7 Comments
author

Good question. Here's an excellent paper that dives into the question of what happened in 2015: https://www.ojp.gov/pdffiles1/nij/249895.pdf. The big difference between 2015 and 2020 was that the 2015 spike was pretty concentrated in a few cities. The 2020 spike is in cities and suburbs, big and small. It's much broader and deeper.

Expand full comment

Thank you for the reply. On its face it seems hard to overlook a — possibly the — only common denominator between the nation’s only surges in homicide in the last decade+: a highly publicized incident of a police shooting, with a racial vector. You’re likely more familiar than I with the wealth of research out there on this, but Vox (of all places) even noted a study on the correlation between size/scale of anti-police protest and increase in homicide post-Ferguson: “That is, there were a few thousand more homicides in the places where there were BLM protests than would have been expected if those places followed the same trends as the ones that didn’t see protests.” (Link to article, with embedded links to the research: https://www.vox.com/22360290/black-lives-matter-protest-crime-ferguson-effects-murder)

I just mention this because it seems possible (to a lay-person) that you’ve adduced a relevant variable (people stuck at home more due to pandemic) but that the variable doesn’t operate in the way argued for. ie, people stuck at home, near their belongings, cars, etc, may deter nearly every form of crime where the presence of additional people is an inconvenience to the criminal (thieves/burglars don’t want to deal with the owner); but for crimes where those people themselves are the targets of the crime (homicide), there is no deterrent effect. This would square with the Ferguson effect theory: police retreat, so you might expect many forms of crime to rise. For most crimes, the people stuck at home provide the deterrent effect previously provided by police. When people are the target of the crime, they (obviously) don’t provide the deterrent. Ergo only certain forms of crime rise. Other forms of violent crime (eg assault) decrease because for example people aren’t out late drinking and doing drugs at clubs and getting in fights afterwards (and perhaps these sorts of bar fight encounters simply aren’t a major source of escalation to homicide).

The main difference, if any, between the post-Brown and post-Floyd episodes seems that the post-Floyd protests were more widespread, larger in scale, and more intense than the post-Brown protests. This would also square both with the correlating homicide spikes (much steeper and more widespread after the larger, more widespread protest) and with the conclusion of the Vox-cited study above.

I appreciate your methodology and rational inquiry.

Expand full comment
author

Yes, totally agree. And child welfare cases as well. Hopefully, the data are available shortly and we can sort this out more thoroughly.

Expand full comment

I want this to be the answer.

> For an explanation for the homicide spike to hold, it must be true everywhere, for almost everyone—homicide spiked in cities, suburbs, and exurbs, in cities with red mayors and blue mayors. This explanation does.

Does this analysis hold for other countries or is US an outlier in homicide spike?

The timing, duration, and severity of covid mitigation measures like lockdown orders and school closures differed significantly between cities and states. If "everyone was at home because of covid" is the causal factor, then should that be pretty evident in the time series data?

Expand full comment

Curious how you explain the identical phenomenon after the Ferguson MO shooting of Michael Brown in 2014. Murder up, crime overall and violent crime both down; no pandemic closures; murder remained elevated for a few years afterwards.

Expand full comment

Thanks for this logical analysis. The causal analysis you paint suggests that domestic violence of all types should have also risen.

Expand full comment

My initial instincts are to agree with that hypothesis. I've read studies that link long term decline in the overall murder rate with increased aggressiveness in responding to DV.

Expand full comment