4y ago
Are you saying that reporting, the activity of journalists, caused the murder spike or are you saying that you are reporting this post to hackernews moderators? Either way I think you should explain a bit.

Why is it verboten for left-wingers to believe this? To me, police pulling back from black communities in retaliation for the protests fits into their narrative. But then again I don't really understand what either side is on about, these days.

addicted 4y ago
Unfortunately the article has a lot of issues detailed by several commenters below. I’m personally convinced that at least a significant part of the rise in crime definitely was related to police pullbacks (both for anecdotal and more data driven reasons…but I need to clean up that data still). This article, however, does not make that case even close to convincingly.
xbar 4y ago
I believe there was a spike. Check out the flatline for Asian murders per 100,000. I am surprised, given the amount of Asian hate that is being reported.
- alsaaro 4y ago
  Curious there was a massive spike in reported hate crimes against blacks (45%), to 2,871 (2020) from 1,972 (2019). Strangely, Alexander writes: "there was no corresponding rise in hate crimes the way I would expect if this was whites targeting blacks, so I think the perpetrators were most likely also black." Alexander is a smart guy, and he almost certainly found the same data I did, so I suspect his use of the word "expect" is doing a lot of work here.
  - derbOac 4y ago
    Theres a lot of that unfortunately. For example, when discussing foreign trends he's similarly sloppy, acting like "I searched Google a bit and this is the first I found". But rates went up in Brazil in 2020 — this had substantial news coverage at the time — and it doesnt take long to find this: https://foreignpolicy.com/2022/05/20/homicide-murder-violence-united-states-latin-america-caribbean/ https://foreignpolicy.com/2022/05/20/homicide-murder-violenc... I also don't think it's reasonable to expect psychosocial effects of lockdown issues to just pop up immediately in March 2020 like a light switch. I'm not really convinced the effect wasn't there but I'm also not convinced it was, and think Alexander was being a bit too armchair about the whole thing in a way that made me wonder what else he was missing.
    - EdwardDiego 4y ago
      Yeah, the expectation of an immediate effect from lockdown struck me as rather odd.
- vanattab 4y ago
  Has anyone here witnessed asian hate in response to covid. I have seen much reported about it but have really ever witnessed any myself.
  - mushbino 4y ago
    I lived in a Chinese neighborhood at the time and there weren't any hate crimes. I feel like news agencies find a hot topic latch on and blow it out of proportion. Fear sells like nothing else.

alsaaro 4y ago

My pet theory is that the 2020 homicide spike may have been caused primarily by Covid unemployment checks, for which almost all were eligible -- at least at their base line level --, eligibility even extended to people who did not qualify for unemployment because they were not employed. Borrowing costs for business owners through PPP, and the Federal Reserve buying assets; juiced consumption so high we even ran out of spare change. All goods were inflated; this includes drug markets. If you use a proxy for drug use as fatal drug overdoes, drug overdoses increased thirty percent between 2019 to 2020 (71K to 91K), an unprecedented jump. I suspect that over the short term drug market supply is inelastic, and the people that sell drugs simply kill each other for market share, increasing the murder rate. This also may explain the precipitous drop in crime in the mid-1990s; the federal government largely stopped mailing people welfare checks, and people didn't have excess money for vice, subsequently fueling the drug trade and its associated gang conflict. https://nida.nih.gov/research-topics/trends-statistics/overdose-death-rates https://nida.nih.gov/research-topics/trends-statistics/overd...

snypher 4y ago
I'm going to disregard your link between welfare recipients and drug use for now; is there anything to show thay this spike in homicide is drug related?
- alsaaro 4y ago
  You are correct to question whether there is a connection between welfare and drug use, perhaps there isn't. But a broader point is that for most people your income is due to someone's spending, so a reduction in income (welfare and unemployment in this case) may reduce spending on all things, necessities and vices alike. As for the connection between homicides and drugs, this is a general assumption, perhaps there is no connection.

srean 4y ago

The book Freakonomics has a chapter that analyses the sharp fall of homicide that followed a sharp rise. It makes for an interesting read. Its topical given the recent overturning of Roe Vs Wade. The authors present many potential explanations that were put forward at that time to explain the phenomena. These included, more cops on the ground, stringent sentences, broken-window policy and so on. In the author's assessment all these hypotheses turned out to be wrong. According to the author's analysis the likely cause was the ratification of the Roe-Wade judgement. The causal chain is as follows, legalized abortion resulted in the reduction of the number of people who were likely to get very little parental support and social support while growing up. Arguably, this population is more predisposed towards crime and its relative reduction led to a fall that aligned with the timeline -- around the time the newborns would have hit the late teens. Attributing an irrefutable cause is hard.

drewrv 4y ago
Crime data is messy and can take decades of debate without ever coming to any conclusions. Beyond the abortion argument, I’ve heard convincing claims that the cause of a decline in crime was due to phasing out leaded gasoline. For 2020, I blame Tiger King for glamorizing a life of crime. I’m prepared to put together some infographics about it and the theory will be as convincing as any other.
- srean 4y ago
  Indeed. I am familiar with the leaded gas hypothesis. Inferring causality from observations without randomizing the intervention is hard to neigh impossible. In Levitt's defense, he uses natural experiments, where natural variation does the work of random assignment of a subject to 'treatment group' and 'control group' and does the causal analysis. This is over and above accounting for the known hypotheses that purport to explain the phenomenon. Short of RCT it doesnt get much better than that.