As a long time retired Social Worker and Investigator I think the reason for Murder Rate plunging might be due to the following 1) Gangs appeal has lost its appeal 2) more knowledge about how to reach hard core damaged teens in the area of trauma 3) cameras all over the place . It seems that the kids today who grew with Technology recognizes that Big Brother is always around. My thoughts are not based on analysis but a lifetime growing up and working around crime .
I would to thank Jeff Archer for putting analysis back into the crime reporting. Unfortunately most folks still think crime is forever growing making us permanently hypervigliant.
I think it is due to kids just staying indoors more. Being locked down during COVID had separated them from the type of activity and bad influences that would of influenced them to get into criminal activity.
Middle schoolers during COVD are high schoolers and skipped the initial gang affinity that starts in the 8th and 9th grades.
What we are seeing is a passive cohort in the population entering what would be their most violent ages 17 to 25.
In 2 to 3 years we will start to see murder and other violent crimes tick up again.
This is what I was going to say. Has anybody been into an urban center after dark recently? Nightlife is dead. The kids don’t drink anymore and even if they did there are homeless on every corner discouraging revelry.
"Middle schoolers during COVD are high schoolers and skipped the initial gang affinity that starts in the 8th and 9th grades."
You don't have to be in school for gang affinities to start. They aren't a study hall / math class activity. They're a hanging out in the neighborhood kind of thing, and there was plenty of that during covid in most neighborhoods.
Yeah I’ve got to think that’s a factor. People are afraid of saying or doing the wrong thing and getting filmed and losing their job as a result. I’ve got to think growing up in essentially a crowdsourced security state deters a lot of young folks from casual criminality.
Hi Jeff, there are endless examples of governments and private entities investing millions of dollars in high-crime communities (e.g., Baltimore's Sandtown) without achieving results.
"This Article is built around a central empirical claim: most reforms and interventions in the criminal legal space are shown to have little lasting impact when evaluated with gold-standard methods of causal inference.” See https://www.crimeinamerica.net/does-anything-work-to-reduce-crime/.
There are multiple examples in the cited article as to a social programs approach not having an impact.
"When it comes to the type of limited-scope interventions evaluable via RCT and other quasi-experimental methods, the engineer’s view appears to be mostly a myth. More than fifty years of RCT evidence shows the limits in our ability to engineer change with this type of intervention (emphasis added).”
The essence of her argument is that it’s tough to engineer social change through programs.
As you are aware, it's difficult to impossible to explain a decrease in homicides, a stance taken by most criminologists after the 20 year decline. Lead paint? Average age? Police endeavors? Some are suggesting that fentanyl deaths have led to a vast decrease in overdoses, and crime.
"Some are suggesting that fentanyl deaths have led to a vast decrease in overdoses..."
Wasn't that also the explanation leveled at the waning crack epidemic of the 1990's? Too many deaths due to overdose, neglect and misadventure drastically reduced the population of drug users and crime in general declined as a result.
Spending Big Bucks does not necessarily mean that the programs were well-designed and effective. There are many examples of effective social programs that were aborted by inconsistent focus/funding as federal administrations change.
The unarguable fact is that the rates of All Crimes have dropped dramatically since the peaks of the early 90's.
Yeah IMO the complete collapse of social and institutional capital (like schools, community social events/programs) during Covid and then the gradual re-emergence of that institutional and social capital seems to me the obvious reason.
My guess (has anyone looked at this) it has to do with the increasing smartphone usage and the demographic that is highest risk of murdering someone (around 22 growing up on high levels of smartphone usage) and so high risk males spending more time alone and entertained ish, combined with a good labor market conditions for lower skilled men. It belongs in the category of teens having less sex, less time spent with friends, less relationships, and being more lonely and anxious but thankfully less murdering.
There's only one candidate for "the Great Murder Decline" title, and that was the one that started in the mid-1990s undoing a jump that started 30 years earlier, not 3 years that undid a much smaller jump that started 8 years before.
It sure is easier to write about what is happening with crime than why it is happening. That's because we just don't know, and there is very little real proof for cause-and-effect. What's more important is, we can say with confidence that murder/violent crime does not go down because of more police because crime diminished with fewer police. Stricter sentences do not lower the crime rate because it went down without any real increase in sentences. Local law enforcement tactics don't play a significant role because crime went both up and down in a nationwide wave regardless of local police programs and strategies. In fact, this data prove that the conventional political arguments over crime are simply wrongheaded.
It is important to understand that murder and assault (the biggest category of violent crimes) are overwhelmingly emotional, not logical, acts. As Jens Ludwig writes in the recent book Unforgiving Places, maybe 80 or 90 percent of these acts are not planned. Reality is totally different from murder mysteries and crime shows. It seems reasonable that COVID increased perpetrators' psychological and mental health problems, and perhaps created increased opportunities for interpersonal friction, that caused murder and to a lesser extent violent crime to rise. And the end of COVID had the opposite effect. But, again, nobody really knows why crime rates increase or decrease and we should all be ultra-modest about our guesses.
Your first paragraph is very confident in saying what you believe had been solidly disproven, then you express humility about attribution in the second paragraph.
If there are a plethora of positive and negative inputs affecting the outcome, and many of them change, it becomes harder to so confidently attribute it to a given factor, or to rule out a given factor.
> "we can say with confidence that murder/violent crime does not go down because of more police because crime diminished with fewer police. "
Let's parse that a bit more carefully. We could say that the statistics show that murder rates (and likely other crime rates) can potentially go down even when law enforcement staffing is also down; that demonstrates that there are other factors in play besides law enforcment staffing levels (etc), and that these other factors can at least sometimes dominate the dynamics. Staffing levels do not control crime levels, but we haven't shown that they have no effect.
That is, the stats do not rule out a hypothesis that "crime levels would have gone down even more in this period if law enforcement staffing had been robust while all other factors stayed the same as the real world".
It would rule out a hypothesis that "crime levels will always go up when police staffing is reduced, regardless of any combination of other factors" - but that's not really something any serious person would suggest.
I find it tentatively encouraging that "other factors" can in cases like this over-ride any positive or negative influence of law enforcement staffing levels - IF we can get a handle on those other factors, perhaps we can continue to reduce crime more than we could by increasing staffing. Maybe there is a lever to use, if we can find it. But we have not proven that law enforcement staffing has NO effect.
Per CDC data, total murders for the first 3 months of 2020 were up 10% compared to the first months of 2019, which is substantial though not huge. (This is to answer a question I asked in a reply I since deleted.)
The last 3 months of 2019 were also elevated by roughly 10%.
Personally, I don't read a whole lot into it, because 10% increases across a few months seem to be within the range of noise. This one continued a bit longer, but to me it doesn't necessarily suggest that anything notable had happened.
It’s probably that cops stopped being on de facto strike after George Floyd. Isn’t there a well documented effect of cops withdrawing from work in response to police protests that is associated with an increase in crime?
Isn’t Ludwig’s book describing what was called “broken windows theory” in the ‘90s? It’s interesting because a)obviously, stopping people from jumping subway turnstiles does nothing to address systemic issues but also b)it seems like enforcing the low-level stuff, which is a much easier lift than eradicating poverty, actually does help prevent more serious crime.
This article to me is a giant facepalm. It is basically the author presenting as conclusion what he thinks to be the causes, then selectively inserting pieces of facts to support his pov. The analysis process is thus devoid of deductive reasoning. That said, I can't fault the author too much because many analyses on violent crimes always have one giant blind spot. The blind spot is that the vast majority of violent crimes are committed by men. Once that blind spot is addressed, it would be much easier to find the answer. I'm not saying this to "blame men". I'm bringing this up as a demographic element to begin deductive reasoning.
You deduct why murder rate has declined by first isolating the demographic of perpetrators to men. From there, you break down the demographic further to young men. Then what kind of young men? What age range? Does class have anything to do with it? What about geographic locations? Does any type of murder have a bigger decrease over others? After breaking down the demographic to the specific type of men who committed murders, we can ask the question: What changed for these men in the last 4 years? Is it economic? Is it changes in environment? Was the spike in murders an exception and not a norm committed by men of a certain age range who have since grown older? Etc, etc.
For people trying to answer questions about violent crimes, you will come to more convincing conclusions why murder rates have declined only when you take away that giant blind spot. Otherwise, you're infusing too many variables about non-applicable demographics into the equation and your entire analysis gets muddled.
Just had a very long debate with someone about this. I think in order to understand why it went down we need to understand why it went up. If you look at most of the crime data during the spike, it was young black males not middle aged people trying to feed their family. Why would it go up then? Well in the most simple sense a general sense of chaos during that time. Covid plus the Floyd era. Post Floyd there was a general sense that you could wild out and not get in trouble. How many videos of looters did we see with no consequences for them? Also a huge part of this is the pandemic unemployment assistance fraud going on. Young people suddenly had access to thousands of dollars without really having to prove that they were searching for a job. What can you do with that money? Wild out. Especially in the city I’m from and general what I saw nationwide was this was gang related stuff. Lil z killed lil y because of a diss track and a perceived sense of disrespect. It wasn’t because the local back pack program shutdown for 2 years. The fact that the combination of lawlessness post Floyd and flexing money from PUA is not discussed more by criminologists is sort of baffling to me. Because this is clear as day to me all other stuff is mostly noise.
I'd encourage Jeff and co. to study the case study of Baltimore, where I lived for many years. Compared to some other cities, violent crime didn't come down as much there between the 1990's and 2014, though the rate did drop somewhat. But, the violence surged in Baltimore well before the onset of COVID, starting with the tragic death of Freddie Gray in police custody, and stayed high, until 2023, when it began to drop dramatically. In the last 3 years before COVID, though, the city government and some community- based violence intervention groups did achieve significant results (e.g., in the Cherry Hill neighborhood). And Mayor Scott has pulled together a web (various levels of government, non-profits, community groups) that's focused on crime reduction, and surely has contributed to the notable improvements. So, take a good look at Baltimore. With regard to the national trends, contrary to Trumpian rhetoric, 2023 and 2024 were very good years for the national economy: very low unemployment, falling inflation, rising wages at the low end of the scale, lots of public construction jobs, etc. Young people who might have been unemployed in earlier years were driving Amazon delivery trucks or holding up paddles (stop, proceed with caution, signs) at bridge reconstruction sites. I hope that we don't regress in the next few years!
Lehman makes the point that police departments, saddled with reduced staffing and increased work loads, seem to be prioritizing murders over all other crimes.
I would note that the NCVS (which measures everything except homicides) has seen remarkable increases year on year even as the homicide rate drops.
Last CPP 1963-78, crime peaked in 1980, fell a bit and rose to a second peak around 1990 before beginning its long-term decline, so figure a 7-year lag (using center of twin peaks). For the CPP before that (1913-27) crime began its long-term decline around 1933, so a 6-year lag. For the CPP before that (1863-77) we seen a general decline in arrest rates after 1860. In Britian, assuming they too had a CPP in the 1830's like here, a long-term decline began after 1842 IIRC.
So maybe. Specifically, I would say the current decline reflects the unwinding of the George Floyd bump, but since that was a CPP event, it's really an instance of the idea expressed above.
I will note that mass shootings, which I track as an empirical correlate with CPPs, declined dramatically in 2024 and seems to be continuing, which is consistent with a peak in such political unrest around 2020, consistent with Peter Turchin's 2012 prediction.
I wonder if all the deportations or threat of deportations from the Trump administration have contributed to the decline at all. At least as far as the 2025 data is concerned.
As a long time retired Social Worker and Investigator I think the reason for Murder Rate plunging might be due to the following 1) Gangs appeal has lost its appeal 2) more knowledge about how to reach hard core damaged teens in the area of trauma 3) cameras all over the place . It seems that the kids today who grew with Technology recognizes that Big Brother is always around. My thoughts are not based on analysis but a lifetime growing up and working around crime .
I would to thank Jeff Archer for putting analysis back into the crime reporting. Unfortunately most folks still think crime is forever growing making us permanently hypervigliant.
I think it is due to kids just staying indoors more. Being locked down during COVID had separated them from the type of activity and bad influences that would of influenced them to get into criminal activity.
Middle schoolers during COVD are high schoolers and skipped the initial gang affinity that starts in the 8th and 9th grades.
What we are seeing is a passive cohort in the population entering what would be their most violent ages 17 to 25.
In 2 to 3 years we will start to see murder and other violent crimes tick up again.
This is what I was going to say. Has anybody been into an urban center after dark recently? Nightlife is dead. The kids don’t drink anymore and even if they did there are homeless on every corner discouraging revelry.
"Middle schoolers during COVD are high schoolers and skipped the initial gang affinity that starts in the 8th and 9th grades."
You don't have to be in school for gang affinities to start. They aren't a study hall / math class activity. They're a hanging out in the neighborhood kind of thing, and there was plenty of that during covid in most neighborhoods.
Yeah I’ve got to think that’s a factor. People are afraid of saying or doing the wrong thing and getting filmed and losing their job as a result. I’ve got to think growing up in essentially a crowdsourced security state deters a lot of young folks from casual criminality.
+1 for surveillance
Hi Jeff, there are endless examples of governments and private entities investing millions of dollars in high-crime communities (e.g., Baltimore's Sandtown) without achieving results.
"This Article is built around a central empirical claim: most reforms and interventions in the criminal legal space are shown to have little lasting impact when evaluated with gold-standard methods of causal inference.” See https://www.crimeinamerica.net/does-anything-work-to-reduce-crime/.
There are multiple examples in the cited article as to a social programs approach not having an impact.
"When it comes to the type of limited-scope interventions evaluable via RCT and other quasi-experimental methods, the engineer’s view appears to be mostly a myth. More than fifty years of RCT evidence shows the limits in our ability to engineer change with this type of intervention (emphasis added).”
The essence of her argument is that it’s tough to engineer social change through programs.
As you are aware, it's difficult to impossible to explain a decrease in homicides, a stance taken by most criminologists after the 20 year decline. Lead paint? Average age? Police endeavors? Some are suggesting that fentanyl deaths have led to a vast decrease in overdoses, and crime.
Explaining crime is the wild west of sociology.
Best, Len.
"Some are suggesting that fentanyl deaths have led to a vast decrease in overdoses..."
Wasn't that also the explanation leveled at the waning crack epidemic of the 1990's? Too many deaths due to overdose, neglect and misadventure drastically reduced the population of drug users and crime in general declined as a result.
Yep. But one of many explanations. Len.
There is a fair amount of evidence for both lead and increased abortion availability. Isn’t the argument in Freakanomics pursuasive?
Spending Big Bucks does not necessarily mean that the programs were well-designed and effective. There are many examples of effective social programs that were aborted by inconsistent focus/funding as federal administrations change.
The unarguable fact is that the rates of All Crimes have dropped dramatically since the peaks of the early 90's.
I'd love Jeff to expand on this, because it's not obvious to me:
"It is also helpful to acknowledge that the factors driving murder down now are almost certainly not the same factors that drove murder up in 2020"
Yeah IMO the complete collapse of social and institutional capital (like schools, community social events/programs) during Covid and then the gradual re-emergence of that institutional and social capital seems to me the obvious reason.
Or the behavioral changes from people living in a collapsing society. People don’t leave their houses much anymore.
My guess (has anyone looked at this) it has to do with the increasing smartphone usage and the demographic that is highest risk of murdering someone (around 22 growing up on high levels of smartphone usage) and so high risk males spending more time alone and entertained ish, combined with a good labor market conditions for lower skilled men. It belongs in the category of teens having less sex, less time spent with friends, less relationships, and being more lonely and anxious but thankfully less murdering.
There's only one candidate for "the Great Murder Decline" title, and that was the one that started in the mid-1990s undoing a jump that started 30 years earlier, not 3 years that undid a much smaller jump that started 8 years before.
It sure is easier to write about what is happening with crime than why it is happening. That's because we just don't know, and there is very little real proof for cause-and-effect. What's more important is, we can say with confidence that murder/violent crime does not go down because of more police because crime diminished with fewer police. Stricter sentences do not lower the crime rate because it went down without any real increase in sentences. Local law enforcement tactics don't play a significant role because crime went both up and down in a nationwide wave regardless of local police programs and strategies. In fact, this data prove that the conventional political arguments over crime are simply wrongheaded.
It is important to understand that murder and assault (the biggest category of violent crimes) are overwhelmingly emotional, not logical, acts. As Jens Ludwig writes in the recent book Unforgiving Places, maybe 80 or 90 percent of these acts are not planned. Reality is totally different from murder mysteries and crime shows. It seems reasonable that COVID increased perpetrators' psychological and mental health problems, and perhaps created increased opportunities for interpersonal friction, that caused murder and to a lesser extent violent crime to rise. And the end of COVID had the opposite effect. But, again, nobody really knows why crime rates increase or decrease and we should all be ultra-modest about our guesses.
Your first paragraph is very confident in saying what you believe had been solidly disproven, then you express humility about attribution in the second paragraph.
If there are a plethora of positive and negative inputs affecting the outcome, and many of them change, it becomes harder to so confidently attribute it to a given factor, or to rule out a given factor.
> "we can say with confidence that murder/violent crime does not go down because of more police because crime diminished with fewer police. "
Let's parse that a bit more carefully. We could say that the statistics show that murder rates (and likely other crime rates) can potentially go down even when law enforcement staffing is also down; that demonstrates that there are other factors in play besides law enforcment staffing levels (etc), and that these other factors can at least sometimes dominate the dynamics. Staffing levels do not control crime levels, but we haven't shown that they have no effect.
That is, the stats do not rule out a hypothesis that "crime levels would have gone down even more in this period if law enforcement staffing had been robust while all other factors stayed the same as the real world".
It would rule out a hypothesis that "crime levels will always go up when police staffing is reduced, regardless of any combination of other factors" - but that's not really something any serious person would suggest.
I find it tentatively encouraging that "other factors" can in cases like this over-ride any positive or negative influence of law enforcement staffing levels - IF we can get a handle on those other factors, perhaps we can continue to reduce crime more than we could by increasing staffing. Maybe there is a lever to use, if we can find it. But we have not proven that law enforcement staffing has NO effect.
I think the humility you describe is appropriate.
Any explanation for the climb in homicide rate in most places prior to Covid? Even the first 3 months of 2020 were elevated.
Per CDC data, total murders for the first 3 months of 2020 were up 10% compared to the first months of 2019, which is substantial though not huge. (This is to answer a question I asked in a reply I since deleted.)
The last 3 months of 2019 were also elevated by roughly 10%.
Personally, I don't read a whole lot into it, because 10% increases across a few months seem to be within the range of noise. This one continued a bit longer, but to me it doesn't necessarily suggest that anything notable had happened.
It’s probably that cops stopped being on de facto strike after George Floyd. Isn’t there a well documented effect of cops withdrawing from work in response to police protests that is associated with an increase in crime?
Isn’t Ludwig’s book describing what was called “broken windows theory” in the ‘90s? It’s interesting because a)obviously, stopping people from jumping subway turnstiles does nothing to address systemic issues but also b)it seems like enforcing the low-level stuff, which is a much easier lift than eradicating poverty, actually does help prevent more serious crime.
This article to me is a giant facepalm. It is basically the author presenting as conclusion what he thinks to be the causes, then selectively inserting pieces of facts to support his pov. The analysis process is thus devoid of deductive reasoning. That said, I can't fault the author too much because many analyses on violent crimes always have one giant blind spot. The blind spot is that the vast majority of violent crimes are committed by men. Once that blind spot is addressed, it would be much easier to find the answer. I'm not saying this to "blame men". I'm bringing this up as a demographic element to begin deductive reasoning.
You deduct why murder rate has declined by first isolating the demographic of perpetrators to men. From there, you break down the demographic further to young men. Then what kind of young men? What age range? Does class have anything to do with it? What about geographic locations? Does any type of murder have a bigger decrease over others? After breaking down the demographic to the specific type of men who committed murders, we can ask the question: What changed for these men in the last 4 years? Is it economic? Is it changes in environment? Was the spike in murders an exception and not a norm committed by men of a certain age range who have since grown older? Etc, etc.
For people trying to answer questions about violent crimes, you will come to more convincing conclusions why murder rates have declined only when you take away that giant blind spot. Otherwise, you're infusing too many variables about non-applicable demographics into the equation and your entire analysis gets muddled.
Just had a very long debate with someone about this. I think in order to understand why it went down we need to understand why it went up. If you look at most of the crime data during the spike, it was young black males not middle aged people trying to feed their family. Why would it go up then? Well in the most simple sense a general sense of chaos during that time. Covid plus the Floyd era. Post Floyd there was a general sense that you could wild out and not get in trouble. How many videos of looters did we see with no consequences for them? Also a huge part of this is the pandemic unemployment assistance fraud going on. Young people suddenly had access to thousands of dollars without really having to prove that they were searching for a job. What can you do with that money? Wild out. Especially in the city I’m from and general what I saw nationwide was this was gang related stuff. Lil z killed lil y because of a diss track and a perceived sense of disrespect. It wasn’t because the local back pack program shutdown for 2 years. The fact that the combination of lawlessness post Floyd and flexing money from PUA is not discussed more by criminologists is sort of baffling to me. Because this is clear as day to me all other stuff is mostly noise.
I'd encourage Jeff and co. to study the case study of Baltimore, where I lived for many years. Compared to some other cities, violent crime didn't come down as much there between the 1990's and 2014, though the rate did drop somewhat. But, the violence surged in Baltimore well before the onset of COVID, starting with the tragic death of Freddie Gray in police custody, and stayed high, until 2023, when it began to drop dramatically. In the last 3 years before COVID, though, the city government and some community- based violence intervention groups did achieve significant results (e.g., in the Cherry Hill neighborhood). And Mayor Scott has pulled together a web (various levels of government, non-profits, community groups) that's focused on crime reduction, and surely has contributed to the notable improvements. So, take a good look at Baltimore. With regard to the national trends, contrary to Trumpian rhetoric, 2023 and 2024 were very good years for the national economy: very low unemployment, falling inflation, rising wages at the low end of the scale, lots of public construction jobs, etc. Young people who might have been unemployed in earlier years were driving Amazon delivery trucks or holding up paddles (stop, proceed with caution, signs) at bridge reconstruction sites. I hope that we don't regress in the next few years!
Lehman makes the point that police departments, saddled with reduced staffing and increased work loads, seem to be prioritizing murders over all other crimes.
I would note that the NCVS (which measures everything except homicides) has seen remarkable increases year on year even as the homicide rate drops.
It is possible there is a link with the creedal passion period (CPP) which was forecasted for 2013-27.
https://mikealexander.substack.com/p/cycles-of-radicalization
Last CPP 1963-78, crime peaked in 1980, fell a bit and rose to a second peak around 1990 before beginning its long-term decline, so figure a 7-year lag (using center of twin peaks). For the CPP before that (1913-27) crime began its long-term decline around 1933, so a 6-year lag. For the CPP before that (1863-77) we seen a general decline in arrest rates after 1860. In Britian, assuming they too had a CPP in the 1830's like here, a long-term decline began after 1842 IIRC.
So maybe. Specifically, I would say the current decline reflects the unwinding of the George Floyd bump, but since that was a CPP event, it's really an instance of the idea expressed above.
I will note that mass shootings, which I track as an empirical correlate with CPPs, declined dramatically in 2024 and seems to be continuing, which is consistent with a peak in such political unrest around 2020, consistent with Peter Turchin's 2012 prediction.
https://mikealexander.substack.com/p/where-are-we-going-politically-part#:~:text=and%20current%20CPPs.-,Table%202%20shows%20sequential%205%2Dyear%20totals%20for%20events%20and%20deaths,Instability,-related%20to%20secular
I wonder if all the deportations or threat of deportations from the Trump administration have contributed to the decline at all. At least as far as the 2025 data is concerned.
Still those are third world murder rates - or rather worse in many cases