But the issue is that people are unhappy with crime/economy/etc. The justifications they give for it are completely irrelevant. It's the job of policy makers to figure out how to make them happy. If they do policy based on facts, and the results don't make people happy, then they've failed at their job.
> They reward politicians who do things that align with their incorrect beliefs about reality, and punish politicians who do nothing and/or deny the “problem.”
The fact that they are unhappy is real. They might be wrong about the reasons, but they are unhappy. The customer does not know what they need. It is the job of the provider to find it for them.
> Is the problem politicians or voters?
The problem is democracy. If you support democracy, you're not really in a position to complain about the inevitable consequences of that system.
Wait - why did you not address corporate crime at all? That's arguably an equally important dimension of crime and stating "crime is down" without adjusting for corporate crimes seems quite inadequate.
Lol. No, I'm usually a subtle irony guy. But not in this case. Corporate crime results directly in massive morbidity and mortality. There's no particularly good reason to exclude it when you're talking about crime rates.
Also, that's not what right wing reactionaries are referring to when they claim that "crime is up." Nor do they support the re-regulation that would be required to suppress white collar fraud.
Well, that's a reason. But I think part of the best reaction to RW reactionaries is to discuss the selectivity in how people consider the societal harms from "crime."
Yes on the first sentence, but you're wrong on the second. It's not the 80s or even the Tea Party era anymore. The current wave of RW reactionaries are all over social media complaining about banks and institutional investors, sounding as much like Elizabeth Warren as anything. There's enough historical animus over taxes and some Obama-era abuses that they still hate IRS agents, but there's certainly a strange new respect for FTC and anti-trust enforcement. Josh Hawley comes to mind as a model of this. I think lots of current RW reactionaries would be happy to see more law enforcement tools in that arena, provided that they were going to be operating solely against Blackrock et al, but they would be afraid that when Dems got back into power it would be used to persecute the pillow guy and a handful of random Trump-loving boat dealers in Alabama.
RW reactionaries basically discount massive amounts of corporate crime (Trump talking bribes, Elon gaming regulations, etc.) even as they obsess over one white girl dying in a car driven by someone without documents. I don't think your description holds up.
In theory, yes, there's some right-wing populism that complains about banks and tech companies and monopolies. In practice, I don't see Republicans doing anything substantial about it, at least not right now. Trump will throw away any principle at all to "make a deal" with anyone who sucks up to him enough, which in practice means that those companies will continue to get away with whatever they want as long as they pay him off.
But even putting Trump specifically aside, Republicans' proposed "pro-American-worker" policies are largely beneficial to those same corporations. Tariffs, repeals on environmental regulations or labor laws. These things have *some* benefit to workers, but the overall net effect is probably negative to regular people and positive to large companies. I don't see Republicans actually engaging in meaningful antitrust action either.
Most of the Reaganite/market above all conservatives are either very old or jumped ship to the Democrats a while ago. Trump is sort of a bridge between the old GOP establishment and the rising populist right that is more open to questioning the neoliberal economic order. I think future rightist politicians will be more amenable to at the very least anti-trust and economic protectionism.
First, maybe so, but that would need to be established. In a real world of limited resources, policies to address the one don't exist independently of policies to address the other. Second, this was a treatment of "crime rates," not a treatment of policy options.
Another reason to treat it separately is that it's very difficult to condense into a rate. You very quickly run into the "is this one crime with ten thousand victims or ten thousand crimes" problem.
Murder, tax evasion, software piracy, and dumping toxic waste in the river are all bad things we should want less of, but they are very different bad things. Their impact and the sorts of things you might do to prevent them are all very different, and it is hard for me to see how glopping them all together helps us think clearly about anything. To decrease the murder rate, probably you want to arrest/keep in prison the most violent young men who are committing other crimes, make it hard for criminals to have guns, hire more police to patrol the streets, stuff like that. To decrease tax evasion, you probably want to hire more IRS agents, and maybe revise the tax code to be less complex. To decrease toxic waste dumping you want more environmental monitoring and more people working for EPA/state equivalents. And so on. The solutions to one of these problems have nothing to do with solutions to the others.
At best "corporate crime" is a way too broad category for this discussion.
Some of it increases morbidity and mortality, but the category also includes accounting irregularities and and uncountable other bureaucratic missteps.
> Corporate crime results directly in massive morbidity and mortality.
Maybe I'm dense, but how? Like the most direct example I can think of, health insurance companies killing sick children and grandmothers for profit by denying coverage / care, which they do to the tunes of tens of thousands of people every year, is legal.
Selling people junk and fast food so 80% of people are overweight (and have ~2x all cause mortality?), also legal.
What corporate crimes that are actually illegal directly result in massive morbidity and mortality?
There's ~5k workplace deaths per year, about half the number that are murdered every year, and roughly 10% of the number that die in car accidents every year. And that's assuming all those deaths are due to corporate malfeasance, when it's probably something like 500 of the deaths.
If we focused on legislatively making it easier and faster to roll out self driving cars, we'd move the deaths needle 50 fold more.
If we passed out free Ozempic to everybody, it would halve mortality in the ~50% of the population who are obese and over 40, and we'd save 600k lives per year.
Way bigger interventions are available before going after that 500 people.
Would you consider examples like the Volkswagen Diesel-scandal? They knowingly and actively cloaked their car's real emissions from authorities. The death count is much higher, but also not directly traceable to their source on an individual level. But epidemiologically, it is undisputed that it increases mortality.
In other cases, there is a lot of "risk-skirting", where management encourages risky behavior that is not per se criminal, but can result in e.g. ecological catastrophes that outright kill people directly or indirectly. For example, Boeing and MCAS (Boeing pleaded guilty in a favorable deal), or various fossil fuel drilling and pipeline companies. Would it be fair to include this or not? And why?
Those are just examples and weren't meant to be anywhere close to comprehensive. And please, seriously, you're going to downplay the magnitude of the effects of the crimes committed in relation to the opioid crisis?
I mean, that goes into definitional issues: the people Scott was answering to (like the neoreactionaries) definitely don't mean corporate crime, they mean the kind of crime people experience directly and first hand. Their general argument is that the liberals are too soft and worried about looking racist to address the real causes of such crime.
I think it's important to point out the selectivity in how those people address the impact of crime. Just answering on their defined term, without pointing out the embedded biases, misses a large element in the societal impact of crime.
True, but also, these different layers affect different levels of survival instinct and Maslow hierarchy. If an exec embezzles money or plays recklessly with investor money, I can lose my savings or pension, which is awful and hurts my future, but in a long term perspective. If a mugger stabs me or a rapists assaults me (for a woman), I get immediate pain and danger, possibly death or permanent damage, and trauma that has been wired into our minds by evolution because it's pretty much THE definitional fight-or-flight kind of situation.
Basically I don't think it's weird that people weigh this kind of crime as something they perceive more dramatically close to home. The utilitarian perspective is "which does more harm at scale", and that's the rich finance guy. But the personal perspective is "which does more harm to ME if I am unlucky enough to fall victim to it, and which is scarier", and that's the petty crime. Another common moral consideration tends to be the fact that the rich guy has more moral choice - he's more educated and has less needs, he does crime only out of greed, whereas the thief or robber may be poor or mentally ill or from a background where he was never taught that crime is wrong. And that may also be true, but ultimately, getting hurt is getting hurt, I can only care so much about what specific excuses the culprit had.
I'm not saying it's "weird." It aligns with pretty standard cognitive biases like availability bias, temporal discounting, proximity bias, intentionality bias, etc. I'm arguing that it doesn't necessarily reflect the actual harm at the societal or even individual levels.
What's the reason you think considering a form of crime that causes great amounts of societal harm shouldn't even be considered as part of a thorough and comprehensive discussion of crime rates?
Nothing at all, just not in response to an article that's literally about how if you say crime is down people will find some type of crime is not down and then generalize it to say crime is not down
Sure. I get that within that narrowed context of responding to particular claims, the distinction isn't directly on point. I still think it's important, however, to point out the larger biases embedded in claims people make about crime rates.
Lol. Funny. But workplace injury resulting from corporate malfeasance are real. Environment and public health harms resulting from corporate malfeasance are real. Product safety harms from corporate malfeasance are real. And then there are indirect effects from financial crimes and fraud. It is a deliberate choice to exclude these factors when describing crime rates. What's your rationale for doing so?
Bit of an ironic response. I just asked a question, I didn't put any words into your mouth! But implications do exist, don't they...
I was just trying to see where you were coming from. Whether it was a starting place of:
"We need to do dangerous jobs to make the items we all depend on. Sometimes people will get hurt. But, evil management can sometimes exacerbate things so that people get hurt at a much higher rate than is needed"
or
"Everything should be able to pop into existence, ready-made, with no danger to human life or limb at any step of the process. But evil management wants to make things in a life-threatening way because they are so evil"
I am glad you are *probably* starting from the first one, even though my clarifying question failed to induce a clarifying response.
"I just asked a question, I didn't put any words into your mouth!"
Not a fair question. It was loaded. It's implying he (might) think that.
What he said (in the first post) is implying there are other sources of "workplace injuries". (Which makes your question, as written, pointless.)
=============
It looks like the "place he's coming from" is "corporate malfeasance isn't some abstract thing. It can actually have an understandable affect on 'regular Joe' by sometimes resulting in workplace injuries".
There's no argument to be made that workplace injuries occur at a higher rate now than in the past. Ditto for environmental impacts. Not sure about financial crimes.
OSHA and the EPA will have data on both these topics. I would guess they all kinda generally go in the same direction, but if you think I'm wrong, by all means....
Lol. Investigate it as you will. I haven't made a claim about trends either way. My point is that trends of more categories of crime should be addressed under a thorough and comprehensive treatment of crime rates.
Trump admin. has moved to end EIS/NEPA so “data” will have to have an asterisk going forward.
I mean, Musk is being granted a national wildlife refuge, home of listed species, that was initially purchased by the Nature Comservancy. This is incredible.
Never Trust the Government. I mean, you do realize that, right? They have a job to do, and it's not "preserve the ecosystem" except when DARPA says it is (and then, you're looking at the Department of Deer Warfare).
Violent crime creates larger second-order effects (fear-driven behavior changes, depression of property values, erosion of social trust, security costs, etc) than white collar crime. Most white collar crime seems avoidable because a) it doesn't generally target individuals and b) you have to first choose to do business with most white collar criminals. I have no fear that the next Bernie Madoff is going to bother me, for example. Violent criminals on the other hand will bust down the door to your home. That generates much more fear and avoidant behavior in people.
The other consideration is that white collar crime exists on a more-or-less continuous spectrum with normal net-social-positive business transactions. The distinction is often one of perspective and of somewhat arbitrary laws about what is and isn't legal: it's not like most white collar criminals stab people. That means that there's no clear way to crack down on white collar crime without exerting some nonzero chilling effect on regular business activity and so people shrug and accept most forms of white collar crime as the cost of doing business. Obviously there are exceptions like SBF and Madoff, but I suspect it's actually a rational attitude most of the time.
I also suspect that violent crime is more strongly auto-catalyzing. Its successes are public (a dead body, a robbed store, a smashed window) which signals to the marginal would-be criminal that maybe they could get away with it too. White collar crime is generally only apparent when someone gets caught and therefore has built-in anti-mimetic properties.
Well, the article wasn't meant to address all things that impact societal safety, health and well--being. It was meant to addess the things like murder, vandalism and mugging which some people call "crime".
Would it be useful to include wage theft? Probably not. The people pointing to high "crime" rates usually are advocating for policies like increased police presence. The rate of wage theft would not be affected by that policy so isn't relevant to the discussion
Sure. But that seems circular to me. I think there are biases embedded in how people separate out different forms of crime, even forms that directly (and indirectly) affect safety and health on a massive scale. We make choices in how we spend limited resources to address societal harms. The choices we make in the public discussion of crime rates, where an entire category of crime isn't even included, arguably leads to a misallocation of resources when we consider rate of return.
I think you're doing something similar by using the shared label of "crime" to complain that people focus more on mugging than wage theft.
It's reasonable to argue(like you are) that corporate malfeasance is more harmful than mugging. Just like it's reasonable to argue that things like agricultural policy (which isn't usually called a crime) are more harmful than both. But pointing out that we use "crime" to refer to both corporate malfeasance and mugging doesn't make your argument more compelling than a similar argument that agricultural policy causes more societal harm than mugging and should therefore get more focus.
Edit:
For example, the following things happened to me while I lived in North Seattle:
1. Someone broke into my apartment and stole things
2. My job did not allow me to take mandated 15 minute breaks
3. I got hit by a car while biking to work and was hospitalized
All of these harms might have been prevented by policies like sweeping the nearby encampment (where I found my discarded backpack) or designing streets better. We typically only refer to the break-in as "crime" in these discussions while advocating for policies like sweeps. Your argument is that we should also include the 2nd harm(but probably not the 3rd) when we talk about ccrime. I don't think that's particularly helpful.
When companies falsify safety data, cut corners on environmental protections, ignore workplace hazards, or market unsafe products, the result can be toxic exposures, injuries, chronic disease, and even premature death. These harms are every bit as real as those caused by street crime, they just show up through polluted air, unsafe working conditions, defective products, or contaminated water rather than a single dramatic incident.
How do people spend their individual resources in response to crime? Economists look to that sort of thing when measuring the costs of crime and asking how much people would be willing to pay to avoid it.
"The people pointing to high "crime" rates usually are advocating for policies like increased police presence."
Really? I did not get that impression. What I mostly saw was people complaining about "Well now I stay in at night and lock my doors, so all the muggings I've avoided don't show up in statistics. We've traded high crime rates for a lower quality of life". They want to go out at night and not get mugged.
Similarly, they (presumably) want low crime without there being cops stationed on every corner.
Now, the commenters pointing to high "crime" rates have not been all that forthcoming with what kinds of policies they have in mind. (See my question to Steve Sailer way back in Scott's first post about crime rates like a week ago. Sailer never responded to it.) Personally I suspect they want a policy in which some kind of unpleasantness is continually meted out on members of their outgroup, who they blame the crime on.
I just gave you my personal suspicion. It would be easy for anyone to deny that it is true and impossible for me to prove it, so to give you any more details is kinda pointless.
But I think when people fantasize in unarticulated ways about unpleasantness being meted out on members of their outgroup, even if it would be cops (or some other special law enforcement agency) doing the meting-out if those fantasies were brought to fruition, that doesn't look the same in people's minds as cops standing around deterring criminals with their mere presence but also reminding non-criminals of the nearby threat from criminals.
People are complicated and brimming with cognitive dissonance too.
If Steve Sailer were king, what would he do about this?
I replied:
Steve Sailer
5d
We've had three periods in my lifetime in which liberals held the upper hand in criminal justice policy: the Warren Court Era, the Ferguson Effect, and the Floyd Effect. All three led to explosions in crime, especially black on black murder.
So, first thing: Let's learn from that unfortunate experience and Not Do It Again.
Second thing: We've seen New York City radically reduce its murder rate by focusing on point-of-use gun control to change the culture of criminals: make lowlifes more afraid of the cops than of each other, so they leave their illegal hand guns at home.
It shouldn't be that hard to learn from the world's most famous city.
Jeez, you're right, I'm sorry. Somehow I didn't get the notification. I retract what I said about you not replying to that question.
But I do see I was correct in a sense, that you want your outgroups punished in some way (in this case, not allowing liberals to hold the upper hand in criminal justice policy). I note also this is a different outcome than increased police presence.
I thought NYC had general restrictions on gun ownership with very few exceptions.
There's a very simple test about which "type of crime" is has "worse" impacts: which sort of criminals would you rather you and your family live near?
If you have to live on Street A, which is home to, in order, a rapist, a murderer, a carjacker, a serial duelist, and a burglar, or Street B, which is home to an embezzler, a wage thief, a tax fraud, a corrupt politician, and a digital pirate, on which street do you think your physical safety and your mental peace is more likely to come under direct adrenaline-spiking possibly-violent threat in the future?
I phrased the test without the right amount of delicacy; my quotation marks around "worse" were meant to communicate "more urgent and pressing impacts to your physical safety". I concur that in the aggregate most white collar criminals directly and indirectly cause more "harm" to society.
When people talk about "crime", they almost always mean "urgent and pressing impacts to my physical safety". White-collar crime indirectly threatens the prosperity and safety of your society in often difficult-to-quantify ways; "blue-collar" crime threatens to delete you tomorrow on your way to the corner store. Of course it's reasonable to be more finely attuned to the latter type.
Living next to someone is one vector for measuring the impact of crime. There are others. IOW, which society would you rather live in, one with a good environment and relatively little poverty and some property crime, or one with a degraded environment and high rates of poverty and a police state that prevents property crime?
I think living besides our prime minister (you probably heard his name, although our country is a smallish spot even on the map of Europe) might be quite OK-ish, still, his actions wreck the whole damn country.
Dictators with really bloody backgrounds can be quite charming in private, as well.
They don't impact physical safety. They're basically like taxes or bureaucracy. If I get paid less because someone in my company is embezzling a ton of money vs it's getting wasted on pointless bureaucracy, it makes no difference to me. But if someone mugs me, I'm worried about a lot more than money.
Something I find a bit annoying about this whole conversation is that personally, I rarely if ever hear people complain about "crime". "Crime" is not a terribly meaningful category.
Next Door will have plenty of posts, if you care to look. Most people are complaining about a particular crime (porch pirates that are continual and not caught are a favorite)... but yes, people do complain about "crime in general."
And if they do, it's usually clear from context what subset of crimes they're talking about.
"I don't go downtown any more, there's too much crime" -- clearly they're talking about the sorts of crimes of which they are likely to be a victim if they go downtown.
That’s difficult to analyse. I would expect that: different size corporations make different size crime; distribution of sizes of corporations (including “number” of corporations) changed significantly over time; regulation changes significantly over time (and therefore the type of crime committed). For example, securities fraud today is different from 50 years ago. An analysis would need a very different technique than just looking at crime rates.
But that relates to cognitive biases. Yes, the threat of some forms of crime is more concrete and tangible. But that doesn't actually translate directly into it being less of a threat to health or safety. That's precisely the reason why corporate crime shouldn't just be excluded from a discussion of crime rates: to address the underlying biases. When we do so, we run the threat of irrationally addressing threat levels.
it is apples and oranges. this argument is about "if the apples are rotten", you are claiming that in fact "oranges are rotten", this being or not being the case has no bearing on the claims being made about apples, even if both are fruits.
Street crime and white collar crime generate differing types of threat, illicit different psychological reactions, and require completely different policies to deal with. they are only the same insofar as both cause social harm and break the law in some way.
My point is that (imo) as a society, we direct our attention towards different categories of crime in a way that is disproportionate to the extent to which those different category of crime impact health and safety. That's what I mean by bias, as in other biases where attentional focus isn't purely "rational."
But I agree, disagreeing about the relative importance doesn't inherently mean "bias."
I think you've made this claim in many different words throughout this thread; that corporate crime should not be excluded from crime rate discussions. However, I don't feel you've really supported that claim with any data or evidence and have only supported it with "vibes".
It's an opinion, that when discussing rates of crime, the definition of crime should be inclusive of a form of crime that has massive societal impact. I'm not sure how one would provide "data" to support that opinion. Do you mean data that corporate crime has massive impact? Do you actually believe that it doesn't, and need to data to believe it so?
you started this with "Wait - why did you not address corporate crime at all? That's arguably an equally important dimension of crime and stating "crime is down" without adjusting for corporate crimes seems quite inadequate."
so argue that it is an equally important dimension of crime in way that is not just "because Joshua Brooks thinks so". It doesn't have to be data driven but at least a logical argument stating expected consequences by not addressing it as a dimension. or whatever it is that is driving your statement
When companies falsify safety data, cut corners on environmental protections, ignore workplace hazards, or market unsafe products, the result can be toxic exposures, injuries, chronic disease, and even premature death. These harms are every bit as real as those caused by street crime—they just show up through polluted air, unsafe working conditions, defective products, or contaminated water rather than a single dramatic incident.
There was a huge explosion of low brow white collar crime during covid as the government did helicopter drops of money thru various programs such as paycheck protection loans and meals for kids, such as in Minneapolis. Lots of dubious NGOs got their hooves in a lot of troughs.
On the other hand, throwing all that cash around with little oversight may have headed off a recession.
Different industries and companies have different rates of compliance with occupational safety and environmental laws, and you can choose whether to work for a company that cuts corners on safety or near a company that is known to get away with pollution. Different states have very different levels of enforcement of laws involving environmental damage, fraud, and the like, and people can move from one state to another.
I expect this happens all the time, but there probably aren't statistics on it.
This place is a hive of billionaire defenders and guys who love capitalism much more than Adam Smith ever did, so I predict you will get a chilly reception.
You are talking to the wrong audience. They are far less likely to be victims of any type of crime, but right now they're far more likely to be victimized by thugs rather than corporations violating workplace safety laws. Why should they care what happens to those below them?
Corporate crime is also less "sexy" to talk about, harder to see, and it isn't as clear whether the people to blame are in this readership's (and Scott's) outgroup, so you're less likely to see it discussed here.
While we're at it, why is "wage theft" frequently defined as paying less than the minimum wage (when did this start?) while the less frequent but possibly more serious crime of employers not paying what they promised gets ignored?
You are a great illustration of the Streetlight Fallacy. Scott says crime is down. You reply by saying, "What about corporate crime? That matters too." You might as well have said, "But war in the Ukraine has increased." Nobody worries about corporte crime, whatever that is. Crime means burglary, robbery, larceny, ... So you are changing the issue.
Depends. I mean when you're watching neighbors stringing up their fellow Ukrainians, and calling it "catching Russian Spy"... that's lynching* by any other name.
*description deliberately changed to a more American-centric referent.
This is a good point. Also very interesting to think about. I bet if you ask Average Joe On The Street if he thinks corporate crime is rampant and out of control, he'll say yes. But he'll say it in a shrugging "whaddaya gonna do" way, as if corporate crime is something that happens on another planet. Contrast with how people talk about street crime, you can almost see the torches and pitchforks ready to come out--even if they have never been the victim of it or even witnessed it.
Homicide and drug overdoses (in absolute numbers and rates) are way down.
I’m a life actuary and I’m not going by “crime stats” but by CDC numbers, because they’re not waiting for adjudication (and it’s tough hiding dead bodies… it can be done, it’s just a lot of work. Only a few overlooked/hidden bodies show up each year.
I’m not trying to make a point about crime, other than various cities trying to claim their great policing did it, because it happened everywhere over 2024 and 2025. (A little bit in 2024, and a lot more in 2025). I doubt all these cities had great policing improvements simultaneously.
>by CDC numbers, because they’re not waiting for adjudication
Does that mean those CDC numbers can't distinguish between murder and manslaughter and you are using "homicide" colloquially? I think that's an important distinction.
The distinction matters for an individual homicide, but unless you think the *mix* has changed a lot then for overall statistical trends it won't matter.
That's what I don't know and why I'm asking for clarification.
I believe that the distinction between murder and manslaughter matters for the question of what effect improvements in healthcare have on the murder rate. Murder implies planning and forethought that includes the knowledge of the state of healthcare, meaning a would-be murderer is going to put in more effort if they know healthcare for their victim is good and available. All else being equal, I'd expect improving healthcare to reduce manslaughter rates more than murder rates, shifting the mix towards murder.
Here are your choices for “external causes of death”:
-caused by someone else, not accidental (I simplify this to “homicide”)
- self-caused, not accidental (suicide)
- Accidents (this is a huge category, including drug overdoses, motor vehicle accidents, falls, drowning, etc)
The data are aggregated from the state level, based on death certificates and subsequent determination (external causes are censored from the public for 6 months in CDC data).
CDC death data are separate from FBI, etc, reporting, which is useful for checking.
Is there some sort of official technical definition of "homicide" that excludes manslaughter? The definitions I'm finding on Google (both formal and informal) all seem to include it.
If someone kills me dead, I wouldn't care what their exact mens rea was. Super from the fact that I'd be physically unable to, the consequence it's the same.
Drug overdoses are absolutely not "down" in a meaningful historical sense. Vs. the all-time highs a few years back? Sure they're down a little. But vs. overdose deaths in the 70s/80s? Still way, way higher: https://www.cdc.gov/mmwr/preview/mmwrhtml/mm6346a19.htm
Notably Scott did not mention overdose deaths at all, even though they are a pretty reasonable proxy for drug use (in the same way murder is a reasonable proxy for other violent crime). I don't think it is "malicious" per se but this difference stands in contrast with the other metrics.
They are down year-over-year after the huge spike up in 2020 and continuing rise in 2021
Yes, the long-term hideous trend of decades is pretty bad, but we barely had any turnaround, especially once fentanyl was in the mix.
Now, again, there was some noise made about widespread availability of Narcan helping with the recent reversal from the OD spikes of 2020/2021… again, don’t think that’s what is going on.
I think the illicit fentanyl supply has been reduced by quite a lot (starting in 2024), and that probably has a lot of knock-on effects. Perhaps some other stuff has been going on as well, but again, I don’t think it’s been due to great policing or drug abuse treatment programs. Not for such a large effect.
The Chinese used to manufacture fentanyl and just mail it to the USA.
The Chinese government cracked down on that a number of years ago, so Chinese manufacturers then made the raw ingredients for fentanyl and sold them to Mexicans, who mixed the final products and smuggled them into the USA.
Lately, it appears that the Chinese government has cracked down on this recent two stage method, so until some criminal perfects a work-around like last time, fentanyl deaths seem down.
Mary Pat and I like talking about the numbers, while most commenters want to talk about how to think rationally without dirtying their hands with any actual historical data.
I've been reading Steve for a really long time, so I know he's been going back to some of these external causes of death, even pre-pandemic effects. (aka, "Ferguson effect" stuff)
I'm a life actuary and have been following U.S. mortality trends for decades in my career, and I'm interested in multiple trends. When I saw the huge spike in homicides/drug ODs in 2020, on top of the COVID deaths, I was pissed. (Oh, and alcohol-related deaths also increased, as well as motor vehicle accident deaths... and some other stuff.) I follow lots of stuff... I work in the insurance industry, so yeah, I'm kinda invested in death.
I have a theory that some of the fentanyl precursors could have been making trouble in China internally, not just through Mexico/Canada routes to the U.S. Something seems to have gotten shut down by the CCP, I think, and not necessarily due to anything the U.S. is doing.
People get so tied up in U.S. politics that they think that whatever they care about must be the driver of whatever they're seeing, as opposed to looking more broadly at potential cause-effect chains.
I've heard some wind about the term life insurance industry being endangered, post the covid19 vaccinations*. You heard anything, working in the field?
*significant increases in deaths of the younger varietal (not preterm).
That's U.S. general population mortality, monitored by the Society of Actuaries (I'm a member, and I've been involved in past mortality studies there as well) -- the life insurance experience studies can be more difficult to come by if you aren't in the industry.
I'm not convinced that drug overdose deaths are a reasonable proxy for drug use per se; isn't the crisis almost entirely related to the advent of much more dangerous drugs than were available before?
That said, insofar as drug overdose deaths are the biggest actual impact of drug use, they may be a good proxy for what we actually care about when we talk about "drug use".
Yeah I think that's another thing to keep in mind - when people say "drug use" in the context of crime/disorder they don't mean high schoolers smoking a bong in the basement when mom and dad are out of town. They mean someone in downtown SF with untreated mental illness passed out with a heroin needle in their hand.
Right. I too encourage people to look at CDC homicide numbers, which come from doctors and coroners, while the demographics come from funeral directors going over a quick checklist with loved ones. There is no political pressure on undertakers to fiddle with the race or ethnicity data of homicide or car crash victims.
There is more political pressure on FBI murder stats, which are collected from local law enforcement agencies. But, in general, CDC homicides and FBI murders correlate closely together, suggesting that dead bodies get counted pretty honestly in this country.
Murders in the United States exploded during the last weekend of May, 2020 due to the triumph of the Black Lives Matter movement over the police following the death of George Floyd on May 25, 2020. For example, the most murders in the storied history of Chicago murder was May 31, 2020 with 18.
In June 2020, 44% more blacks across the U.S. died by homicide than in June 20219.
Traffic fatalities exploded during the month of June 2020 as drivers noticed that the cops weren't pulling people over as much, much less searching them for illegal handguns. In June 2020, 53% more blacks died by motor vehicle accident than in June 2019.
Eventually, American elites started to figure out that Black Lives Matter was losing popularity -- I started to notice less propaganda in the second half of 2020. Murders and car crashes started to fade consistently and murders dropped like a rock after Trump's election in November 2024 signaled the end of the second BLM era.
When people talk about whether crime is up or down, what they typically want to do is make an argument over whether the BLM Era of 2014-2023 was a good idea or not. People who emphasize that crime is down in February 2026 vs. June 2020 or 1991 or some other cherrypicked starting date typically want to say: "Don't Blame BLM. It Was a Good Idea. We Should Do It Again In a While."
People who say crime is up are typically arguing that American elites made a huge blunder by falling for BLM.
I, for one, am of the opinion that the 2020 were stupid and counterproductive in a lot of ways, but they probably did not cause the murder increase. That was a general increase in homicide that started months before George Floyd died and was clearly downstream of COVID somehow, and then, yes, an actual "Floyd Effect" in the Summer but that was almost certainly a "viral video" effect, people reacting to the video in ways that made them more likely to commit murder.
The most obvious tell is the shot spotter data, which yes, shows a big spike in murders immediately after Floyd's death - immediately as in a few *days* after. It's absurd to think that police pulled back and criminals noticed and starting murdering people in large numbers in only a few days. Even your example of May 31, only 6 days after Floyd's death, is incredibly implausible as any sort of "protest effect" if you put any serious thought into it.
Which, of course, easily explains why we don't see any such effects in Los Angles homicide rates after the 1992 riots. Internet video sharing as we know it did not exist in 1992. And easily explains the lack of a similar international effect, we wouldn't expect a bunch of, for example, Chavs or Yardies in the UK to particularly care about what happened to some Yank in Minneapolis.
Many poisons don't actually show up in a toxicology report. I'm not sure I'd say you're getting good data from the CDC on anything except blunt trauma, bullets and knives. The Darwin Awards are often a good example of "what people will do if you point a gun at them" -- "the easy way or the entertaining way?"
People subjectively feel that crime is up, falsely, as you have illustrated (and as they basically always have when surveyed on it, as best I understand it). And they also feel that disorder is up.
You argued (accurately) that crime is way down. Some people refuse to agree. Others shift to "well, disorder is up". Has anybody even established that the latter is true (thus making this a motte and bailey) or is this just retreating to an unfalsifiable claim? Why should I accept that people aren't just as wrong about this?
One of my favorite comments from that discussion was the guy in Honolulu saying one of people's main complaints is crime. Honolulu is the safest city (looking at the homicide rate specifically) in the 69 most populous US cities (i.e., over 300k people). Their homicide rate is less than half the national average. Crime more broadly is also low there. I went there a few years ago and at least my best guess is that "disorder" is also much lower than the major city I was living in at the time. And yet they are (allegedly) seized with the same panic as everywhere else.
"Why should I accept that people aren't just as wrong about this?"
I think it is reasonable to just believe that people are wrong about this. I do that.
But I think there is also value in trying to figure out WHY people are wrong about this. They may have something in mind that they have a difficult time articulating. For example, I believe this is true for people who prefer the "warmth" of analog records vs digital music such as streaming and CDs. No, the vinyl record isn't qualitatively better in any measurable (for most people) way. But there is something to the preference.
If we care about what might be bugging people who think crime is up then it makes sense try to figure out what it is that is bothering them. We may then decide we don't care ("I'm upset because the *other* team is winning politically ...") but maybe we will care at least a little bit.
I think that using an easily measured and understood phenomenon as a proxy for something else that seems psychologically related is pretty widespread. People still think that inflation is out of control. It isn't crime or disorder, it's people feeling defensive and territorial because the future seems uncertain and other people seem to be getting aggressive about competing for scare resources. But that's complicated and messy to think about. Instead, most people are just saying to themselves (not necessarily fully consciously) "I'm feeling socially anxious, let's look around and find a reason why. Oh, crime, of course." I think we all do this. I could cite research on it if anyone wanted me to.
There's something called the "Two Factor Theory of Emotion" which states that emotional reactions start as a physiological response (the heartrate increases, bloodpressure rises, muscles tighten, etc.) and then the mind starts looking around for a reason in the environment. We have a strong tendency to attribute our reactions to the first thing that comes into view. Someone once tested this theory by placing an attractive female survey taker in the middle of two different bridges: one rickety suspension bridge, and another that was solid cement. She stopped male hikers, and them answer some fake questions, and they gave them her phone number, telling them to call if they had any questions about the survey (of course, they actually called the researchers). The hikers were far more likely to call her if they had been on the rickety bridge--the only way to explain that is if they misattributed their physiological reaction to her instead of to the bridge.
See:
Schachter, S., & Singer, J. (1962). Cognitive, social, and physiological determinants of emotional state. Psychological Review, 69(5), 379–399. https://doi.org/10.1037/h0046234
Dutton, D. G., & Aron, A. P. (1974). Some evidence for heightened sexual attraction under conditions of high anxiety. Journal of Personality and Social Psychology, 30(4), 510–517. https://doi.org/10.1037/h0037031.
I don't think it's that deep. Bad-faith political actors on the internet have an incentive to spread around false narratives that serve their agenda. For some of them, one such narrative is that crime is increasing. Because most people form their beliefs not by rationally thinking about the evidence but by social diffusion from the people they trust, those who share a political tribe with the aforementioned bad-faith actors become convinced by them. So people don't believe the narrative because they're searching for some easy-to-understand proxy for their general sense of malaise. They believe it because their favorite political influencers told them so.
I think the Internet is a factor currently, but also people have been erroneously thinking crime is up for long enough that it can't be the main reason.
And it's usually the same except-in-my-neighborhood thing that makes me think it's not the "disorder" story. If people thought their lived experience of graffiti and crazy people = crime then they wouldn't think "crime is really bad, but not around here".
Note that the "but not in my neighborhood" has applied to how good the public schools are for a long time. It is common for folks to rank public schools in general lower than the public school their children are attending.
Yeah, I think this trend is really common. Also fits the "Vibecession" data pattern where people think the economy has gone to sh** even as they acknowledge their personal situation has improved.
Agreed, the general problem is that politics is the mind-killer and how exactly that manifests differs from time to time. In our time, the internet is the primary vector.
I haven't been able to find the poll, but I believe this phenomenon is true of economics as well: There's a big gap between what people say their financial situation is and what the think the overall economic health is, and the gap has grown over time.
Even if I assume people would answer the questions that way in that situation, I don't think that fits the questions where people also report rarely or never being concerned about being the victim of any particular type of crime.
A poll by Pew Research or whatever would compress people's conflicting sentiments about their varied experiences (e.g. between work and home, or home and travel) down into one thing, so it's interesting to consider which thing they would choose to express to a pollster and why they would choose that thing rather than the other thing.
But here, for example, where we can write basically as much as we want, people still tend to minimize one and focus on the other, and I think it's for exactly the reason(s) noted in the comments immediately above.
For the crime thing, though, it can also be that stuff is locked up in stores that wasn't locked up ten years ago (personal observation. I don't have data and thus this is an anecdote ... maybe I'm wrong about the stuff being locked up.).
"Crime is down a lot. That is why we have to lock up merchandise so that it won't be stolen," sounds like gaslighting to a lot of people. Who are you going to believe? Me (with my crime stats) or your lyin' eyes?
There are many explanations for why stores appear to be locking up certain items and not all of them involve increasing crime. The explanation that one buys into depends heavily on what their trusted pundits are telling them.
Moreover, the choice to focus one's attention on a particular piece of evidence, such as this one about stores, depends on the narratives that they believe or want to believe. For any hypothesis, there are always pieces of evidence to point to and say, "See!" and a clever pundit can pick and choose the evidence to make virtually any case seem convincing.
Because of this, I don't think it's so easy to disentangle "I feel that crime is up because I notice that items are locked up in stores" from "I notice that items are locked up in stores because I feel that crime is up."
You say "stores appear to be locking up certain items". The word "appear" is a tell that you are using skepticism to avid the obvious conclusion. They are locking up certain items. I can see it. So can everyone else. I asked the store manager, and she told me it was a corporate policy. Do you think I hallucinated, and just so happened to have the exact same hallucination as everyone else, including the people working at the store?
> There are many explanations for why stores appear to be locking up certain items and not all of them involve increasing crime.
There are many possible explanations for everything. That doesn't mean they are all equally likely, or that we can't get any knowledge of what is happening in the world. Are you this skeptical of all possible conclusions? Or, do you selectively choose to apply total skepticism of all knowledge when you don't like the conclusion you would otherwise reach?
When I go to the drug store, deodorant is locked in a plastic case. I need to ask an employee to get a key to open a box and watch me as I take an item from the shelf.
Am I a bad faith political actor for noticing this? Has CVS been tricked by bad faith political actors into thinking this was necessary?
Economically, there isn't any reason to prefer the change in price levels over the last week, the last year, or the last decade as the measure of inflation. Well, ideally we'd use the instantaneous rate of change but that's not measurable. The government, and hence the media, reports inflation in yearly rates by long convention but there's nothing fundamentally special about years here.
People have a hard time adjusting when a unit of measure changes, just they would if they were trying to figure out if they could lift something when the definition of a "pound" had recently changed.
Which isn't to say that restoring the price level to 2020 like people want wouldn't be a much worse cure than the disease. But I don't feel like the complaint is actually unreasonable.
>Well, ideally we'd use the instantaneous rate of change
This is trivially false. If we imagine the price of everything suddenly doubled in one day and inflation is 0 for the next four years, the negative impact of inflation would harm people throughout those entire four years.
The actual impact of inflation on an individual is the amount of purchasing power they lose as a result of it. However much money you have in the bank, you effectively lose the inflation percentage of that (100k in the bank *4% yearly inflation costs you $4k per year). However much your income is, you lose that much income (if not adjusted for in your raise). If you would have gotten a 2% raise, but your boss gives you a 3% raise instead to compensate for inflation, but inflation is actually 4%, you've effectively got a 1% pay cut, and 3% pay cut compared to the counterfactual case where you could have had a 2% raise with no inflation.
For each individual, if you had perfect information of what their income would have been in the counterfactual case there was no inflation, you could compute the total harm to them from inflation by integrating it across their lifespan. They were be permanently that much poorer. You can't just say "well, you lost $3k this year, what are you complaining about? Nevermind that you lost $50k last year, that was last year." The harm remains, but isn't realized until their savings (which have the same numerical value) don't go as far as they buy things.
People can easily tell they feel uncertain/anxious or like something is wrong. It's not always easy to figure out why. But inflation and crime get headlines, which means it's likely that people can grab on to them and then try to justify it as an explanation afterward. Human memory is unreliable and can even be rewritten to justify current beliefs.
I often think about inflation "being out of control", and the statistics that get used to disprove it
i imagine a scenario where the government prints a quintillion dollars in a single day, and for that year inflation is like +100,000% or whatever... but then maybe next year they print nothing, and there's no inflation
if i lived in that society, it would probably take me the rest of my life before I stopped claiming "inflation is out of control", and i would feel justified in making that claim even if there'd been 0% inflation for decades prior
I think I would not feel unjustified in saying that, unless prices actually decreased back down to the level they would have been at if inflation had been more like 5% annual
Technically, your concern isn't about inflation, per se, it's more about price stability. To the average consumer, that may be a distinction without much difference, but to an economist...
I don't mean to dogpile, but the inflation discourse frustrates me. It feels like it often goes like this:
"Inflation is out of control!"
"No it's not. High inflation happened years ago and it's back to reasonable levels."
"But prices increased massively and my wages didn't keep up! Why would I care about the current rate? The problem is affordability!"
"Yeah but the affordability problem happened years ago and hasn't had any sharp increases since then."
"So? My wages *still* haven't caught up to the price increases!"
"Oh, so you want deflation? That would actually be bad for the economy because blah blah blah."
The person pooh-poohing the problem isn't saying anything incorrect there (other than attributing a desire for deflation to the other speaker) but they're completely dismissing the correct things the other person is saying, just to feel smug. The affordability crisis isn't over. Perhaps there is nothing else any government can do about it now, but dismissing the problem out of hand because inflation is back to reasonable levels is crazy-making to the people still feeling the squeeze.
It isn't just to feel smug, though. Your two actors are using different definitions of "inflation" (or "affordability", or whatever). The second is referring to a population phenomenon: economic change is something that happens to populations of people, not single individuals. Meanwhile, the first person is using something like "Inflation is what happens to me!"
There is a sense in which inflation can happen to a subset of people within a larger population, but identifying exactly who those people are has proven surprisingly hard (even harder is convincing them that they are a minority).
> But I think there is also value in trying to figure out WHY people are wrong about this.
Sure, but I think this is well explained by a combination of sensationalism and polarization in traditional media, and general amplification of emotional content on social media.
As has been discussed here before, people mostly don't believe things based on careful study of the truth. They reason socially. If they hear about crime all the time, they think it's up. The news loves to broadcast about crime. Politicians love to attack their opponents over crime. You hear about dramatic stuff like murders even though murder is one of the rarest crimes.
You still have to account for changes over time. Fear of crime isn't a constant, it rises and declines. Those changes can't fully be accounted for by constants in human nature.
I'm not really sure what you're trying to say here. My explanation doesn't suggest people will always think the exact same thing, so much as that their thoughts are mostly driven by trends in media reports.
A lot of more-partisan sources will be inconsistent over time in how they highlight or downplay crime, based on who is in power or what policies are being discussed, though.
Isn't it obvious why people think crime is up? There are fewer white people now and more black and brown people.
It's a predictable consequence of a big demographic shift: different racial groups don't trust each other as much and don't feel as safe around one another. And there are existing stereotypes anout black and brown skinned folks being more associated with crime that even non-whites often absorb (eg. through television and media)
People are mostly reporting that crime is going up in the US but not near them and that they feel pretty safe. If they thought this based on how scary minorities are, presumably they would be more scared for themselves.
That's an interesting point, although I do notice the defintion they give of "near you" is "less than a mile from where you live". Imagine how parents who live in the suburbs of a big city might answer this question. They have the impression that the cities have gotten more dangerous and disorderly, and maybe that's even the reason they moved to the suburbs (perhaps even looking for the whitest neighborhood unconsciously) but they don't see this as affecting the area within a mile of where they live.
This survey might also be an indication of the role that politics and journalism have played in amplifying this effect. Perhaps originally it was a marginal effect, whether real or imagined (that crime was going up). But once that happened, politicians capitalized on that and began using the "crime is up" narrative in their platforms. It became part of the party line, and that provided a financial incentive for Rupert Murdoch and other right-leaning media owners to encourage more coverage of crime anecdotes. Covering that makes more money at that point, because it's telling people what they want to hear, validating their fears and re-enforcing their beliefs. Soon enough, people are watching (and resharing) a rising number of crime stories across the nation, but most of them aren't happening near where they live.
I commented in the Disorder thread on what I believe it is that is bothering people: That crime hasn't gone down sufficiently as a result of widespread anti-crime technologies, and in particular that police ignore crime even when there is ample evidence to convict (for example, you have video evidence of shoplifting, but the government declines to do anything about it). Historically, police not doing anything about shoplifting might be acceptable because they didn't have enough evidence, and there was a societal notion that criminals would eventually slip up and get caught; however, we live in an era in which people perceive criminals constantly slipping up, and those responsible for catching them simply decline to do so.
Yes, I accept that London is much safer than it used to be, but I am absolutely furious when the police refuse to investigate crimes where the perpetrator has been literally recorded doing it.
That's a very good observation. There are cameras on every pole and in every pocket, there are trackers in every other device, so there's a reasonable expectation that this will produce swift and inevitable punishment.
However, I think this ubiquitous surveillance is also preventing the efficient elimination of disorder. Some decades ago, a cop or a deputy could simply trash a hobo's tent and dump the owner on the road halfway to the next town, Will Teasle style. The same cop or deputy could also detain a small-time dope peddler or a petty vandal and give them a wallop without booking them. Nobody would really see how the sausage is made, but everybody would appreciate the more orderly streets.
But in 2026 the LEO is wearing a body cam, the perp is filming them, his friends are filming them, random bystanders are filming them. The existing sentencing process is not calibrated to handle this torrent of evidence. It implicitly assumes that the cops have some leeway in how they enforce the law and if any evidence of them abusing this power comes up during the trial, it's because it was egregious enough.
The perception of crime is one of those things that confounds public opinion research. In 23 of 27 Gallup surveys since 1993, at least 60% of U.S. adults said there was more crime nationally than the year before.
It's worth noting that public perceptions on most other social tends are *not* totally unmoored from reality. Generally speaking, peoples' perceptions of which trends are getting better or worse are correlated with real numbers. Given that background, it's somewhat of a mystery why peoples' perceptions of crime trends are so distorted.
Tangential: the most salient explanation I've heard for the preference of records is that operating a record player has significantly more friction. You don't just have the music available at the press of a button, so if you're not familiar with the concept of active listening, putting on a record is more likely to put you in the right space to do just that, as the instantaneous availability of digital music (and being served on platforms full of distractions) makes it easier to drop and do something else during.
> But I think there is also value in trying to figure out WHY people are wrong about this.
I think that's for neuroscientists to figure out, not some random web blog. Either way, it doesn't make sense for Scott to get mad about this. How is he still getting surprised about how people function?
"Overall crime rate" is not what most people care about. "Crime rate that might affect me" is the common concern. Murder can be down dramatically overall with more than all the decrease being among active participants in the drug trade - everyone who wasn't involved in the trade's "chance I might be murdered" is up even though murder rate is down.
"Crime avoided by costly preventive measures" goes into most people's perception of crime, but not into crime statistics. Even if the rate of theft is the same, a town where no one locks their doors and you can just walk into the stores feels a lot safer than one with bars on every window and armed guards in every store.
> "Overall crime rate" is not what most people care about. "Crime rate that might affect me" is the common concern.
True, although as I noted elsewhere people pretty consistently report that crime is not that big of a problem in THEIR neighborhood while thinking it's bad in the US overall.
> Even if the rate of theft is the same, a town where no one locks their doors and you can just walk into the stores feels a lot safer than one with bars on every window and armed guards in every store.
Yeah, agreed here. Which means you can make people think crime is up when maybe it's that you're not willing to pay for more retail employees so you have to do these things instead.
Maybe a pithy way to put it - instead I’d doing The expensive thing they used to do to mitigate shoplifting (hiring more employees) since crime is down they’re doing the cheap thing instead (locking shelves) but that makes customers feel like they’re doing the expensive thing, so they feel crime is up.
I doubt that's the motivating factor, though - companies seem almost (or maybe literally) pathologically committed to reducing labor spending regardless of whether it makes any sense.
""Overall crime rate" is not what most people care about. "Crime rate that might affect me" is the common concern. Murder can be down dramatically overall with more than all the decrease being among active participants in the drug trade - everyone who wasn't involved in the trade's "chance I might be murdered" is up even though murder rate is down."
I lived in Washington when in it was the murder capital of the US. The well off "civilians" were still very worried about crime even though the ratio was something absurd like 6 out of 400+ murders occurred west of 8th street NW.
I understand it's mostly the same sort of localization in Chicago, which also has a bad reputation for being out of control now.
I noticed when I used to be on next-door that people posted videos all the time that were for the most part just making everybody unhappy. I don’t think anybody was ever apprehended as a result of any of these grainy videos. But each morning here they’d come: I camera caught the car that was going by and emptying the mailboxes last night, here’s my camera showing a guy walking up into my garage and taking the beer out of my fridge, here’s a snapshot of the homeless guy who was banging on my door at 3 AM, here’s a picture of people who rifled through cars on our street last night.
Gadget lovers were over the moon about this.
But nobody ever got any interest from the police. At least not individually. Once in a while, you know a male theft gang might be arrested, so maybe it’s the case that they were collecting geographic data and then when it reached a tipping point, acting on it.
Stolen cars were often found, often in a sort of Bermuda triangle for gang crime but there wasn’t generally anyone to prosecute.
In one memorable case, where a homeless guy was yelling every night and just disturbing the peace in one location and frightening people by being at their doors and windows, the person who managed to get the police out, was for some reason actually cited for harassing the homeless person.
The post uses this specific example (responses 6o a post about whether crime is up) to make a more general point about online discourse.
A lot of online writing (including Scott's crime post) is debunking what other people think. The issuenis that we lump together similar but different claims together under the same general umbrella. This means that a debunker, like Scott, has to focus on specific claims to falsify. This leads to people with similar, but different claims, to complain that he only debunked some noncentral claim. But if Scott doesn't address the false claim, he is not adressing something that others feel like is the central claim.
Does this blog need to be "debunking" people? Isn't it simply enough to find a discussion, do your own research on the subject, and then publish it? People are free to disbelieve the findings, that's their problem.
I think the point is that one article cannot cover everything. Like, when you write an article about X, someone will probably say "but why didn't you mention Y?" and probably suggest that it's because you have an agenda. But of course, if you mentioned Y, the same objection could be made about Z.
And this becomes more tricky when X and Y *are* actually related... but so is Y and Z... and also Z and Q... because things are connected, in general. So wherever you draw a line, the very fact that you drew a line can be used to attack you.
Here, X, Y, Z, Q could refer to things like "national statistics of violent crime", "national statistics of white-collar crime", "national statistics of general nuisance", "statistics of San Francisco specifically", "things that people do to protect themselves from potential crime they expect" etc. Each of them a perfectly valid topic. But it is annoying when you choose one of them, do the research, write the article... and the comments are mostly about why didn't you talk about some other topic. (Especially if you already preparing a separate article on that other topic.)
Sort of. I think Scott has one of the best reputations for being earnest and sincere and unbiased in his search for truth, so there is certainly no need for him to "both sides" every issue so that every aspect is evenly and fairly covered. But lying by omission is a real thing. It's one of the primary reasons people have grown to distrust a lot of media. You can publish a lot of technically true articles that give a heavily distorted view of a topic.
When I read about nutrition - one of the few things I know about more than 99.9% of people - everything I read is wrong.
When I read about … I dunno something I know more than a vast majority of people about but not at the 99.9% level, and it’s all still seemingly very completely wrong all the time.
So then when everyone tells me crime is down - but all the Walmarts and cvs has everything on lock … and my own experience mentioned last week where my own big box retailer reports basically 0% of shoplifting … what am I to believe? Them, or my own lying eyes.
Crime is probably down in big ass cities since the 80’s or something.
Murder is down.
But crime? Crime? Like vagrants and assholes doing slightly illegal things and people stealing petty amounts of shit? It’s not down.
For some reason it’s called disorder now just like homeless are called unhoused.
> So then when everyone tells me crime is down - but all the Walmarts and cvs has everything on lock … and my own experience mentioned last week where my own big box retailer reports basically 0% of shoplifting … what am I to believe? Them, or my own lying eyes.
Said this elsewhere, but this could also be a function of retailers employing fewer people that would otherwise discourage shoplifting.
> But crime? Crime? Like vagrants and assholes doing slightly illegal things and people stealing petty amounts of shit? It’s not down.
How can you be confident about this? Your views on nutricion are presumably not from your own personal experience of watching people eat.
Staffing is horrifically down amongst all retailers compared to even ten years ago. But that seems like a ‘ she was asking for it ‘ argument.
I’m confident because I see it! We literally see it in our stores and neighborhoods.
Life wasn’t an unlivable shithole 30 years ago. And it’s not now. But it’s not drastically better.
We see crime not being taken care of, we see it not being reported … I understand it’s anecdotal, but like Scott touched on (somewhere) everyone is lying about everything all the time.
> But that seems like a ‘ she was asking for it ‘ argument.
Eh, I'm not saying anybody deserves to get robbed. I'm just saying this could explain all of the locked shelves, rather than some independent sweeping crime trend.
> I’m confident because I see it! We literally see it in our stores and neighborhoods.
Shoplifting can be down and shoplifting in Walmart and CVS could be up. Shoplifting inequality could easily be decreasing which would increase exposure of the average shop to shoplifting.
The illegal things you are gesturing towards haven’t shifted to being called other things such as disorder, you can just look up shoplifting statistics, or assault statistics, or burglary statistics, or lighter types of crime statistics, they all have specific legal definitions.
A lot of people just refused to put in effort to engage and it looks like Scott felt frustrated.
I'm one of those people who think "crime is up, cops are hiding it" but since Scott addressed this *in* *his* *title* I would have needed to put in a lot more effort than I cared to in order to treat his argument fairly. Instead of just saying "no, I don't care what you prove" I said nothing, and feel happy with that decision.
> People subjectively feel that crime is up, falsely, as you have illustrated (and as they basically always have when surveyed on it, as best I understand it). And they also feel that disorder is up.
Agree this is what it comes down to. People always feel like crime is up. These posts from Scott have amounted to a back-and-forth between people claiming crime, or whatever crime-adjacent thing, is *really*, by some definition, up, and then someone pulling some stats to test it and finding that it's not up, leading people to come up with a *new* definition for the thing they care about and why *that* is up.
Usually ends with something sufficiently not quantifiable that it can't be proved wrong.
I'll just say - whatever time period people think are the good ol' days when crime was low - look at what people said at that time. They probably thought crime was higher at that time than before, back in *their* good ol' days, when crime was *really* low!
"But the inverse evil trick is saying something 'directionally correct', ie slightly stronger than the truth can support. If your enemy committed assault, say he committed murder. "
I'll add one variation (?) on this, which is the logical fallacy where a term gets used to mean two separate things.
One example that I see a lot are taking "casualties" counts from the Ukraine war and treating them as "fatalities" where casualties also includes wounded.
Another is when discussing education overhead and turning "administration and other non-teaching staff", which includes IT staff, librarians, janitors, etc. into simply "administration".
I think a lot of how to deal with this depends on whether we think the folks engaging are searching for truth/correctness or just want to win.
There's atrocity sorta fallacy too like after Oct 7, lots of claims of what Hamas did was going around, and there was the baking, and beheading babies one, Noah smith actually blocked people for saying we don't have evidence of baked babies yet as them "denying atrocity", another dude blocked me when I said after two years it seems Hamas did not indeed target children and the kids killed seem to have been fron stray reckless killings. To preempt any "Hamas apologist" thing, i stated clearly at the start that Hamas is evil, oct 7 was bad and all, didn't stop anything. Person literally stated I was lying about atrocities while admitting i was right that there was no 40 babies killed, it was a discussion about how it's fine to sometimes ask for evidence of some of the worst crimes about Oct 7, and I mentioned how the 40 babies ended up being a lie and he was lie "liar we know Hamas intentionally killed children even if it's not 40", this was crazy on three dimensions, he shifted from babies to kids, admitted the 40 numbers is incorrect and his actual evidence didn't show intent too! Same with mass systemic rape, tjis is actually something i believed for like a year after Oct 7, only to find that evidence basically shows maybe very few like 2 or 3 rapes happen and we cant actually prove those either and nothing to indicate mass or systemic rape.
Oct 7 is a great example here because you can't question anything, they could say Hanas dropped poisonous manner that melted people skin off and questioning it is basically considered implicit support for the atrocity. Trying to say the baked and beheaded babies thing is without evidence, and it's why point this out, you are hamas supporting because why would you try to draw the distinction at all. It's like why would you care about the distinction being drawn at all too if it's indeed still bad.
The Noam Chomsky Khmer rouge denial too, is basically "i think this number is unlikely to be true cause they are extreme" during the height of the event, and after a while, he admitted the numbers then were correct but it's still "Genocide denial" for him.
Noah Smith blocked me years ago for referencing Matthew Yglesias correcting him on Tsarist economic growth. Every once in a while I want to "like" one of his tweets and remember I can't.
Yeah this is interesting. I'm sympathetic to the point you're trying to make and even I get this feeling while reading your post of: "why does this matter so much to you?" I sense this weird pull at the back of my brain that I think these other people calling you a Hamas supporter must also have. Arguing about the facts here feels in bad taste somehow, but I agree with what you're saying that this doesn't make any sense. It's worthwhile to be accurate about what happened on October 7.
If I'd have to guess why this happens, I think it's because it's very hard for our minds to see the difference between you disputing some of the facts, and you disputing all of the facts. Or perhaps in arguing against some of the facts, it sounds like you're making a moral argument pro hamas. It seems like one of these system 1/system 2 heuristic type of bias. I have to force myself into system 2 cause otherwise I get a visceral reaction against your text.
Yeah, most of this are not actually me but observations on the discourse. Like stuff I see, the Noah thing was not one i directly participated in, but I was following him and had his notifs on then so i had front row to the entire thing.
The main one that was mine was actually a meta discussion specifically about conduct on the Oct 7 discourse, and to make myself clear i stated to start with Hamas was evil, terrorist, oct 7 was a war crime and massacre and unjustified, etc. didn't stop much. That discourse basically was like i said i think Oct 7 can be so sensitive people that are shocked by some of it might not be so quick to believe all the claims and that some of this reluctance seem to have been good in hindsight and I listed ones like the baked babies and Hamas beheaded 40 babies, which he agreed tgat indeed Hamas did not do that to 40 babies but that I was evil for denying Hamas targeted kids even if it's not beheaded 40 kids, it's the directionally correct thing Scott mentioned. And he attached a picture to a source which I hadn't even seen before and he would have known if he didn't just reply to my one comment and block. But also the supposed evidence didn't even prove it but like it would be pointless to try to argue indeed I wasn't arguing that Hamas didn't kill kids on Oct 7, just that they didn't seem to have deliberately targeted them as claimed.
This was actually over a year after Oct 7 cause for the most part, I didnt enter the discourse about that event for a year cause we didn't have evidence, that is true of why I thought a lot of the claims are being made without evidence but it's also why I didn't comment because I didn't have evidence to the contrary and people didn't seem in the mood then.
Thing here is the stuff is a big event, it's actually important we know the details of such a pivotal event. That was important, it's true it's a huge crime against humanity but there's a reasons claims od baked babies took off, it adds different stuff to the discourse. Like did Hamas truly did it just to kill jews or because they thought they could destroy israel or did they indeed do mass rapes and go around baking and beheading babies. This is important, it's the reason we study Holocaust and stuff.
Unfortunately, relying simply on the facts, while the best long term strategy, often makes things worse in the short term, as various factions realize that their agenda may be weakened. Since humans compete on an in-group basis by instinct, there really isn't any workaround, except persistence.
You don't get to know WHICH dastardly trick is currently being used against you before you choose which heuristics you're going to deploy to detect dastardly tricks.
Arguments with supporting data are (and should be) somewhat more credible than arguments without supporting data. This heuristic will occasionally cause you to give more credence to an argument that turns out to be wrong, but it's still better overall than not having that heuristic.
Also, cherry-picking still usually involves some risk of getting caught, though it is typically a smaller risk than with outright lying.
I suppose ideally the crime-is-up people would clearly state how their claim can be measured and then we measure it. You can try to be clever in this setup and name a metric that's prohibitively expensive or borderline impossible to actually measure in the real world, but one can hope that it would be obvious that that's what you're doing.
How dare you use your platform to argue with people who are not me. For some reason, there are people on the Internet who are responding to people other than me rather than making content that only responds directly to my beliefs and interests. This is intellectually dishonest.
There is another difficulty, which is that to most observers, a good-faith attempt at precision looks identical to a malicious streetlight. The surface form is the same: “That’s not quite right, here are the numbers.” Distinguishing motive requires context most readers don’t have. This is what makes this protocol both effective and routine in politics.
More generally, opinion often comes first, with the supporting logic second. That "streetlight" is just as often shone by someone selectively illuminating the part of reality that happens to fit their intuition. Thus, the intent is not always malicious. We just live in a world of invisible gorillas.
And unfortunately the number of malicious dark arts users who actually are trying to manipulate people with these sorts of arguments outnumbers the number of good faith actors like Scott. Being suspicious of such arguments is good epistemic practice (though concluding that they're obviously bad without confirmation either way is bad practice).
It's good and useful to identify and call out rhetorical sleights of hand that are used to obfuscate the truth. Identifying them and giving them names is good for this.
But you don't want to get tangled up in the labels. They should not get in the way of the core mission, which is to find the facts most likely to be true using the best available evidence, and debunk 'facts' that have been contradicted by the evidence.
The oldest sophist trick (at least the oldest in my subjective experience; I think I noticed this being done to me when I was still had a single digit age), is to jump to an adjacent topic the moment you realize you can't support your position on the current topic. Then, if you get into trouble on the adjacent topic, you can jump back to the original one, and most people will forget the trouble you had with that one. To anyone who's not paying close attention, it will look like you've come up with new points and your counterparty will have to labor mightily to re-explain.
It appears the labels, useful in many contexts, can be used to assist this oldest trick. I think the solution is to just call it out explicitly when it happens. Force a concession before you allow a change in subject, or a change in rhetorical framing.
There's some tweet I forget exactly, but it goes something like: "to an average person, if their package is stolen, or if they stand around all day looking at their front porch / install a ring camera so they can grab the package the instant it arrives, those are all the same, even though only one would appear on crime statistics."
Basically, to a normal person, if they have to take extra steps to keep themselves safe from crime, that's the equivalent to crime being "up" even if it doesn't get statistically reported.
Where I grew up, we didn’t lock our car doors. Later, I moved to a “bad neighborhood” and still didn’t lock my car door. Somebody stole stuff from my car there. I didn’t report it. But after that, I always locked my car.
My impression is that crime in that neighborhood is worse than where I grew up.
Scott could not conclude that, based on his methods. There’s no police report; and me locking my car door shows up in no data. He has to conclude that I’ve been swept up in a moral panic, because the circumstances I’m actually responding to are invisible to him.
For about a year, there’s been a tent set up in the woods in a park near me. Camping isn’t allowed there, it’s for people to walk their dogs and jog, and children to play soccer and baseball in the park’s fields. It closes at sunset; nobody is supposed to live there.
*I* know about this particular instance of disorder, but I haven’t called the police about it. So there’s no police report. And the victimization surveys don’t ask about homeless people you’ve seen. So he won’t show up on, say, the NCVS.
That tent is completely invisible to you and Scott’s hard-nosed empiricism.
Is it “awfully suspicious” that neither I nor any of my neighbors have reported this guy? We’re just minding our own business, like normal people. You think this tent was hallucinated by my moral-panic-fevered brain?
I guess the only way for my fellow citizens of a scientific frame of mind to believe me about the disorder I see around me is to lodge a formal complaint with my local police. Then it’ll be reflected in the statistics. I don’t like that, but I guess that’s where we are.
If it's just one type of disorder, then yes, that's not suspicious. But the problem is if every measurable type of crime and disorder seems to be going down, and all the immeasurable ones are going up. What about them being immeasurable makes it so they're the ones that happen more?
Do you have a good argument that the fractions of break-ins that are reported has declined over time? Or are you just pointing out that statistics can never prove anything with absolute certainty because they can always miss something?
The article in question addressed that by looking at homicides and car thefts, because they are almost always reported.
- Crime (as in somebody is actually victimized and experiences harm) is probably about the same
- Disorder (graffiti, drug selling, homelessness, loud noises, hooliganism) is up but hard to measure and distinguish from cultural differences and bias
- Lawlessness (people who don't care about the law and will break it if it benefits them and they can get away with it) is up a lot. (A very small number of lawless people have a huge effect)
- Defensive measures taken by normal people to protect themselves from crime and disorder are up
- The US taxpayer spends a lot more money to address disorder than in the "good old days". Because of this, taxpayers would feel betrayed if disorder stayed at the same level -- because they're paying a lot for it to decrease
- A small increase in crime visibility due to social media.
- Conservative political operatives online think they can score points by making crime seem worse than it is, because liberals "own" disorder and lawlessness. This part of the "moral panic" narrative is totally right.
So the increase in complaining isn't necessarily due to actual crime victimization. But that doesn't mean it's not legitimate. It's partly illegitimate, partly justified.
Crime is probably about the same as when? 2000? 1960?
Thanks for clarifying your position.
Keep in mind I haven't read all the comments on this article and I'm only asking about the point made above about crimes not being reported. Scott attempted to address that concern in the article and made a convincing argument that it's not reports that are declining, but crime itself. Do you have a good reason to think a smaller fraction of crimes are reported now? Ideally, an argument stronger than you know what conclusion you want to prove and you're looking for any possible gap in Scott's argument. The problem with hunting for gaps is you're only showing it's possible, but not showing it's likely. Given that Scott showed murder and car theft are almost always reported and are down, and that there is no obvious reason to think there would be a smaller fraction of crimes reported today than 25 years ago (before cell phones and video recordings were ubiquitous), it seems unlikely that the fraction of crimes reported are down.
If the fraction isn't down, your argument about not reporting your car break-in doesn't affect Scott's arguments. If the same fraction of thefts were reported 25 years ago, then thefts really are down.
It seems to me that the major point of the parent post is that people would interpret "successful crime" as somewhat equal to "prevented crime" - events that can't show up even in perfect statistics because they didn't happen; but they didn't happen only because people took undesirable preventative actions; in both cases the "threat of crime" was real.
If a store locks up its goods and they don't get stolen, that's technically a reduction in shoplifting, but it's in no way an indication that people feel that the threat of shoplifting has decreased, quite likely the opposite - the fact that they didn't lock them up earlier but do now is some evidence that the threat might have increased.
If threat of stealing stuff from cars increases in some city, people will start habitually never ever leaving any bags in cars because their windows would get smashed otherwise (I've seen these habits change over time in both directions). Now once that habit gets ingrained, the number of thefts from cars would decrease, and the number of smashed windows would decrease - but that doesn't mean that the threat has decreased, it's just that there are fewer opportunities now. What people actually care about is the chance of coming back to a smashed window *if* they left their stuff inside, but what the crime stats show is this valuable metric multiplied by how often people leave stealable things in cars. What people want is to feel safe enough that they could intentionally just leave things in their car as the risk is insignificant, but that would increase total crime stats by providing more opportunities.
If someone decides that they don't feel safe anymore in the park and stop their evening jogging, this can only reduce the crime statistics (because they might have gotten assaulted in that park before, but now it's not going to happen), but it doesn't mean that the threat of crime went away - actually, the person *did* suffer a real harm from crime, namely, that their freedom to jog was meaningfully restricted by the threat of crime.
So crime statistics count "crimes which happened despite countermeasures", but people actually care about the general *threat* of potential crime and if we want to compare different places and times with meaningfully different level of countermeasures, we need to adjust for that; if we want an apples-to-apples comparison, we need to estimate a "what if" of what the risk of crime would be for people who take the exact same (strong or weak) level of precaution.
I think you're right that their point was the threat of crime and I misunderstood that.
There's a problem with the threat of crime theory as well that needs an explanation. People taking precautions against crime is a negative feedback loop. This should act as a multiplier. If criminals collectively are trying to commit a lot more crime, but people take protective measures, the actual crime rate only goes up a little bit. The potential crime is some multiple of actual crime.
But crime has been decreasing for decades, and by the same logic, that should mean the threat of crime has decreased even more than actual crime. E.g. as the city becomes safer, people go on their evening jogs again, and stores hire fewer security guards, so the actual crime rate decreases less as people reduce precautions.
A second question is what does it mean for the thieves if property theft is down due to precautions? If theft is down by half (I'm making up example numbers), but attempts at theft have doubled, thieves are getting half the reward for twice as much time and effort. You would expect that to discourage theft attempts. On the margins, thieves should give up. People generally spend less time on things as they become less worthwhile. So we would want some explanation for what's driving more crime attempts even as it becomes a worse option.
You didn’t give an example of crime going up, you gave an example of crime being more common in one place than in another place, and finding that out the hard way. Nobody said crime is equally low everywhere.
Yup. My example is meant to illustrate how disorder can be meaningful to people and yet invisible to Scott’s methods. I can’t say from personal experience that disorder has gotten worse in any one place, since I’ve moved around the country a few times.
The people where I live now definitely have a sense that disorder has gotten worse since the 90s. But I wasn’t here then, so I can’t compare.
Imagine 2000 people who don't lock their car doors live in your old neighborhood. 1000 move to your new one.
Of the people who moved, 900 of them never have anything stolen. 100 of them have some stuff stolen from their car, 90 of whom don't report it, and 10 of whom do.
Of the people who didn't move, 990 of them never have anything stolen. 10 of them have some stuff stolen from their car, 9 of whom don't report it, and 1 of whom do.
From this data, Scott concludes that there are 10x more thefts in the new neighborhood.
The real world will be messier than this. But most of the time, if the rates of people having stuff stolen and not reporting it rise, the rates of people having stuff stolen and reporting it would also rise. If we suspect this isn't the case for some reason, we can find other sources of data to test alternate hypotheses. Scott used data from the National Crime Victimization Survey, which asks random people if they were victims of crime, to address the very problem you're concerned by.
The point of my example wasn't the theft, but the door locking. Look at the first comment in the thread. People who feel like they have to watch their porch all day to prevent packages from being stolen are having a bad time, but the NCVS doesn't notice, because the package didn't get stolen.
If minor crimes and the defensive behavior that protects against minor crimes are in an equilibrium with each other, then you would only expect a rise in lawlessness to result in a rise in reportable crime when people's ability to defend against it starts to fail, and the equilibrium breaks. (Or in cases like my own, where I thought I was too cool to have to defend myself.) But you would expect people to complain more when lawlessness rises and they're having to take a lot of defensive measures.
It gets a tricky, because people who are in a moral panic _also_ take defensive measures, and they also complain. That's why corporations locking up their goods gets mentioned; presumably they're only sabotaging their business that way because they're losing more money in shoplifting than the defensive measures cost in lost sales and payroll. Cold calculation, not moral panic.
> If minor crimes and the defensive behavior that protects against minor crimes are in an equilibrium with each other, then you would only expect a rise in lawlessness to result in a rise in reportable crime when people's ability to defend against it starts to fail, and the equilibrium breaks.
Not necessarily. The thread-heading comment implies equilibrium has already broken, because crime is “only” down due to everything being locked up or surveilled. That could happen if the cost of crime has gone way up or the cost of surveillance has gone way down, regardless of the underlying lawlessness.
The theft would. The subsequent door-locking wouldn't. If you look up at the first comment in this thread, it's about people's defensive behavior changes in response to crime not being reflected in the statistics. That's what I was trying to explain.
The crime discourse is increasingly sounding like the "God of the gaps" argument to me. When someone presents data showing that crime is not up, certain people insist that the crime must be where we do not yet have data for.
+1. Really does not feel like the pro-crime people have a grasp of basic Bayesian reasoning or truth-seeking, it should be obvious that crime is down way before reading Scott's post, never mind after.
You could consider it part of the cost of crime. If we increase the amount of time and money we spend trying to prevent crime (cameras, guards, locks, etc.), but see the actual crime rate be mostly unaffected, the cost that the perceived crime level has imposed on us has increased. I don't think society does a great job of considering those costs when we implement programs concerned with safety, so it results in a lot of extra resources going towards things like crime prevention, which may not actually help at all, thus increasing the true costs of crime, even if actual crime levels aren't changing.
It may take less than you'd think, if crime is actually relatively rare, and the costs of prevention are spread among everyone. According to https://www.safehome.org/safest-cities/ there was about $27B in property crime in 2023, which is a ton, but only breaks down to around $80/person. I can believe that the average American feels more than $80 in losses from crime prevention than they do from the actual crimes. This is even more true when you consider that there are only about 6.4 million property crimes per year, or one for every 50 people.
I'd definitely believe that people think there is more crime or disorder, or whatever you want to call it, just because our prevention costs increased, regardless of the actual impact it has had on crime.
Yeah, one of the big complaints I hear is about how a whole bunch of merchandise at the store is locked up when it didn't used to be. At least in my area, that is inarguably true.
Now, we can argue until we're blue in the face about why that's happening, whether it's necessary, whether these companies are behaving rationally, etc. etc. And that's all interesting. But at the end of the day I still have to spend 5 minutes waiting for an employee to come get a circuit breaker out of a locked case at Home Depot, which I didn't 5 years ago. And that's equally annoying regardless of why it's happening.
That is indeed very annoying but it's a corporate decision. You're not going to solve *that* problem by hiring more police or taking a more "tough on crime" approach to politics.
Not sure about that. About 5 or 6 years ago there was a noticeable increase in shoplifting at local stores, and today, there's a noticeable increase in the amount of goods that are locked up and require employee help to access. I realize technically this is "post hoc ergo propter hoc" or something in Latin, but to me it seems very plausible that the former caused the latter.
I don't think the reduction will be immediate. Even if shoplifting immediately dropped to zero it would likely still be at least a few years before all the locks were removed. But over the long term, yes, I do think a reduction in crime rates will lead to a reduction in anti-crime measures.
If we're only arguing about the object level issue of whether crime is currently lower than it was 5 years ago, then to be clear I do think you're right about that so maybe we're wasting our time here. My point here is that my life, and a lot of people's lives, are noticeably worse now in certain regards than they were in the 2010's, and that part of this worsening traces back to crime even if the crime spike itself is mostly over.
This is exactly wrong. The problem was solved. Many places used to be tough on crime, and the prior state of unlocked circuit breakers was perhaps even a manifestation of that policy that everyone took for granted. If we lock up enough criminals for long enough, we can reverse it.
For circuit breakers in particular, part of the issue is that there are newer styles of breakers with added safety features that are a lot more expensive, so there's more incentive to steal them and hurt the store more when they are stolen. Old-school basic circuit breakers look like they cost around $7 at my local Home Depot, while ACFI breakers cost $60+. A bit over a decade ago, US electrical codes started requiring ACFI breakers for circuits feeding bedrooms.
That said, a bunch of other stuff that doesn't have this problem also seems to be locked up these days.
This is technically true, but only if 'hiring more police and taking a more "tough on crime" approach to politics' does not actually result in an environment reducing the incentive to crime. Which might be the case! Frequently "do X" politics does not result in X being done. More police hired might go to places with low crime and high influence, and thereby do nothing for the overall crime rate.
Surely we can agree, though, that if shoplifting is no longer a risk and all else remains constant(!), those locks and plexiglass panels will at least eventually not be included in new stores and the percentage of products locked up in stores will decline?
No, I don’t think that’s necessarily the case. The TSA was put in place in response to 9/11. It’s been 25 years since 9/11 and terrist attacks involving planes have been very rare in the US since then, but we still have the TSA despite it being very inconvenient. If shoplifting goes down after the introduction of locks on merchandise, then corporate is going to assume that the locks are doing the trick and keep them in place regardless of whether they or the police are what has stopped it.
I agree with Melvin that the incentive structures are lined up differently for private companies vs the federal government. If the locks were actually doing the trick, what I would expect to see is cycles of fewer locks, shoplifting rising, and more locks. There's always pressure to remove a cost, celebrate the additional profit, and get a promotion or polish that resume and look for a job that's a step up. Things that are effective but costly have to prove themselves effective on an ongoing basis in such a corporate culture. In the federal government, DOGE notwithstanding, the existence of a workforce is its own reason for existing and being funded.
I think it’s helpful to consider exactly how corporate is going to make its decision. The problem is the visibility of the cost vs the benefit. As far as I can tell, there are basically two costs to locking up your mechandise: the cost of materials and installation, and the cost resulting from the inconvenience of having to unlock the merchandise for every customer who wants to make a purchase. Of the two, only the first one is going to show up on a balance sheet and it’s a sunk cost. Removing the locked cases is an additional cost so from a balance sheet perspective removing the locked cases might look like they cost more than keeping them even if that’s not the case.
Keep in mind that corporate managers aren’t stupid, they understand that there is a cost to keeping locked cases in place, but they can only really make decisions based on information available to them. Inventory loss due to theft and other causes is easy to measure, but the cost of keeping merchandise behind glass because that’s mostly going to be lost sales and that’s going to be noise against all the other factors that impact sales. When you have a measurable cost vs an unmeasurable cost the measurable cost is usually going to win. I used to work in business intelligence and I used to see this sort of thing all the time.
If it's just "venting annoyance at having to wait 5 minutes" then it doesn't really matter why it's happening. But the majority of these discussions are either explicitly or implicitly arguing for *policy change*. And the policies that would address the annoyance are dramatically different if the cause is "shoplifting is up and stores must protect themselves" vs "stores are employing fewer people, and so are less able to watch for shoplifters and have installed countermeasures even though shoplifting isn't any more prevalent" vs "nothing materially has changed, but stores are being pressured by higher ups to look like they're Tough On Shoplifting" vs "there has been a surge in funding for anti-shoplifting measures despite no surge happening so the money is being spent because it's there". It's irresponsible at best to advocate for a specific fix while not caring about the difference between those (or other) causes.
I don't know why more people don't talk about the locked up merchandise as a consequence of reduced staffing? It used to be the case that my local CVS had between 4-6 staff members on the floor at all times, stocking, working the registers etc... They could keep an eye on the merchandise. Now there are between 2-3, and instead we have self check out and locks. I imagine this is cheaper than hiring staff.
Yes, and people adjust their behavior to create a thermostatic effect on crime. Part of why car thefts and carjacking are down, for example, is widespread adoption of habits like not leaving things in your car, using a wheel lock if you have a Kia/Hyundai (visible everywhere in major US cities), and not idling in your car. But from the POV of the car owner, it feels like things are perpetually getting worse, since "back in the day" you used to be able to just leave your stuff in your car, idle outside your friend's house and scroll on your phone while waiting for him to get home, not bother with a big awkward metal bar on your steering wheel, etc. The loss of those things is a real quality of life effect that won't show up in crime statistics but will show up in "do you feel safe" and "do you think crime is up"?
Matt Yglesias has a post where he says that voters' definition of "corruption" is basically "politicians doing things I don't like," even if that's just voting for a bill I wouldn't have supported. I think crime is similar: "crime" means "stuff other people do, or might do, that threatens my quality of life," whether or not it actually results in a Bad Thing happening.
Thank your for writing that post as thoroughly as you did. And of course people who want to be upset will continue to find ways to be upset. Please don't stress out about it. The title of your post was very clear, and your post dealt with the title topic.
If you want to write another post about disorder, please do so! If you don't want to that's fine also :-)
My 2 cents, this is why your “much more than you needed to know” posts are the gold standard. If you edited both of those posts into one post, it makes it harder to argue against you. I recognize this is more work for you and the algorithm favors multiple shorter posts, but it’s probably more persuasive to anyone that actually cares enough to read in depth on the issue.
I agree with all your observations here, but as Scott said I think 'part 2 on Disorder coming next week' probably optimizes across the different trade-offs here.
Or at least, it's worth trying that format a few times to see how it works.
That and explicitly saying "Hey, I'm rebutting against this specific argument from this source, but I acknowledge there are other dimensions to the broader topic which people will feel I've missed".
Seems like a combination whataboutism and goalpost moving, but where you can choose to move only one inch in some direction and dodge the argument.
But in reality lots of people have lots of different ideas so maybe nobody does either of those. You land in one argument spot and there will always be people 10 inches to the right left or some other dimension that can reasonably say "you didn't hit me!"
In reality I liked your last few posts and changed my mind on crime rates, but yes, one single mega post would be better.
This is why I think your "more than you wanted to know about X", such as guns, posts are so good. They cover a large enough space to really prevent that issue.
Trust in data was a strange and fragile thing, incredibly historically anomalous, and enough entities responsible for trust and data screwed their Mandate of Heaven so hard that Uriel now has severe PTSD and a bad limp. Now we are in the time of ~~monsters~~ epistemic agnosticism bordering on nihilism.
I'd like to say enjoy it, the water is fine and the cookies are good, but I can't really be sure of that, either. Tend your own garden and try not to think what happens when we cease to understand the world.
>no, crime rates really are down
A big flashing header that you're only talking about the US might dissuade at least a few comments about Europe's issues, where the data collection problem runs screaming past Poe's Law and various forms of disinteresting biases. Or for a *slightly* less CW example than crime rates, Ed West's quite reasonable article about the many ways the UK population estimates range by millions: https://spectator.com/article/we-have-no-idea-how-many-people-are-living-in-britain/ Illegal immigration? Bad data collection? Fraud? All that and more!
Related to the crime question, the occasional mask-off moment of a leftist actually saying they think of crime like weather probably doesn't do anyone any favors when trying to convince anyone to Scott's right that crime is down: https://x.com/hecubian_devil/status/2025731262132920820 One idiot like this fuels a million confirmation biases that "the left" (writ too broadly) really doesn't care and so why trust any numbers from "the left" (again, writ too broadly) at all?
For me it seems you are in the business of finding truth and reporting it regardless of its political direction. It's a commendable position and rare. I sometimes disagree with your posts but I don't believe they are ever bad faith. What's my point?
Do you have listen to us?
I don't want you to get disheartened about what you are doing and stop or even slow down. I guess all writers have to listen to their audience at some point to hone their craft but are you at the point where you can just trust your own instincts and ignore your audience forever more? I think I'd vote for that - even if your audience was just me! Or when writing about correcting misconceptions do you have to listen to the feedback to get more insight into the misconceptions?
> I don’t know how to get around this. On the one hand, it’s a problem if people are saying false things, and nobody can correct them without getting mobbed by a bunch of people accusing them of committing malicious streetlight fallacy, muddying the debate, using Dark Data Journalism to steamroll over lived experience.
> On the other hand, it’s a problem if malicious streetlight fallacy can never be challenged, because perpetrators can always defend themselves by appealing to some hypothetical group of people who think Mexican immigration is worse than Central American immigration and are lying to convince people that it’s Mexican immigrants specifically.
This is precisely what cruxing was invented to address, yes? Your interlocutor has to be able to point to something tangible - and preferably predictable - that's the honest source of disagreement, so that addressing it honestly *resolves* the disagreement. If they know what the point of contention is but refuse to offer it, that's the essence of bad-faith argument. If they don't know what the point of contention is, they have work to do in making their beliefs pay rent. And if they're yet another in the endless string of anons with bespoke niche disagreements, then sometimes you have to pick your battles.
And ofc, lionizing ideological consensus or illegibility for their own sake rather give the game away at the beginning.
Right, but that's a method for a dialogue between two people.
When you are writing a persuasive essay to a large audience, every member of that audience may have a different crux.
And since the article seems to be written to persuade them, if it doesn't address one person's *personal* crux, then it *feels* like it's intentionally misrepresenting them.
Addressing a broad audience is a difficult proposition.
People experience things relative to expectations, not absolutely. People may be angrier about disorder than the absolute amount of disorder merits, if they expect less disorder, for some reason.
One reason I can think of is that, relative to the “good old days” of the early post-war era, the US taxpayer has paid several trillions of dollars, at least, to fix disorder.
If it’s the 1950s and there are no expensive programs to address disorder, you’re not as mad about it as if it’s the 2020s and they’ve been spending tons of our money on fixing disorder for decades and the best they can say is that it’s no worse than before.
If you hire a plumber and pay them a lot and after a lot of work, the pipes are still clogged? It only makes you angrier for the plumber to say, “Well ackshully, the pipes aren’t any worse than when I started, so what are you complaining about?”
I wanted to say this on the crime posts, but never got round to it.
Those people who were arguing against you on the crime posts should have been welcomed! But welcomed as only one datapoint among many.
You can't demand that everyone believe the truth, even when it's true. Imagine if crime has fallen, as you suggested. It still must be true that for some fraction of people, crime has risen. Their viewpoint is necessary and valid! Your post was arguing that they are in the minority - and that was amply borne out in the comments. Lots of people offered contrasting viewpoints to the "crime up" commenters.
But the crime up commenters could definitely have been right about their own experiences. I saw nothing in the comments to disprove them.
The crime up commenters were wrong to think that their personal experience disproved the larger statistical trends that you were describing in the posts. But they were still right for them, and I think everyone should have been more chill towards them.
I wonder about crime perceptions skewing because of distribution of crime, to be honest. Like, I lived in ATL for ages and it feels noticeably less safe in the good neighborhoods than it did 10 years ago, but if crime went down substantially overall but had a minor increase in randomness, then the bad neighborhoods would dominate the stats in having less violent crime but the median commenter-who-never-went-to-those-neighborhoods-anyways might be more worried about random crime. Might be hard to fully map out to test though. But feeling less in control of one's ability to avoid crime is much more about where crime happens than about total amount of crime happening.
That'd probably work as a proxy in a lot of places, though might vary regionally. I might try to dig up the appropriate ones if I can.
I'm just trying to put sense to the subjective difference in how things feel on the ground. But it also is tied into a just sort of weird antisocial vibes aspect I think, where folks aren't as convivial with other folks on the street as they were pre-covid, and the mental health crisis homeless are significantly more visible in cities like ATL (though they're aggressively trying to shunt them out for world cup prep, so there's also a humans-being-shifted-more-than-normal aspect at play too). But I know that there's a lot of places I'd have walked alone a decade ago that I'd def feel more nervous about, and I'm not inclined to be overly precious about such. People on the street feel less predictable in interactions than before.
I wonder how much this relates to the phenomena of "Summer of the Shark", where what gets reported can make people anxious about something that hasn't even changed (or is improving). In general, I think it's a mismatch of scale, where people expect the frequency of hearing about bad thing X is proportional to how likely they are to experience it themselves. The US is huge with a population of ~350 million though, which means that you only need a base rate probability of about 1 in a million chance of occurring per person per year to have a daily example of said thing to harp on about. If the probability is higher than that, then even if it's decreasing, the news can continue to harp on about it. If the news itself is going from a low rate of reporting to a high one, it can even reverse the apparent direction of effect by making it seem like it's increasing, because you are hearing about it more often, even as the actual rate is decreasing. Social media only makes all of this worse because you can get into some bubble where the only thing people talk about is 'horrible thing X just happened again'. This constant information input is providing you with the integral of bad thing X happening over time and space, instead of the local rate, but it's getting treated as local rate.
Yes, I expect this is mostly driven by media ecosystems, and in particular a backlash against progressive politics (inasmuch as 'crime is high, we need conservatives to fight it' is a long-term stable meme in our culture).
I'm brave enough to admit I'm a work-from-home computer jockey who rarely goes outside, my priors on how much disorder there is in the country is almost entirely driven by a murky cross between my media exposure and my ideological commitments.
Given the audience for this blog, I have a strong suspicion that this is a source of a *lot* of commenter's priors on the amount of disorder, but I haven't seen many others admit it.
Still, how many people used to swim in those shark-infested waters compared to before? If people are still afraid enough that those waters are still emptier than before, than there's still worries that needs to be addressed. And before the waters are full again with visitors and without shark attacks, no statistic about shark attacks would matter.
On a meta level this is an issue I have with discourse generally, and people who specifically make "caring about the truth" a core part of their identity specifically.
One of my favorite pieces of rationalist writing is Zvi's series on simulacra. He basically posits four ways of communicating, each further removed from caring about object-level truth:
1) "There's a tiger across the river" - there is a tiger, the jungle cat, on the other side of the river
2) "There's a tiger across the river" - there might be a tiger across the river, that's irrelevant to me. I don't want to cross the river.
3) "There's a tiger across the river" - people in my group are saying there's a tiger across the river so I'll be popular if I also say there's a tiger across the river
4) "There's a tiger across the river" - I think I can curry favor with groups that help my goals if I advocate for crossing the river.
While this framework was formative in how I think about communication, it's extremely flawed. Because nobody says "there's a tiger across the river" to say an interesting thing about a tiger and a river. Nobody. That's not real. Otherwise why didn't they say, "Ducks don't have thumbs" or "My resting heart rate is 82 BPM?"
Statements are never made absent the context of some goal. And so people who think their goal is "truth seeking" without doing a lot of qualification are actually just unaware of their goals.
Which isn't to say that the truth is irrelevant. It's more relevant than ever. And I'd rather people try than not, so good on the rationalist community for doing so.
But good luck coming to a consensus. And not because of malice. Because we are creatures in the dark shining tiny flashlight beams on the smallest possible part of an elephant, thinking the rest of the elephant is probably irrelevant.
Behind every communication there is an intent. When you write a post about crime statistics, it might feel as though your intent is "debunking fallacies", but there's a reason you're debunking fallacies, too, so there's an intent to that. You're trying to influence somebody, somehow. Maybe it's out of an autistic truthseeking for its own sake, sure, but I doubt it. The tone of the discourse suggests something more like: assertions that crime is high are being used to justify stances which you find threatening; therefore disproving those assertions ought to weaken others' interest in those stances; therefore you try to publicly weaken them. There are some ancillary reasons as well, of course: probably a bit of autistic truthseeking; plus 'blogging is your job' so you gotta post something; plus I imagine there are a lot of downstream points and cases you want to make that rely on the facts being established vis-a-vis crime rates so it's a building block in a longer-term project. But ultimately, it is a *goal* of yours to changes others' opinions on this matter.
Meanwhile the people replying to you have a goal also. They feel threatened by the state of crime they feel they live in, and they wish to debunk claims that this state is not real, because if it is not real then their feeling of anxiety and unsafety is justifiably ignorable, and nobody ever (EVER) wants society to be able to justify ignoring their anxiety. Some of these peoples' fear is caused by, well, their fear being actively fomented by e.g. fox news / twitter panics / doomscrolling.
But some of it is real also, and real *despite* the arguments he made. For example it is essentially a fact, not a misconception, that "bikes left out of in San Francisco will almost certainly be stolen within a day" (note: my anecdote here is a couple years out of date since I do not life in SF; it is just an illustration anyway), whereas that is not the case in the recent past or in many other cities. If you say property crime is down, yet it feels like my example fact is true, then they simply see you as mistaken in a fundamental sense. And it is easy, trivial really, to extrapolate a reason why you are mistaken: in this case, because bike theft got so prevalent that everybody stopped leaving their bikes out, and also because the bike theft is so unhandled by the police that people stopped even reporting the, so via multiple mechanisms the lower rate in the statistics is DUE TO the greater prevalence of crime in actuality. And of course there are a million other similar arguments for all the other crimes. Nobody cares about the *actual* rate of crime when they're calculating what to be mad about; they care about the one that they experience in their life.
So anyway we have to look at the goals of communication. Scott's goal is to establish that crime rates are lower (and implicitly: tell people whose experience of unsafety is greater to get over it). Their goal is to have their feeling of unsafety acknowledged (and implicitly: signal that he can choke on his statistics). It's probably true that he is more 'right' in a scientifically-certifiable sense, but nobody cares, because communication is about purposes and his purpose includes negating their fears.
> I don’t know how to get around this.
If you actually care to get a different reaction, stop negating anyone's anxieties. Only synthesizing gets a positive response. If you don't care then, fine, but don't get aggrieved by the result of not caring, because you signed up for it.
> " Scott's goal is to establish that crime rates are lower (and implicitly: tell people whose experience of unsafety is greater to get over it). Their goal is to have their feeling of unsafety acknowledged (and implicitly: signal that he can choke on his statistics). It's probably true that he is more 'right' in a scientifically-certifiable sense, but nobody cares, because communication is about purposes and his purpose includes negating their fears."
This is *BY FAR* the smartest and most useful thing anyone has said on this series of posts, including Scott himself.
Because as a living-above-a-notorious-homeless-encampment-and-working-downtown-and-thus-seeing-crime-every-literal-day-Seattleite, crime is very fucking definitely not "down" ENOUGH. Not enough for me to use the bus system to commute to work (a stop on my route was closed for years due to crime! That route had OSHA-violating levels of airborne fentanyl!), not enough for me or anyone with dogs or kids to use the park neighboring my home, and certainly not enough to feel "safe."
So why the everloving fuck is Scott - who says himself he avoids going into San Francisco - spending so much time and energy attempting to convince everyone to believe "crime is down?"
What can possibly be the benefit of convincing people it's "down," when there's still much too much of it?
> What can possibly be the benefit of convincing people it's "down,"
Even if the destination is very far off, it is still important to be clear about whether we are walking towards it or away; if we are walking towards it, we want to walk faster, not turn around and reverse all our progress to date; cries of “this isn’t working! we are nowhere near where we want to be, we need to radically change direction” are counterproductive, and we need to push back on that and call for walking faster instead.
I will double down on my point: that you have to parse purposes, not facts, to understand communication. If Scott wants a different reaction he needs to synthesize his purpose for the original post with their purpose in responding. Which is to say, 'listen'.
This is tangential to your point, but it reminds me how insane it drives me when I read about polls saying that 72% of Americans feel that the country is heading in the wrong direction.
But we are not going in the right direction. The reductions in crime are seemingly unrelated to policy, as they've happened worldwide. If people really want crime to be all but eliminated, then this centuries-long movement towards policy and culture that provides mercy for the downtrodden needs to be reversed first.
I never said they're unaffected. You were implying that social progress was responsible for the decrease in crime rates, when the crime rates are decreasing worldwide regardless of how regressive cultural attitudes are locally. Crime rates improving is simply an inevitability with the changing world, but that doesn't mean changes in policy were responsible for it. There are countries like Japan who have much less sympathy towards liabilities and minorities, and they have far less crime than we do despite having less wealth to work with. Radically changing direction from the current liberal policies towards such people could offer massive improvements.
I think you're hedging here by half-admitting that maybe Scott is right about the overall crime rate but then railing against him for even bringing it up when crime is still too high, and (I think wrongly) impugning his motives for bringing it up.
Suggesting that Scott's secret motive is to "tell people whose experience of unsafety is greater to get over it" is totally crazy and off the mark in my view. Correcting misconceptions is itself a common human goal. If everyone thought that beluga whales only dove down 500 meters deep, and you found evidence they can dive over 1000 meters, then of course that is interesting information to tell people. There's no hidden leftwing/rightwing motive. People being wrong about something is interesting to us all on its own.
I don't think that Scott 'feels' that as his motive. He feels like there is a factual misunderstanding and also he feels people debating him on it and he wants to win. But *functionally* that result feels like ignoring/negating those people's experiences. The fact that he does experience his personal motive as equivalent to its functional result is the reason he's getting so much pushback--because other people do. The point is that progress will be made when understands how this looks at the level of purposes and does something about it.
I mentioned this in my other reply: the thing to analyze in the whale example is not "why the person thinks they spoke" (to correct factual misunderstandings) but "why the person spoke", that is, why the person wants to correct factual misunderstandings. The two reasons may have basically nothing to do with each other, but people's responses are based on the truth of the action, not what the speaker believes about the action.
Perhaps the whale-fact-speaker thinks they're correcting people but it is part of a general pattern of trying to sound smart and superior in unwelcome ways. The audience's reply (say, rolling their eyes, or feeling admiration) will be to the reality of the action, not the speaker's belief. Of course the audience might incorporate what the speaker believes they are doing into their response (e.g. "I understand you're trying to correct facts but it feels like you're just trying to sound smart.") That depends on what their motive in responding is.
Also let me add. The 'purpose' aspect of your whale example is: what is the speakers' purpose in saying something about the whale at all? Why are they trying to share interesting facts or accurate knowledge in the first place? For attention? Out of boring? Out of a mild compulsion for correctness? Because it is normal to do so? Etc. There must be some reason because they said it instead of not saying it. If you only analyze it at the level of facts then you have entirely elided the important part of the interaction.
> Why are they trying to share interesting facts or accurate knowledge in the first place? For attention? Out of boring? Out of a mild compulsion for correctness?
Yes, those are potential reasons and that was my point about that. A desire to correct misconceptions is a common motive on its own. You could (for example) fairly describe someone with a mild compulsion for correctness as seeking truth.
The point that there's some reason Scott thinks this is worth saying is fine. The problem is you aren't just doing that; you're making assumptions that he has particular left-wing, soft-on-crime motives, and in particular that "people whose experience of unsafety is greater" should "get over it".
You are very much misunderstanding/pattern matching to the wrong thing.
I don't know where you got anything about me ascribing left wing motives, I think no such thing. I am saying that functionally when you say 'no crime is going down' you are invalidating the concerns of people who feel otherwise and that's why they react so strongly. This is just a simple and predictable mechanism; there is no political angle to it. I'm not claiming that he understands or intends that he's trying to invalidate them ; I'm saying that he *is*, because that's what refuting people's anger does. He believes (perhaps) that he is refuting misconceptions, but by choosing to invalidate them (instead of entertain them them) he is clearly stating how he feels about the emotional stances he is replying to, which is that they don't matter to him. If he had the opposite purpose he could easily write an equally factual article that supports them instead. He seems to basically not understand how any of this works, or he would not be confused about the reaction.
Let's refresh on Scott's stated motive for writing this series:
>"My goal isn’t to deny anyone’s lived experience, nor to discount the importance of solving these problems (I support the death penalty for boom box carriers). It’s to push back against a sort of Revolt Of The Public-esque sense that everything is worse than it’s ever been before and society is collapsing and maybe we should take the authoritarian bargain to stop it. On an emotional level, I feel this too - I can’t go downtown without feeling it (one of many reasons I rarely go to SF). But I don’t like feeling omnipresent despair at the impending collapse of everything. Having specific thoughts like “house prices are up since the pandemic, so it’s no surprise that there are more homeless people, and more of the usual bad things downstream of homeless people”, rather than vague ones like “R.I.P. civilization, 4000 BC - 2026 AD” isn’t just more grounded in the evidence. It’s also more compatible with living a normal life. I’m not a pragmatist who thinks you should be allowed to lie or do a biased survey of the evidence in order to live a normal life and escape despair. But I’m also not some kind of weird anti-pragmatist who makes a virtue out of ignoring evidence in order to keep despairing. "
Scott actively avoids going into downtown San Francisco because of his feeling that downtown San Francisco is worse and collapsing, a feeling presumably based on literally *seeing* how San Francisco is worse than it was 10+ years ago and sure looks like it's collapsing.
He states that he doesn't like the feeling of despair at the impending collapse of everything, so he decides to literally avoid San Francisco, a place where despair is a completely appropriate response to the visible environment, and write about crime going down *nationwide*, so that he doesn't have to"feel the despair."
And shares it with his readers so that they, too, can - if you will - "get over" said despair.
My personal opinion is that this is a misfire and very much not a good use of his attention or time, and upon reflection, is particularly irritating if he's correct about the other subject on which he's trying to persuade folks, upcoming advancements in AI.
If AI is destined to advance as much as he believes, then there actually *is* an impending collapse of everything for the many (most?) people whose livelihoods will be replaced by a bot without any alternative. Right now the plan for their survival is being drawn up by the Underpants Gnomes, so the aggressively cheery articles about "crime is down, get over your despair like I did" aren't sitting well with some of us.
Why should Scott have to write about San Francisco instead of the US? The national crime rate is a perfectly valid topic and it's more relevant to all his readers outside of SF.
> And shares it with his readers so that they, too, can - if you will - "get over" said despair.
He shares it with his readers so they can have an accurate and informed view of the national trend.
If crime is worse where you live, by all means, you should be upset about it. But it's not fair to criticize people for talking about a trend because the trend doesn't hold somewhere. No trend will be true for everyone.
This is silly, and both you and Scott should know better.
National crime trend statistics are irrelevant to an individual's risk, which is why, e.g. car insurance companies use extremely granular information about your location when assessing a price to make you whole should your car be stolen, broken into, or crashed into by an at-fault person committing the crime of not carrying insurance themselves. They don't care about the national averages when it comes to assessing an individual's risk of being victimized in an auto-related crime, and neither Scott nor anyone living in places with crime and blight above the national average should be comforted by a dropping national average of crime if *their* *city* *is* *so* *awful* *they* *actively* *avoid* *being* *in* *it*.
He can write one on San Francisco, as that's more or less the same thing; he's apparently aware that the city is effectively unusable enough to openly state he avoids going there himself.
(San Francisco is more relevant for the rationality community, but I feel like I heard a disproportional number of horror stories from Seattle considering how few people I know who live there.)
Vancouver, BC, Seattle, Portland, San Francisco, and LA are all looking and feeling pretty dire, especially compared to what they were like 10+ years ago. (My personal observation is informing that statement. More vacant storefronts and buildings, more litter and graffiti, more open drug use and illegal camping, more human waste, etc. Like, sorry, I don't care what a study says, you can literally *see* it, check historical Google street views, it's *different.*)
Official crime statistics about these places should be viewed with a bit of skepticism, as there's a strong ideological motive by city leaders to downplay the volume and impact of crime, disorder, and blight. Seattle has a whole division of street team workers dedicated to cleaning up especially egregious messes made by trespassers/loiterers/indigents, as well as trying to move them along when they bother businesses too much. They're quietly literally cleaning up crimes, which prevents those crimes from being officially reported as the crimes they are. Lies, damned lies, and statistics, etc.
It also seems unclear that the studies are controlling for all the new measures being taken to thwart crime - if shoplifting is steady despite a recent trend of stores spending a lot of money locking up items and hiring more security and using increasingly advanced camera systems (Target is basically the CIA now), obviously crime would be *much* higher without those strong *new* deterrents in place. Criminals haven't gotten less crime-y, victims are putting in greater effort to minimize the harm done to them.
This argument frankly sounds like a schizophrenic arguing why the CIA is actually stalking them. Sure, the data might not agree, sure there is no tangible evidence of this, but I saw a broken twig on my way to work which was clearly a mistake by one of the agents stalking me. Also, why does that schoolbus drive by my house every morning? If the CIA isn't stalking me, why have I seen that Walmart employee look at me twice in the past month?
If your "lived experience" trumps data, why even bother collecting data in the first place? After all, it'll only be used to justify your position in retrospect.
I think "staying attached to reality" and "practicing good heuristics" is Scotts main motivation for the crime posts, not anything political. I guess what one person considers good epistemic practice is another persons autistic truthseeking.
I think the issue is that disorder invites the potential for serious crime in a way that makes the true crime rates irrelevant if you're living with the disorder.
I saw a tweet thread once that pointed out that a common anti-social behavior, listening to loud music on public transit, carries the implicit threat of death if you confront the listener. Everyone has seen the many different stories of unstable or psychotically antisocial people who stab the person who asks them to turn down the music; and so when you're around someone breaking social norms like that, you're not just "dealing with the hassle of having your focused disturbed", you're dealing with an implicit death threat; or the threat of getting punched, or at the very least being screamed at and menaced.
And the people who violate social norms know this, and it's part of why they do it: it's fun to dominate the crowd. Will you turn your music down if someone asks you, or will you fistfight them, or will you kill them? Only you know the truth, and they don't, and so you live in comfort, with everyone treating you like a potential murderer, as if you're a warlord and they're peasants.
Even if "real crime" is "down", "situations that invite the threat of real crime" does not seem to be "down".
Someone else pointed out that, say, sharing a train with a loudly raving, clearly deranged person, triggers feelings of physical insecurity in people (and rightly so), even if that person eventually does not stab or shoot anyone.
IDK why Scott does not address such comments. He treats perceptions of crime as misunderstanding of hard statistics instead of something that is reciprocal to perceptions of personal safety. Which may be satisfactory for very exact minds, but it is also societally and politically tone-deaf and unproductive.
How would you address these comments? Are there any stats for people raving on the subway? Maybe it's also less common. Maybe it's more common. Maybe it's the same
The argument is that these unmeasured crimes likely track the measured crimes that he does discuss. This is part of the “crime as proxy for disorder “ post.
If your complaint is that he addressed some forms of disorder, but not the specific one you wanted, that's the point of this post.
In some cities here in Europe, there are attempts to make "perception maps", where people can say "this and this corner or park feels extra unsafe" etc. Subjective, but better than nothing. Over a longer period, this could actually translate in something that can be analyzed. Then we would see whether these really track each other or not.
Because he’s pretty obviously making a data based argument which contradicts vibes and the comments keep trying to push it into places where there isn’t data to look at but only vibes that could be used to support whatever the speaker wants them to, even while the original argument showed us that the vibes are way off the reality where the data does exist.
Why don't people who play loud music on transit have to live with the same threat of death that an unstable or psychotically antisocial person may become annoyed with them for it and stab them, scream at them, or menace them over it? Are they all just real brave?
> Why don't people who play loud music on transit have to live with the same threat of death that an unstable or psychotically antisocial person may become annoyed with them for it and stab them, scream at them, or menace them over it?
Because the cluster of "people who listen to loud music on the subway" also clusters with being more likely to be unstable, antisocial, violent, etc in a way that isn't symmetrical, and similarly, the potential outcomes are assymetrical for both sides.
Anyone with a career is unlikely to listen to loud music on the subway AND unlikely to confront the loud music listener, not just because the person may be a psycho, but because even if they weren't, if the police were called, it's more likely to hurt the career-haver than the music listener.
The police can basically raze the career-havers life to the ground arbitrarily if they "assaulted" the poor music-listener by asking them to turn the music down and it turning into an altercation. The music-listener is immune to that threat, they have nothing to lose.
So career-havers can't do anything for two reasons - 1) the guy may be a psycho, and 2) the police aren't on your side and are more likely to ruin YOUR life than to improve the situation if they get involved.
The fact that this is true is a large part of the current problem. No good actor can act on behalf of the commons or career-havers, and any bad actor can act with impunity, and the police are literally worse than useless - they won't fix any problems, and are more likely to ruin any career-haver's life who tries to do anything about it.
Public transit users are such a tiny proportion of Americans (Census says 3-5%) that it makes me suspicious that this is always the go to example for public disorder.
If public disorder were extremely common surely people would be using more common examples that actually affect the other 95% of Americans.
I think there's a common perception that the character of crime has changed, independent of its quantity. And like, rationalists are going to hate this because it's probably impossible to quantify. I worked various crappy retail jobs from 2015 to 2022 or so. I couldn't tell you if the dollar value lost to theft went up, down, or stayed steady during that time. But the behavior of shoplifters changed, a lot. There was a distinct sense of... I guess you could say shamelessness that crept into the whole thing. "Person stuffs a few items in a purse and furtively sneaks out" and "Man sweeps whole shelf of merchandise into a cart and walks out while making unbroken eye contact with the clerk, knowing damn well he can't even confront him" are both shoplifting. But, call it irrational if you want, they *feel* very different to be around. Even if the dollar value is the same.
People have been sneaking onto public transportation for as long as it's been a thing, and I don't know if fare evasion rates are up, down, or sideways. But watching a group of teenagers jump the turnstiles *while a transit cop stands there watching them and does nothing* (true story) has me thinking thoughts of societal collapse and degeneration of public morals, in a way that wouldn't be true if they were doing it sneakily.
This shift in brazenness has probably been influenced by social media. When clips of people shoplifting started going viral, it spread awareness of what had always been the case: shoplifting is easy to do and difficult to stop. It's not necessarily the case that many more people were willing to shoplift, it just might be that many more people know it's easy to get away with it, because they've seen videos of people doing it.
I agree with the form your non-apology takes here, people mad at you for defining a topic you want to talk about and then talking about it are getting over their skis in assuming malicious intent or incompetence.
What I think you might *instead* consider from this experience is that, if people thought you were spotlighting when you weren't and you found it hard to navigate between these two failure modes in a public space, maybe have a little more charity and sympathy for the 'dark journalists' and others who look to you like they're employing these tactics.
We all know about the fundamental attribution error... you did something that looked bad to outside observers, but did it for good reasons, and it wasn't actually that bad. But when you observe someone else who looks like they're doing that same bad thing, they must just be a bad person doing the bad thing intentionally and maliciously.
No doubt there *are* bad people in the world trying to do evil, but we could all stand to take a bit more time considering possible motives and context *before* jumping to that conclusion. If your own innocent and well-intentioned actions can end up being perceived this way by outside observers, then maybe others who look like they're employing these 'dark techniques' are also just doing their best in a complicated world.
I am *always* this guy, given my job, but I really think it's important.
Crime is ALWAYS a very technical thing, and it's ALWAYS going to be subject to "well, actually...", that is literally the point of why the justice system is set up the way it is: we punish specific violations of statues, where very specific "elements" all must be proven beyond a reasonable doubt to the satisfaction of a jury, precisely because we don't want to punish based on "vibes."
One can literally see someone gunned down on a street corner in broad daylight and correctly "well, actually" that act. After all, it doesn't take too much imagination to see that the act of shooting someone MIGHT NOT BE A CRIME. Kyle Rittenhouse can tell you a bit more.
Labeling a particular act, or category of acts, as a "crime" has very specific consequences for a debater. They're now playing by the rules of law, where pedantry is the name of the game by design. Imagine someone turns to me and says "we should punish all criminals more harshly, and in particular, people who smoke cigarettes." The natural pushback is "wait, smoking cigarettes isn't a crime." What I'm saying there IS rather pedantic: I'm saying that within the jurisdiction I'm in, the general act of lighting up a cigarette does not technically meet all the elements of a specific statutory crime. That's precisely that it means to say an act "Is or is not a crime."
I realize I seem like a pedant when I do this, but the whole point of crime is that it's not categorical. Most general acts (tackling someone, screaming really loudly, sending an email, uttering the words "I am literally going to kill you, no joke", firing a gun) are criminal in some circumstances and not criminal in other circumstances, and we generally need to point to specific laws and findings by juries to tell the difference. I hope we don't need much creative thought to imagine circumstances when all the acts in my parenthetical are very serious crimes, and other circumstances in which they're perfectly legal.
Rittenhouse was acquitted because the court deemed that he was defending himself from attackers, and those attackers were thus indeed committing crimes.
that's not necessarily true. One can have justified self defense even if one's attackers are NOT committing a crime. As a toy example: let's imagine two people are both told that there is a deranged shooter in the neighborhood with a gun, going door to door killing people. To face this threat, they both grab guns, and step outside, see one another, and shoot. In one universe, A misses and B hits, B kills A. In another universe, B misses and A hits, A kills B.
In one universe, a jury for person "A" could easily acquit them, based on the defense "I thought the other guy was the deranged shooter I heard about." A jury for person B, one universe over, might likewise acquit on exactly the same basis. These sorts of self-defense scenarios occur, a little less cleanly, all the time, and critically as well, they're jury dependent. They necessarily hang on what a jury would conclude based on the facts presented to them.
I don't mean to say that in Rittenhouse's case, his attackers WEREN'T committing a crime...its very possible they were. But the entire encounter is illustrative of how hard it is to talk about "unprosecuted, uncharged, and unreported crime", how many crimes was that anyway? one assault for every person who struck him? moved toward him? Threatened him in any way?
also important point of further pedantry: the court didn't "deem" anything in the rittenhouse case. The jury gave a verdict of not guilty, and juries don't need to explain their verdicts. I agree with you that the jury PROBABLY thought he was properly defending himself, but we don't know that by any means, and it's a little hard to "thusly" conclude anything about his attackers as a matter of law or fact
There's certainly a lot of edge cases when it comes to crime. And it's even possible that when you add up all the edge cases there's more of them than non edge cases.
But surely this can't mean that we can pedant ourselves out of being able to meaningfully say things like "Tijuana has more crime than Tokyo".
I think even greatly uncontroversial cases dwarf all other possible crime on a scale that's difficult to articulate. I'm writing a blog post on it now. FTX defrauded a million people. Was that a million separate crimes? There are only about 100,000 crimes reported in Minnesota every year.
We end up in the land of pedantry due to neo-reactionary assumptions of the sort scott is addressing. We can talk about how much "crime there is in Tokyo" with how many crimes are CONVICTED or CHARGED or REPORTED since these are objective criteria and often discernable even without reference to statute (you can say how many "assaults" were reported without worrying too much about the statutory definition of assault in Tokyo)
But we end up in the land of pedantry when we assume bad faith on the part of these counters, and instead claim to be able to measure how much crime there "actually" is. I have to ask for the metric then. How does one count the frauds that are unreported? how does one know about every drug deal that transpires in the shadows? how does one partition the thefts? Most every murder is an uncharged assault. Most every robbery is an uncharged theft. Most fraud is a ton of little frauds, and even a lot of theft is a bunch of smaller thefts.
I wonder how much of crime perception comes from parents not talking about crime to their kids. Then when they become adults and start reading the news, they get the feeling: "Oh my God! Crime has been so much worse lately!"
I vividly remember in middleschool the first time I became aware that our country was 'at war' and what that actually materially meant, and I thought it was some type of extraordinary alarming state of affairs that I needed to be hugely concerned and scared about all the time.
Of course, the US has been 'at war' for over 90% of its existence. Social studies class eventually taught me that this was not a state of exception, but rather extremely common and normal. But I don't think anyone gets explicitly taught this about crime.
I worry about how Kids These Days learn about the news.
In my pre-internet childhood I started learning about the news because the TV was tuned to the news at 6pm every night, and then when I got slightly older I learned more about the news from reading the newspapers that were always strewn around the house (because I was bored, and there wasn't a magical box constantly feeding me the things I was mathematically most likely to engage with).
My kids don't get any ambient exposure to the news, so I have no idea how they're eventually going to have to learn about it all at once, probably from someone with a specific agenda.
One thing I cannot find reference to in the commets on both posts is the possibility of changing laws / redefining the entire concept of a crime? Famously, it is one lever that governments have accessible, and since they tend to be judged on the published numbers (up until some threshold where the divorce between opfficial numbers and reality is noticeable to the naked eye) there's a strong incentive to use it.
This should be something we can check, also: have there been redefinitions of felonies into misdemeanors, and misdemeanors that become legal, etc over time? Could this be visible by discontinuity analysis in the data or is the resolution too small and it's been death by a thousand cuts over time?
I asked opus 4.6 (deep research). intro quote:
"" California has reclassified, downgraded, or decriminalized dozens of criminal offenses since the 1960s, with Proposition 47 (2014) representing the single largest mass reclassification in state history — converting roughly 40,000 annual felony convictions to misdemeanors overnight. These reclassifications have created a minefield of statistical discontinuities that make longitudinal crime analysis treacherous. The primary distortion affects arrest data (felony vs. misdemeanor counts), not UCR Part I offense counts, because the FBI defines offenses independently of state severity classifications — a theft is still a theft regardless of whether California calls it a felony or misdemeanor. Yet indirect effects on policing behavior, reporting incentives, and offense categorization ripple through every dataset. Neither the FBI nor the California DOJ publishes adjusted time series; researchers must navigate at least seven major structural breaks spanning 1976 to 2024, with the NIBRS transition adding yet another layer of complexity."" "
matching crimes across jurisdictions is a famously vexing problem in the law. I generally think it's best to think of crime as a purely jurisdictional artifact, rather than some general proposition. There is no such thing (in america...with a few very weird exceptions...) as a "common law crime"
so, what is "murder" in one jurisdiction is not the same as what is "murder" in another. There are certain acts that might be "murder" in Minnesota that might not be "Murder" in Arkansas...because we have different murder laws. Crimes must be specifically articulated in a statute...and every jurisdiction writes its own statutes.
now, for differences in murder these are weird edge cases. If I say "imagine a murder", the thing you envision in your head is probably murder everywhere, but as we work down to the sort of ambient crime that is everywhere, this gets complicated. Some things are felonies in one state that aren't even crimes in another (Marijuana comes to mind). Other things are misdemeanors in one state, and felonies in another. Some jurisdictions have a thing called "Fraud" which is different than "theft", but in my jurisdiction (minnesota) they're literally the same crime. So in minnesota "theft of the shoplifting kind" is statutorily the same crime as "theft of the kind where you pad your hours in an attempt to defraud a subcontractor for some mega-corporation."
confusingly, not every jurisdiction even has the same definition of "felony" and "misdemeanor." In minnesota we even have a whole separate kind of crime, the "gross misdemeanor", sorta halfway between the two.
True about matching exactly - but luckily we could just estimate a dampening effect based on the downscaled acts numbers? For CA, it means at the federal level and local laws. And if someone was _very naughty_, maybe check for retroactive coding corrections in the databases - that'd send a spanner in the analysis...
This is a failure in analysis I think: This assumes mistake theory, when actually the situation:
Is mistake theory up to the point where someone would have to challenge what they want, then it was conflict theory the whole time.
See: The right wing free speech absolutists who were all about your right to the freedom to say whatever you want, in whatever way you want, on whatever platform you have access to, until EG. people started making jokes about CK being truly, completely committed to the right to bear arms, upon which suddenly it wasn't about that after all.
This is another POSIWID situation, where judging from what people say vs. judging from what they do produced radically different goals.
EG stands for exemplie grattia I think, traditional abbreviation for for example.
CK is charlie kirk, but it was pleasing to use CK after EG, it gave the run on sentence a pleasing symmetry for me.
Just an observation that nobody is fully mistake maxing or conflict pilled; people flow form one state to another based on the circumstances and their feelings on the matter.
Sometimes the challenge to what I want is never going to be affected by what I say (see the number of elections where any seat is decided by less than 1,000 votes), and I just want to pretend to be a good person for a while. What do I do then? I have the same problem Scott’s described, and I knew something was wrong but I didn’t know what. I’m hoping someone finds an answer
> and I just want to pretend to be a good person for a while
I'm pretty sure that's what charity and volunteering are meant for. Personally, I just stopped caring about being a "good person". It's pretty liberating, you should try it out some time.
Quit pretending and just do what you know is right all the time. It isn't hard once you work out your axioms, which is a fun thing to do in it's own right.
Scott, you're talking as if there's a clean and clear distinction between malicious streetlight effect and a legit attempt to address a confused issue.
But the distinction between these two always turns on prior beliefs about 1) the empirical facts, and 2) what one values in the situation. In other words, there is no "context free" way to distinguish these two.
You can call out errors of empirical claim, if you have sufficient arguments and data, but you can't dictate the larger import of your correction.
I think the reason you heard so much about disorder is that people like to complain, and if you put up evidence that something they like to complain about is not so bad they look for ways to legitimize complaining.
Seconded, with the addendum that it's not just complaining for complaining's sake, it's also indictment against an ideology/politics you dislike, and towards one you prefer.
I think the issue is that "decreasing over time correctly-chosen timescale" is not the same as "not so bad". Things can be decreasing and still bad!
It's perfectly legitimate to complain about things even when they're not as bad as they might have been at some point in the past, and that's true of left-coded complaints like "inequality" just as much as it's true of right-coded complaints.
The problem in these debates is that "actually this problem is not as bad as it was at some point in the past" is taken as meaning "stop complaining about this problem you rubes", which gets people understandably upset.
<The problem in these debates is that "actually this problem is not as bad as it was at some point in the past" is taken as meaning "stop complaining about this problem you rubes", which gets people understandably upset.
Yes, I agree that's the problem. But I would locate that problem entirely in the people who get upset and then discharge their upset in the form of posts spewing disgust and annoyance about problems that are not the one under discussion. Doing that is not a person's only option when a blog post that clearly is not saying the reader should just stop complaining sets off an irritating fantasy that it does. And it someone really is not able to do anything but that -- well, this isn't group therapy, you know?
I would say that your problem and the problem of guys like Cremieux and countless others are that you are mistaking statistics with knowledge. You feel entitled to speak about fields you did not study, research or got direct experience of. None whatsoever. Your contribution could charitably described giving your vibes about the validity of the statistic at hand. By chance your vibes invariably are pro-business and everything-is-fine-but-you-are-too-dumb-to-realize-it. You are just increasing the noise and drowning the signal. Stay in your lane and the world will be grateful
Taking the framing that communication usually takes one of the three forms of information-exchange, rapport-building, or doing-battle, talking about crime tends to fall under doing battle. When doing battle, the other party does not absorb and integrate facts. Rather, it attacks them. Thus, people will either question the source or the framing of the data, or, barring that, fall back into some sort of conspiracy about how data is being suppressed at the data collection level. I would love to see an example of someone saying "oh, I think I was wrong about this." You won't find many such instances, though, because this is one of those worldview-defining positions of the right-wing mindset. People that think that crime is up, because this is pumped into their heads 24/7 by Fox News, will never be convinced by facts contrary to those beliefs. The left seems less wedded to their view on this (that crime is going down) and seems mostly to just take the opposite view out of principle. However, I believe it's always important to put the facts out there and to try to reach the minds that might listen, even if it is mostly futile. So thanks for doing the work, Scott.
"My plan was to publish a post one day on crime, and then the next on disorder, but I got so many negative comments the first day for talking about crime without mentioning disorder that I guess in the future I’ll include in the post that disorder is a separate topic and I’ll talk about it later. I don’t know a better way to thread this needle."
Is there a reason why you did them as two separate articles instead of one "Much more than you wanted to know"? It seems like they would have been stronger together and avoided much of this issue.
Like, the first essay on crime was ~2k words and the second disorder essay was ~2.5k words, while your "Prison and Crime: Much more than you wanted to know" essay clocked in at ~13k words (1). It seems like you could have combined both of those essays into one bigger essay, plus some more, and those have been better received in the past and avoided this issue.
Well he said he wanted an easy article to link to whenever someone said "crimes are way up!". I'm fine either way, mostly because the gap between posts is short, but I guess he'll be back to the MMTYWTK format know.
I appreciate the attempt to apply reason to this issue. The problem is that the government, the media, and academics are not trustworthy. It would not be shocking at this point to discover that officials were too lazy to even put their thumb on the scale tactfully and decided to fabricate the data outright. People trust their eyes and ears, and to some extent what they see online, because that's the only real information available. If the numbers contradict their general sense of the situation, then they will assume there is something wrong with the numbers or the analysis. If Pravda publishes numbers saying food production is up, but the shelves are in fact empty, would you trust the news? Would you even bother to quibble about the meaning of "up" vs "adequate"?
I'm old enough to have now been an adult through two different noticeable declines in violent+property crime in the US -- mid 2000s through mid 2010s, and now the faster/deeper decline starting in 2023 -- as well as the 1980s/90s crime levels which briefly recurred during 2020/21.
During all of those periods I resided in either a core city (one very large, one much smaller) or a suburb of same. And then in addition for the past 15 years my work has taken me regularly into rural parts of the US having population densities drastically lower than anyplace I've ever personally resided in, which has been eye- and mind-opening on some topics including this one.
Fully granting that this is simply one person's anecdata, here's what comes to mind from pondering the above life experiences and this topic:
(a) OF COURSE the rate of general public disorder tracks with the rates of non-trivial crimes. That sure as hell has happened where I live now, as it did (in each direction) during various previous decades when I lived in various parts of the US. I asked a few people in my age bracket about this and to a person they first waited for the punchline to my joke setup before realizing that I was asking a serious question. [I may, possibly, be somewhat known for such joke setups.] Their collective actual answer can be summed up as, "Um, _yes_....so hey is everything okay with you?"
(b) The obvious reality that those two things correlate does not _prove_ that they are linked nor if so, how (which comes first). I've never yet seen a really persuasive case on the specifics nor the direction of such a linkage. That said, having at this point seen it come and seen it go multiple times I would if in policymaking office take it as just common sense that those two things do go together somehow.
(c) Also OF COURSE, none of this is going to boost Democrats in any national elections during at least the remainder of my lifetime. Two reasons for that.
(c1) One is that American progressives and fellow-travelers have well-earned our comprehensive loss of public standing and trust on this specific very-high-salience topic. Even if we were all now reckoning with our errors and omissions on this, which way too many of us are very much not, the hole that we spent a generation digging for ourselves is too deep be filled back in quickly.
(c2) The other is that many people, not limited to Americans and not limited to social conservatives, are simply hard-wired to be sure that crime and disorder is higher than it used to be and/or than it reasonably should be. That's a big part of "very-high-salience". This is not a new fact, is not because Internet nor because "if it bleeds it leads" in the MSM [I used to believe that one] nor because social media; etc. Why it is true I'm not at all sure. But it is true, and will remain so long after I'm no longer around to compose grumpy-old-guy comments online.
> OF COURSE the rate of general public disorder tracks with the rates of non-trivial crimes.
Just an idea: Maybe this is true locally, but not globally. It could simultaneously be true that:
1) In any given city, the general public disorder is positively correlated with non-trivial crime. The streets containing more homeless people and drug junkies are better to avoid, because you are more likely to get attacked there, etc.
2) Over time, the ratio between the general public disorder and non-trivial crime changes globally. There is less non-trivial crime these days, globally.
To put it simply, a street with homeless and junkies is *more* dangerous than a clean street; and it always was. But a 2026 street with homeless and junkies is *less* dangerous than a 1976 street with homeless and junkies.
So one side goes like "of course, homeless and junkies = danger, and I see many homeless and junkies around me, what do you mean that the crime is low?"
And the other side goes like "yeah, but you are less likely to be stabbed by them today, than you would be in the analogical situation 50 years ago, that is what we mean by the crime being low".
Today you may see stolen goods sold openly on the street, but 50 years ago someone would have robbed you instead. You see the bad part, you don't see the alternative.
Well sure, that seems simply obvious as a matter of simple fact.
Seems though pretty much like just another useless well-actually type point. Irrelevant, at best, to real-world public/political response. “People used to get drunk and kill each other in their cars way more than they do now” does not impress anybody who’s furious that in the year 2026 some yahoo with a DUI on his record was able to retain his drivers license and now has obliterated a member of that person’s family.
Of course. No matter how much things improve, people want them to improve *more*, and I think that's okay.
It's just that when people take "things are extremely bad today, the worst ever" as a fact, and start looking for a scapegoat -- because that is the typical human reaction -- when I wish more of them took the more abstract perspective. (Ahem, make something something great again, etc.)
I have thought more about this since I wrote the comment you quoted and read your next post (and please edit to use my name and link to the full comment, and also for any time you quote me in the future). My conclusion is that I don't actually think it is a matter of the actual crime rate or the sense of disorder, but rather the tolerance of crime and disorder. When I see the police walk by people selling stolen goods on the sidewalk and do nothing it fills me with a blinding rage that seeing 100 thieves arrested will not. When I see videos of people stealing from CVS with impunity, and the security stands by and does nothing because if they do they will be punished by the law while the criminal goes free, it stirs a hatred in me that no level of crime in the statistics ever could. This is theft of my tax money. This is treason by the government that is supposed to represent me. Not only do they do nothing about the criminals, but actively protect them.
No, I don't want to waste money on that idealistic crap. I used to be of the mind that law enforcement was overdone and should be done with a lighter touch, but boy was I wrong. These people only speak one language, so I just want to see the batons come out for the people bipping cars and stealing from stores.
As a toy example, imagine that there are only two types of crime, let's call them "big" and "small". Every year, there are 100 big crimes and 100 small crimes. The police only has capacity to investigate 100 crimes. Investigating a big crime or a small crime requires the same amount of time.
Would you prefer that the police investigate 100 randomly selected crimes, or 100 big crimes?
The response that considers only the first-order effects would be like: Of course, if you can't investigate all crimes, start with the big ones, they are more serious!
But if you consider the second-order effects, you can see that if you make "only investigate the big crimes, ignore the small crimes" the policy, it will have a huge psychological effect on both the small criminals and their victims. You have practically made the small crime legal! So you can expect it to grow, but even if it somehow doesn't, it will still be psychologically different for both the criminals and the victims.
So maybe a better policy would be something like "investigate 75 big crimes and 25 small crimes", to avoid the effect of de facto legalizing the small crimes.
I am sympathetic to resource constraints and suboptimal allocation issues. Getting that right is extremely hard. But that isn't what is happening in cities like SF. They are overflowing with revenue. And even in the most resource constrained city and worst crime wave imaginable I can't accept that the police will go after people defending their property while letting the criminals go.
Then why not name your articles "assault and murder are down"?
Just "crime" would include every illegal staying in country and fraud in paperwork. People also mentally measure financial "crime" and drug crimes including scams in their dollar amounts and not number of cases.
First I don't see how this maps onto the streetlight effect. Second and more importantly I wonder why most discussions have to be so adversarial. Why can't we talk about things more civilly, like we're sitting next to each other in a bar. I sometimes feel like the evil overlord is pushing everyone's buttons to make them angrier (or something). Don't listen to the evil overlord! :^)
> Why can't we talk about things more civilly, like we're sitting next to each other in a bar.
Because attacking people who make your life worse is an effective strategy for getting what you want? Sure seems to be working out for the right. People are rallying around hatred, and finally taking effective action against the people and systems that denied them their rightful position in society. Hard to argue that it's not working.
>>Because attacking people who make your life worse is an effective strategy for getting what you want?<<
Is it really though? Doesn't this just cause more anger and hate?
>>Sure seems to be working out for the right. People are rallying around hatred, and finally taking effective action against the people and systems that denied them their rightful position in society. <<
Is it the right that is rallying around hatred and such, or is it both sides? It doesn't seem to me that anyone is getting what they want. At least that's my take locally (I live in rural America). I woke up this morning feeling like grace was close at hand and that the right path is through love and God and the cosmic all.* And I hope to keep that throughout the day and pass on that love to those around me.
Be at peace my friend.
*I've been reading some fairly woo filled books, "A Course in Miracles" at the moment. Back to the good words of the book.
> Is it really though? Doesn't this just cause more anger and hate?
Not when you finish the job. That's why the left could never win. All that power, and they just never did anything meaningful with it. Their very ideology prevented them from imposing their will through force. The right isn't going to repeat their mistakes.
Many of the people you're dealing with here are simply liars, to themselves and to others. You cannot "thread this needle" because this needle is not willing to be threaded.
I think a steelman (of sorts) of the "it's still the reporting bias" position is that the police are getting better at lying. In the past, it was possible to obtain some independent crime stats, and so the police wouldn't lie too much; but now, essentially no stats can be trusted at all, and the only data one can obtain is via personal experience.
(I've talked to some global-warming deniers who advanced the same argument: that true climate data is essentially unknowable, due to political pressures.)
I appreciate your willingness to seriously engage critics. It's probably very core to your thinking style, and it's a big part of your distinctive appeal as a writer.
At the same time, I think you might be subjectively better off if you had a little more condescension towards people who refuse to engage in good faith. Or even those who try, but are incapable of meaningfully understanding.
The whole purpose of that post was to respond to people who say "But what about reporting bias?" in a way that took their objections seriously. If someone responds, with no further clarification, like:
"Yeah, but have you considered that maybe the numbers are wrong?"
That person can safely be thrown into the "not worth considering" bucket. It seems like these kinds of responses bother you, that you interpret them resulting from some failure on your part instead of theirs. IMO, it would be better for you to just feel a little jolt of derision and then forget about them.
It's a bit risky, if you do too much of that you end up being pretentious + wrong + uncorrectable, but in small enough doses condescension contributes wonderfully to mental well-being.
I’m not sure this is something you can do on your own. It takes both participants in a dialogue to keep the dialogue aimed towards truth.
In general, people need to try harder to engage with what their actual interlocutors actually say instead of things they pattern match to, things other people with similar political views say elsewhere, things they could have said that there is a ready answer for etc etc. This is really hard, especially when you are emotionally attached to the subject, and takes continuous deliberate effort. I, too, am too often guilty of responding to others with rhetoric instead of logic. We as the commentariat need to do our part.
I like how you go about categorizing fallacies. The "directional" fallacy of exaggeration and fallback, though, is the motte-and-bailey fallacy, not a new one. The Streetlight Fallacy is good and new (for me) though.
There is something like the motte-and-bailey fallacy which is, I think, legit. We might call it the Trump Error Tactic. Someone says there are 1,000 rapes by immigrants. Opponents then say, "No, there are ony 900 rapes by immigrants, you liar!" The tactic has worked, because most people think there are 50 rapes by immigrants because the opponents control the media and are careful never to mention the numbers. Trump uses this all the time.
I think this is the right observation. Most people don't actually care that crime is up or down, rather they care if it is high or low. 4 murders per 100,000 people might seem "high" to a person in 2026, even though it's a near-historical low, because their frame of reference isn't "back in 19XX", it's "compared to X neighborhood/city/country". People travel more, are more mobile, and of course the internet exists now so people regularly talk to others all over the world and learn how exceptionally high the murder rate in the US is.
In this context, the "crime is down" motte for the "crime is low" bailey in the soft-on-crime vs hard-on-crime debate is almost cliched at this point -- pretty sure I've seen it as a punchline in some dystopian fiction before, though I fail to remember any specific examples. I think Scott is right that people responded based on what they *thought* Scott was doing (apologizing for crime and disorder as part of a familiar refrain they've become conditioned to react poorly to) as opposed to what he was actually doing (making a very narrow argument against another very narrow argument).
Out of all the pixels used to discuss crime over the past few weeks here, one thing I haven't seen discussed is whether people are just programmed to worry about it a lot, in the same way a gazelle is probably programmed to worry about big cats, and that maybe this leads people to perceive crime in our locales as generally bad, because otherwise, why would we be so worried about it? From an evolutionary standpoint, you can imagine why this is advantageous: you're better off as a gazelle over-reacting to that rustle in the grass because the costs of under-reacting if it turns out to be a leopard are really really high. Likewise, people are better off over-categorizing individuals around them potential muggers and rapists and so on because the costs of under-categorizing are potentially catastrophic. This leads us to perceive urban areas in general as more dangerous than they actually are and to engage in a bout of what Tyler Cowen calls "mood affiliation," where we search for arguments to support an emotional response we've had, rather than try to form beliefs based on FACTS or LOGIC, and that emotional response is fear or worry.
The other thing I would point out is the following: the major increase in crime and disorder that occurred in the US in the '60's and '70's was at least in the minds of some people enabled by the social upheaval that coincided with it (not saying it was all the hippies' fault or anything, just that the Boomer-fueled social revolution of the time had some negative consequences, one of which was higher rates of crime and disorder).
For better or worse, that social regime is still in place. The cultural norms and assumptions that emerged in the '60's and '70's are still around and largely unchallenged. We never had a Thermidorean Reaction or a Bonapartist Coup that turned back the clock. The Jacobin Club is still open and Robespierre's head is still on his shoulders (I mean, Reagan won a couple elections, but he was no Bonaparte, obviously). I think maybe there's a kind of unconscious association in people's mind that says "because the cultural attitudes of the '60's never went away, the associated crime it enabled never *really* did, either. It's just been temporarily suppressed by technology like Ring cams or modern trauma surgeries" or what have you.
Although I liked both posts, I am still a bit perplexed at splitting them up. Yes, there's value in clear standalone reference posts, and maybe there was some Speed Premium concern behind the scenes...but it seems quite predictable, even without d20/d20 hindsight, that there was gonna be a ton of Well, Ackshually conflation of low-level crime/disorder with capital-C Crime. Integrating both posts into a single cohesive piece, maybe a light rebrand as Falling Crime Rates: Somewhat More Than You Wanted To Know - there, preemptive disarmament of common retort. And much less likely to be screenshotted out of context by the Cade Metzes of the world.
Can't do anything about the "still asking questions already answered by t-shirt" phenomenon though, sadly...
You article was fine. Doing them separately was fine. This is a situation where you should ignore your critics.
You were not doing a "malicious streetlighting". There were specific arguments that people were making, and you were narrow and specific in what you argued against. The example you give of the bad thing you're accused of is helpful in highlighting the differences. If the claim is "immigration on the Southern US border is up", and the retort is "There are fewer Mexican immigrants", anyone can see that these are different claims. You responded directly to arguments about crime being up by citing crime statistics and arguing about crime. That OTHER people wanted to have a DIFFERENT conversation about crime-adjacent things and chose to excoriate you for not having the conversation they wanted you to have is them behaving badly.
You identify an 'on one hand/on the other hand' conflict in the end of your post. It's misplaced. There is no conflict to resolve. You were right, and the people who wanted to talk about something related but different were off-topic.
Agreed. It’s worth adding that people like myself, whose response to the post was, “an utterly unsurprising conclusion, but it’s nice to see it rigorously documented,” have little incentive to comment on the post. Since crime is down on average, people whose “lived experience” is rising crime are going to be a minority of the population, and quite possibly a minority of Scott’s readers.
Framed as a game of Motte & Bailey Chess where a certain strong claim in the bailey is false and a certain weak claim in the motte is true:
Directional Correctness: 1. White is in the bailey and makes the strong claim. Black meets White in the bailey, and refutes the strong claim. 2. White retreats to the motte and makes the weak claim. Black points out White's retreat and refusal to concede the strong claim. 3. White accuses Black of Malicious Streetlighting. Black to move.
Malicious Streetlighting: 1. White is in the motte and makes the weak claim. Black pushes White into the bailey, and refutes the strong claim. 2. White points out Black's push and refusal to address the weak claim. Black accuses White of Directional Correctness. 3. White to move.
Scott, you have now written multiple posts in which you define "crime" and "disorder" as two separate and distinct things and use this distinction to try to tell us that everyone who thinks crime is up is an idiot who doesn't understand the facts, but of course we can probably all agree that disorder is up, that's completely different and irrelevant.
You define "disorder" as follows:
> Disorder takes many forms, but its symptoms include litter, graffiti, shoplifting, tent cities, weird homeless people wandering about muttering to themselves, and people walking around with giant boom boxes shamelessly playing music at 200 decibels on a main street where people are trying to engage in normal activities.
Just a super quick rundown:
1. "litter": Littering is a crime.
2. "graffiti": Vandalism is a crime.
3. "shoplifting": Shoplifting is a crime.
4. "tent cities": Camping on public property is a crime. So is trespassing on private property.
5. "weird homeless people wandering about muttering to themselves": While most people find this annoying, the muttering isn't really what makes most people nervous in this scenario. What makes them nervous is when some Jordan Neely wanders within earshot mumbling about how he's going to kill some white bitch today, because it makes them worry that this clearly violently psychotic individual is going to murder or assault them or their loved ones. Murder is a crime. Assault is a crime.
6. "people walking around with giant boom boxes shamelessly playing music at 200 decibels on a main street where people are trying to engage in normal activities": Violating noise ordinances is a crime.
I hope that helps you to understand why people didn't find your posts very helpful!
For example I believe disorder is probably down, since measurable crime is down.
I feel what Scott is saying is "disorder is hard to measure, maybe it's down or maybe not but can we agree the measurable things are down please" and instead people are saying NONONONO CRIME IS UP AND ALSO DISORDER IS UP.
In 2010 the deodorant wasn't behind bars and there weren't hypodermic needles all over my kids' playground. Disorder is up where I live, relative to the pre-Woke days.
No doubt the disorder isn't evenly distributed. But are you saying that shit isn't more fucked up in your neck of the woods than it was 15 years ago, back before the Ferguson Effect and Defund the Police? Are you sure you're remembering correctly?
Sure, crime goes down when nobody leaves the house anymore. Want to look at those stats? They're grim. I work from home and don't go to movie theaters or restaurants anymore -- in this, I reflect broader trends that certainly contribute to lower rates of crime victimization. This is not because everything is hunky-dory; it's because movies aren't worth watching, restaurants have tripled their prices while service has gotten worse, and downtown has turned into a George Romero flick.
Multiple people including myself made this observation in the comments of the previous post and I was really expecting to see it in a classic "Highlights from the Comments". I really thought I was misunderstanding something because I was sleepy, or maybe he'd phrased something poorly. The fact that we got this unprecedented refusal to engage is... concerning.
Anyone can have a bad day that leads them to write something that was dumb in hindsight, especially if they've moved to San Francisco and formed a progressive-heavy friend group. The doubling-down days later suggests an ideologically blinkered bubble that has corrupted his epistemics. The way they selectively change definitions to pretend that something isn't a problem is an example of the sort of thing he used to be against.
Yes. "Home invasions are actually DOWN since we turned into a Mad Max hellscape in which every family whose house is not protected by multiple machine gun nests is immediately raped to death by marauding biker gangs. Everyone without the machine guns got raped to death way back in 2023, but now in Q1 2026 home invasions are a real rarity! So the claim that crime is up gets five Pinocchios!"
Including disclaimers in the post about what common concerns you're not addressing can be a good idea, but shouldn't always be necessary. In my opinion, the problem is that a significant percent of blog readers expect your articles to address their specific perspective, and you simply can't write a single readable article that engages with a thousand different arguments at once.
"So you use FACTS and LOGIC to prove that something similar-sounding-but-slightly-different is definitely false. Then you act like you’ve debunked the complaint.
My “favorite” example, spotted during the 2016 election, was a response to some #BuildTheWall types saying that illegal immigration through the southern border was near record highs. Some data journalist got good statistics and proved that the number of Mexicans illegally entering the country was actually quite low."
The empirical world cannot provide proof. (Logic can in closed logical systems, e.g. proving mathematical theorems, but these are distinctly different epistemic domains than the empirical domain.) Not gonna stop harping on this because it's a basic stats 101 insight.
I feel like you're defining proof to only mean "mathematical proof" whereas the "proving something means providing overwhelming evidence that is convincing to most reasonable observers" seems just as valid to me? English isn't my first language though, maybe I'm wrong here ? 🤔
> this argument neutralizes a real and influential group of people trying to make the contrary argument that murder/crime rates are up, and to push policy based on that position
Isn't a big part of the problem here that we're trying to argue about numbers as a proxy for policy rather than arguing about policy directly? There's an assumption here, not quite stated and not quite fully believed, but smuggled in anyway, that if crime is "going down" over some timescale then we shouldn't do anything about it, and that we should only be dealing with problems that are currently going up.
This is of course silly; even if there's less crime than the recent spikes it's still far too much, far above the theoretical minimum, and even far above the practical minimum that could be reached with reasonable trade-offs.
Reasonable questions to ask when trying to decrease the crime rate include what tradeoffs would be reasonable (higher taxes? civil-liberties-violating random searches? draconian sentences for minor crimes?) but the question of whether crime rates are higher or lower than one, three, five, ten years ago are not that important; it can help us to gauge whether certain policies are on a right-ish or wrong-ish track, but there's a lot of other factors that go into the crime rate so it's unlikely to be that useful.
If we could dispense with the smuggled assumption that if crime is going down then we don't need to do anything about it, then we could discuss the numbers dispassionately and without trickery.
I think this just goes to show that there's no short cut for having the argument. You can write a post about X, people can argue that you're making fallacy Y, you can argue you're not (and maybe they're making fallacy Z), you go back and forth, clarify your positions and hopefully somebody somewhere changes their mind. There will almost never be an "argument" that is won after just one post, or one post followed by a knockdown accusation of fallacious reasoning.
The recent posts on crime and disorder are interesting and informative and they provoked some useful discussion (as well as, perhaps, some not-so-useful discussion). I don't see any need to reassess the way you approach issues like this - those posts were a success!
This is why the internet broke our politics. Not because of evil algorithms but because this sort of effect causes everyone to perceive the points made on the other side to be unconvincing strawmen.
We are always intensely aware of which subgroups we belong to but find it hard to make those distinctions about groups we aren't part of -- think parents mixing up goth, emo and punk or the like -- and we hate being mixed up like that.
This leads to an inevitable problem whereby some people broadly on the left [could be right, picked one for illustrative purposes] make some incorrect argument and to people on the right it seems to directionally capture what is problematic and mistaken about the left and they rip it to shreds. The problem is that to people on the left that wasn't even directionally their view, it was those crazy extreme greens while they are a sensible pro-growth lefty of whatever.
And honestly, I don't have a good solution. In the before times there was a clearer sense of what is within the mainstream of a given political ideology but now that we all have megaphones there is an unending drip of statements that are infuriatingly wrong coming from what you can reasonably label as 'that side.'
This creates a horrible trap where both sides can spend all day just telling you "can you believe people on the right/left said X" and each side will infuriate the other as it feels like they are attacking a strawman but they are real arguments out there that people are believing so....
I just stopped associating with anything. You can't really get offended by groups or get offended by being associating with a group if you simply do not identify with anything in the first place. If people want to murder each other for stupid reasons, while that is my problem, it's one to be solved by figuring out how to navigate the situation without dying, not by getting attached to certain groups.
I meant I don't have a good solution for society. Yes not identifying too much or just being too contrarian to belong is great for Mr individually (well great in this regard...makes it hard not to annoy people)
The problem solves itself. The strongest group will eventually neutralize the weaker ones, and then you will have a period of relative cohesion and peace. Until entropy kicks in, of course.
I think people just absorb from an early age a verbal template for discussing collective problems in a Grown-Up way, and it includes saying that some statistics have changed a lot recently, which is bad. It's basically a news hook, like journalists typically put into op-eds and such. In reality, the feeling that a problem *exists* typically precedes any real knowledge of how the problem has changed over time. And there is a certain wisdom in this: if you're often accosted on the street by crazy violent guys, it isn't really that relevant to figure out whether the rate of accosting has gone up or down by 20% in the last year or two.
"So just say what you actually think is bad and stop pretending to care about quantitative window-dressing!" I agree. And I think it's good for people to rebut incorrect things that people actually say, even if those people don't really care about their own claims in the first place. But I think the realization that "problem is going up" is basically a trope, a literary device, helps to explain why narrow counterevidence rarely seems to chasten anyone.
> I guess in the future I’ll include in the post that disorder is a separate topic and I’ll talk about it later. I don’t know a better way to thread this needle
Please don’t beat yourself up about this - you are already more intellectually honest than 99.99% of writers, by posting follow-ups that surface your commenters’ objections and seriously engage with them. That’s why we are here!
Your existing process is exemplary, and here it is working as it should when you get more pushback than you expected.
Crime is always up over some arbitrary time span and always down over some other carefully selected time span.
For example, mainstream liberal publications are addicted to start counting crime in 1991. Why such a random-seeming year? It was the peak of The Wire Crack Wars. Once you start noticing this 1991 addiction, you can't unsee it.
Similarly, when people say that crime is up, what they often mean is that it went up recently, but not necessarily very recently. For example, the temporary triumph of the Black Lives Matter movement in the last week of May 2020 due to George Floyd's demise caused black deaths by homicide to increase from 903 in June 2019 to 1303 in June 2020, according to CDC WONDER mortality data. That was a 44% increase in the number of blacks dying by homicide.
Even more strikingly, black deaths by motor vehicle accident rose from 506 in June 2019 to 772 in June 2020, a 53% increase, as cops responded to the BLM hysteria by retreating to the donut shop, which encouraged people to drive more recklessly and pack their illegal hand guns as their fear of being stopped and searched declined.
Black deaths by homicide and car crash remained very high into 2022, but then the BLM fervor started abated and cops were encouraged to get back out there and start pulling people over and patting them down. And guess what? The end of the Floyd Effect saw black deaths by homicide and traffic fatality fall sharply, especially after Trump's re-election.
So, what lessons should be drawn from the Ferguson and Floyd Effects?
A common moral is that because crime is now down in 2026, we should completely forget about how elites drove black murders and traffic fatalities through the roof, twice, in the name of Black Lives Matter.
Personally, I find crime trends interesting, just as I find baseball stats interesting. But, then, I WOULD, wouldn't I?
Murders in the U.S. dropped rapidly in the mid-1990s after crack, flattened out for a decade, declined again after the economy crashed (murders, like traffic fatalities, tend to be Deaths of Exuberance as opposed to Case and Deaton's Deaths of Despair), and bottomed out in 2014.
So then elites got all worked up over Ferguson because the crime problem was, apparently, SOLVED. So murders went up in 2015-2016 due to the Ferguson Effect, then dropped in 2017 and especially 2018.
The murder rate seemed to start going up in mid-2019 as "New Jim Crow" thinking came back into elite fashion, then exploded during the last week of May 2020 due to the Floyd Effect. (Both the Ferguson Effect and the Floyd Effect were worst among blacks and were mirrored in traffic fatalities, due to to less proactive policing.)
Murders stayed high from May 25, 2020 (have I ever mentioned that date before?) through mid-2022, then started to drift downward as BLM fell out of fashion in the months before the 2022 midterms as Democrats started to worry that Tri-State Voters were blaming them for the increase in crime.
Murders continued to fall and dropped very hard after November 5, 2024.
It's almost as if murders and traffic fatalities go up or down depending upon whether American elites are telling the police (who, as their name suggests, tend to be highly political) to back off or go hard against crime.
Scott, I think you're doing great and appreciate what you're doing :)
I have nothing to add to the debate, but I have a nitpicky question: is there a difference between the "evil streetlight" thing in the post and good old strawman fallacy? "Defeating a weak version of an opponent's claim" seems to fit the streetlight description well here, no?
I do use the term 'directionally correct', albeit spelt out, for explaining things to my normal friends. Just the other day I said something like 'AI isn't as advanced as you might hear from CEOs, but you'd be more correct believing the CEOs than you would be believing the news', essentially a re-phrased version of directionally correct.
Still, I think that outside of low-time and low-information zones, such explanations ought to be avoided. But not everyone has the desire to spend all of their time studying esoteric things, so it shall remain used, at least by me.
The quality of American discourse on crime is very poor because crime in America is so fundamentally intertwined with questions of race.
For example, according to CDC WONDER data, in 2024, black males aged 15-34 died by gunshot homicides 22 times as often per capita as white males of the same age, 47 times as often as young Asian men, and, perhaps most interestingly, 5.8 times as often as young Hispanic men, who tend to be comparable in income and education, but not at all in terms of getting shot.
These numbers tend to make Americans very uncomfortable and drive many to switch to Shoot the Messenger Mode. For example, I've been asked countless times over the decades to explain what's WRONG with me that I'm not ignorant about basic statistical facts about my society, unlike all the good people who are empirically clueless.
Personally, my view is that knowledge is better than ignorance.
For example, tonight in the State of the Union Address, Donald Trump is expected to emphasize how cutting down on illegal immigration would reduce the crime rate. That's true, to some extent, but it's much less true for the worst crimes such as murder than most Republicans want to believe, because the murder rate is so incredibly high among Foundational Black Americans.
Of course, the liberal Establishment is at least as unrealistic, as we saw during the Ferguson and Floyd Effects, when elites got tens of thousands of incremental black lives murdered and splattered on the asphalt by making clear to the police to stop policing blacks as aggressively. But how many people know about not just the huge increase in murders but also the huge increase in car crash deaths among blacks when BLM was riding high in 2015-2017 and 2020-2022?
Okay, like, we're all in agreement here that a few ethnic cleansings would probably decrease the crime rate by some extent. You still have to convince people that ethnic cleansings are morally defensible and justifiable under the circumstances.
How is it morally defensible to demand less rule of law for blacks in the name of Black Lives Matter and then you immediately wind up with 44% more blacks dying by homicide and 39% more dying by motor vehicle accident in 2021 than in 2019?
If the crime is mostly about race, does the decrease of crime mean that the ethnic profile of America is changing in a good direction?
Or is it more like, if you believe that the ethnic profile of America is moving in a bad direction, and that the crime is mostly about race... then you are forced to believe that the violent crime is increasing, regardless of data?
As the American Establishment increasingly forgot in the early 21st Century how disastrously they'd screwed up during the soft-on-crime 1960s-1970s, they increasingly fell once again for the myth that the big problem facing African Americans was too much law and order.
So, from Ferguson onward in August 2014, the Establishment promoted the Black Lives Matter movement. That led to BLM triumphs over local police departments in the St. Louis area in the fall of 2014, in Baltimore during the Freddie Gray riots in April 2015, and in Chicago from November 2015 onward, all with disastrous local effects on the black-on-black murder rate (and similar if less spectacular effects on traffic deaths as fear of being pulled over declined).
After BLM terrorists murdered 8 cops in Dallas and Baton Rouge in July 2016, Trump got elected and BLM fell out of fashion. The national murder rate dropped into 2019, but started to inch up again.
Then came the events of the last days of May 2020, and the black-on-black murder and traffic fatality rates soared within days and remained high for several years.
My first suggestion: Let's not due that again. Let's allow ourselves to learn the data.
Let's say the crime rate goes up. And let's say that everyone takes heroic measures to stop crime. They lock up all the goods in the stores so that nobody can steal them. They raise taxes and spend the money on extra policing. They stay home at night and they avoid dangerous neighborhoods. They install expensive burglar alarms. These measures succeed in reducing crime, which is now back to its original level.
Is it fair to say that crime has gone up? It is literally true that in this scenario, the crime rate is no higher than before. But it's also true that this is the result of an equilibrium which leaves everyone worse off because of (what would otherwise be) an increase in crime.
This is why the malicious streetlight effect matters. "Low crime rates" is the similar-sounding but different concern. People are actually concerned with the effects of crime. High crime rates lead to bad effects from crime, of course, but so do "normal" crime rates that are only normal because people drastically reduced their quality of life to stop crime.
(Also, why didn't you take into consideration changes in age of population (older people commit less crime) and changes in racial balance? Those drastically affect crime rates.)
I'm not expressing confusion. I'm saying it reflects a problem, a bias, that corporate and white collar crime are so often excluded when people talk about crime.
I don't see why Scott should give even a semi apology, and don't think it will help much with this problem for him to tell people in advance, "don't worry, I promise to post next about disorder." What's going on is that people are cranky and walking around with various resentments, and they feel a craving to express their feelings and related opinions. Then if somebody says something that appears to undercut the validity of some of their negative views, they feel a craving to argue. And because a lot of what's going on is mostly driven by irritation, they don't bother trying to argue well and fairly, but drag in stuff that's adjacent to the topic but not part of it -- for example, dragging in the Sidewalk Homeless Turd issue in a discussion of crime. Those doing this can distinguish between Sidewalk Homeless Turds and Crime with 100 IQ points left over. But they don't bother to do it because they are pissy.
Look, I have myself done the exact thing I am complaining about here. I get it. At the same time, I think that staying in emotion-driven mode in a situation where the task is to discuss and think fairly and clearly is by far the most common thinking error, and absolutely dwarfs all the rationalist-identified ones, and we should all try hard not to do it. It's not hard at all to recognize that you are in that mode: If someone's data immediately makes you angry, do some introspection. If the idea that their data might actually be *accurate* makes you feel sort of desperate and even more angry, then you know for sure that you have bonded with an opinion based on something other than good evidence. It sits well with you somehow. . You should probably abstain from discussion of the issue under dispute until you can figure out why you married Comforting Opinion and get at least a temporary legal separation.
In short, I think discussants, not Scott, should change. Don't type just to scratch your irritable itches. Jeez.
> Those doing this can distinguish between Sidewalk Homeless Turds and Crime with 100 IQ points left over. But they don't bother to do it because they are pissy.
Public defecation is generally illegal across the US. Whether it's a "crime" or a "civil infraction" depends on the exact jurisdiction; it could also be prosecuted as things such as public nuisance, disorderly conduct or public nudity, which are usually crimes.
While the exact classification as "crime" or "thing that is almost like a crime but officially not a crime" may differ from jurisdiction to jurisdiction, I don't think people need to get a law degree before they're allowed to complain about people around them doing things that are obviously both illegal and harmful.
Well Melvin, my point is that it just gums up the works to post about how bad disorder is in a discussion of crime, and that people posting here are capable of distinguishing between crime and disorder, and should not drag in problems better categorized as disorder. Perhaps I chose a bad example, or at least an imperfect one, if in fact producing a Sidewalk Homeless Turd is sometimes treated as a crime. In that, case, I should change my example to something like Being Disheveled and Stinky in a public place. Seems to me, though, that you must surely grasp my point, which is that there a lot of unpleasant things other people can do that fall under the general heading of disorder rather than into the crime category. If you are going to be a Turd Categorization Stickler about my example I will happily let you win regarding that sub-sub-issue.
Yes, Being Dishevelled And Stinky would have been a better example.
But I do think it's worth being aware that most of the examples put forward of "disorder, not crimes" are themselves still crimes, as argued by another posted elsewhere in the thread.
Maybe there really are circumstances where you really do see an increase in dishevelled and stinky people in the streets who are nonetheless not committing crimes. Maybe it's just an anime convention or something.
"I found that this was true - illegal immigration had shifted from Mexicans to Hondurans/Guatemalans/Salvadoreans etc entering through Mexico." My favorite part of this "favorite" thing of yours from 2016 is that - when presented with this clear evidence NOBODY jumped to the obvious conclusion that the US should pay MEXICO to secure their (much, MUCH smaller) Southern Border - serving both countries' security interests for a tiny fraction of the effort and expenditure in one go.
"Hey, a lot of people are crossing through and trespassing on your property and mine. But they're all coming though that gate you have. I'll happily by *us* a combination lock that you can put on there to secure the gate...how's that sound?"
In this post Scott is using "directionally correct" in a way that I don't think is consistent with its 'original' usage or intent, or at least not the way i've seen it used everywhere else until now.
In this post he takes it to mean "slightly stronger than the truth can support. If your enemy committed assault, say he committed murder. If he committed sexual harassment, say he committed rape." By this definition the claimant is being hyperbolic and asserting the extreme version of the 'actual' situation. It's implied the claimant knows what is actually true and they are trying their luck at a juicier accusation in the hope it sticks. Isn't this just a motte and bailey, where the claimant reserves the right to retreat back to a claim the motte is 'in the same direction' as the bailey if the bailey gets undermined?
To my knowledge "directionally correct" was popularised by Peter Thiel in 2019 or a bit before. By contrast to Scott's usage, this was a way of saying e.g. "there may be exceptions, nuances and edge-cases my argument doesn't handle, but in general it's the best argument". In other words, the argument is good enough such that if we followed it to it's eventual policy conclusion, we would be heading in the right direction.
The massive difference in intent between the two usages, and the very high Pr(misuse of phrases by people posting online) is enough to make me much, much more reluctant to react in the way that Scott seems to be in this post.
No matter how Thiel (I guess?) meant it, the vulgar use of "directionally correct" that Scott uses here is the more common way the phrase is used these days.
I think the root cause of the disconnect is that complainants' concerns are mostly inchoate, but imprecisely rounded off to "crime" or "disorder".
Rebuttals to what is said fall flat because the actual problem isn't well-reflected by the specific terms used.
Everybody's talking past each other and, given the parlous epistemic environment of the 2020s, both sides dismiss each other as irrational, deceptive, confused, &c.
Scott, you are a psychiatrist. You should get the hint that "crime is up" actually means "I feel unsafe" and then take a swing at why people feel so. Why one does not really see the 1960's style very carefree strolling down the street?
It's not lying, it is expressing an emotion in ways that are socially customary.
Can we say there is just generally more aggressivity in the air? Of course that can never be proven by the numbers-based approach you prefer. We might try measuring stress hormones tho.
Perhaps the correlation is the other way around. Perhaps 1960's people were hardened by actual crime and thus carefree, and today we got soft and anxious because nothing bad ever happened to us.
Or perhaps the psychological effect in play here is "I, personally, myself, am getting older and more frail, and less in connection to the youth culture so I cannot tell which parts of it represent real aggression and threats of violence and which are just playing around and fronting, so thus I personally feel unsafe regardless of whether this represents an actual rise in the amount of general unsafety".
Within echo chambers, people are bombarded by so many data points (whether true, misleading or outright fabricated) from "their side" that pointing out that a claim is inaccurate or even completely made up sounds like nitpicking or bad-faith argument. The claim you are knocking down is never a "pillar" of their overall view such that knocking it down would cause them to reconsider the whole structure; it's more like "I have 9,000 reasons to believe this, I'll give you two," in which case correcting the two just doesn't move the needle.
In fact it ends up sounding almost like a conspiracy theorist. Our usual response to "Why is the flag on the moon waving" or some anomaly with the video of the twin towers collapsing is not to take it seriously and investigate but rather to dismiss it due to our having been immersed in the overwhelming knowledge that a man did land on the moon and that 9/11 was a terrorist attack.
I think that's what's going on with "Their actual concerns were about disorder, open-air drug markets, tent encampments, and seeing people fencing stolen goods. They thought I was being deceptive in trying to trivialize these by saying that a similar-sounding-but-slightly-different concern, major crime like murder and assault, was down." I doubt it's that they're concerned about encampments rather than murder. Rather, they've taken in so many data points about crime and disorder generally that it's built up a "forest" of "Crime and disorder are shooting up everywhere because of [insert explanation here]" and then the point about murder rates decreasing sounds like trying to mislead with statistics.
I've had people of all ideological stripes tell me, when I showed them that a particular data point they had claimed was just flat-out wrong, that "you're missing the forest for the trees." I think that's the larger trend here, and it's easy to see how this is driven by internet-based news and social media.
Hate and angry sells. I think keeping us divided is part of the goal. My only solution is not to pay attention to the news. (I wish there was a better solution.)
Nothing to do with the above, but just because it's my personal question for people who worry about how divided we are-- do you try to talk to people on "the other side" (whatever that would mean in your case)? Have you had successes with substantive dialogue? I mean significant, ideally 1-on-1s.
I don't feel like I'm on any side. But yeah I talk to people on both sides. It's hard to have substantive dialogue unless there is a decent amount of trust from both people. But if so then sure I'll push back on what I think are wrong or misinformed views... but then again who am I and perhaps it's I who is misinformed. So listen and disagree and be friends.
Frankly there were a number of replies to your posts that clearly didn't read it, or understand it "did you read the post" that I would have considered handing out bans. At what point is "but you didn't consider X" when you have an explicit paragraph addressing X a violation of two if the gates (true, useful). If they aren't going to read your stuff, why should we read theirs. Your restraint is nothing but admirable.
The fair conclusion I've seen is that violent crime is down or steady, but only because people have taken costly steps to reduce their vulnerability to it that they wouldn't have needed to take some years ago. Going from ten robberies a year in a town where everyone leaves their doors unlocked to nine robberies a year in a town that would have one hundred robberies a year if people still left their doors unlocked isn't a reduction in crime by any meaningful understanding of the term.
I think there's some talking past each other on this issue.
Comments: "I experience an unacceptable level of crime and disorder in my daily life, and I'm upset about it!"
Post: "Statistics show crime is down. The things you are talking about are not really crimes."
Comments: The statistics must be lying.
Post: The commenters are being disingenuous and invoking logical fallacies.
The problem is people don't like being told crime is down because it appears to diminish the importance of an issue they feel is highly relevant.
In the past a higher level of crime may not have been as salient an issue as it is now because there were other issues of greater importance. In the present time, many other problems are not too bad so the highest priority issue that people encounter is "crime" even though there is on an absolute level less crime than in the past.
So, crime is down, but people are justifiably unhappy with the residual crime they encounter and want it to be a high priority public issue until the crime is further reduced.
<The problem is people don't like being told crime is down because it appears to diminish the importance of an issue they feel is highly relevant.
I agree that's the dynamic, but I think that both Scott and other commenters should call out people who are slipping into some gray non-rational zone where disorder is sorta a crime and they are infuriated by the dogshit in the park and anything that infuriating has to be a crime, therefore we should not be talking about why crime is down, but about dogshit, which is also a crime because they hate it, and also because DOGSHIT, QED
Here's Shankar calling out somebody for that kind of thinking in another context:
Person A: My opinion is that most cultures will think it [exploitive sexual misbehavior by the wealthy] is not a big deal
Person B:Considering there have already been real consequences for high-ranking figures like former Prince Andrew and members of the British government, that prediction has already resolved to false. Also it's not just about the sex, it's also about Epstein's Russia and intelligence agency connections.
Shankar Sivarajan: You're being disingenuous. The rest of the comment makes clear A’s talking about the differences in sexual mores.
I understand what you are getting at, but I think part of the disagreement is your example. You say dog shit in the park is not a crime. Which is technically true, but a distraction in that it is actually against the law to not pick up your dog's shit. I don't think it's really a satisfying argument to base your argument on whether the unlawfulness is handled by a civil or a criminal court.
I would like to see the law against dogshit enforced. I don't care whether the violators go to jail, get flogged, or pay a fine as long as *something* happens to them to dissuade them from this behavior.
The sexual stuff you mention seems to me an entirely different cultural issue. I don't accept the analogy.
Apologies if this point has been made already, but how would you (or anybody) know if crime is up or down if the reporting/recording entities are untrustworthy? How would we know the level of unreported crime? How would we know the level of reported crime that is unrecorded by the police? Why are you so sure the streetlight is operating at all?
This is why a lot of folks focus on homicides and auto theft as proxies for violent and non-violent crime trends. Both aren't likely to see a lot of unreported crimes.
Prioritizing which datapoints to use for an analysis *is* choosing which datapoints are directionally correct. Pointing out an analysis has conflicting data points is insufficient to determine if an analysis uses motivated reasoning, you have to show the conflicting data points are better proxies for the intended effect than the initial analysis.
My point is that it's significant that people talk broadly about crime rates and exclude types of crimes that cause significant amounts of societal harm.
both failure modes you describe are really motivated measurement: you choose what to count after you already know what conclusion you want. "directional correctness" is the more socially acceptable version because it lets you weaponize genuine victim status as a shield against precision. the fix probably isn't better fact-checking but declaring your priors before you pick your metric.
I don't use the term directionally correct this way and have not heard it used this way? I use it more in a context of high epistemic uncertainty or where there's a bunch of highly non-ideal options only. "Directionally correct" in this context means "least bad."
If I'm trying to get to New York from Houston and the only offers of rides I have are to A) Dallas, b) San Antonio, C) El Paso, none of those will get me to New York, but Dallas is the most directionally correct. It's at least going in the sorta correct direction to get to New York as opposed to basically the complete opposite way like the others. I will be a bit closer to New York if I take the ride to Dallas.
But the issue is that people are unhappy with crime/economy/etc. The justifications they give for it are completely irrelevant. It's the job of policy makers to figure out how to make them happy. If they do policy based on facts, and the results don't make people happy, then they've failed at their job.
> They reward politicians who do things that align with their incorrect beliefs about reality, and punish politicians who do nothing and/or deny the “problem.”
The fact that they are unhappy is real. They might be wrong about the reasons, but they are unhappy. The customer does not know what they need. It is the job of the provider to find it for them.
> Is the problem politicians or voters?
The problem is democracy. If you support democracy, you're not really in a position to complain about the inevitable consequences of that system.
That's one of my favorite folk tales/parables.
Here's a slightly longer version:
https://www.uua.org/lifespan/curricula/grace/session14/115603.shtml
Wait - why did you not address corporate crime at all? That's arguably an equally important dimension of crime and stating "crime is down" without adjusting for corporate crimes seems quite inadequate.
Should havd been "adjusting" not "a listing"
use the edit button
Thanks. New to this platform. Didn't see an edit button on my phone interface. Switching over.
Yeah, there's no way to edit comments from the phone app. There are lots of weird little functionality gaps between the web version and the app.
i'm pretty sure this isn't true? will test it right now
edit: edited from my phone, by pressing the 3 little dots and clicking 'edit' and then appending this message
If this is meant to be very subtle irony, color me impressed.
Lol. No, I'm usually a subtle irony guy. But not in this case. Corporate crime results directly in massive morbidity and mortality. There's no particularly good reason to exclude it when you're talking about crime rates.
The reason to treat it separately is that it requires a completely different policy response from violent crime and petty crime.
Also, that's not what right wing reactionaries are referring to when they claim that "crime is up." Nor do they support the re-regulation that would be required to suppress white collar fraud.
Well, that's a reason. But I think part of the best reaction to RW reactionaries is to discuss the selectivity in how people consider the societal harms from "crime."
Yes on the first sentence, but you're wrong on the second. It's not the 80s or even the Tea Party era anymore. The current wave of RW reactionaries are all over social media complaining about banks and institutional investors, sounding as much like Elizabeth Warren as anything. There's enough historical animus over taxes and some Obama-era abuses that they still hate IRS agents, but there's certainly a strange new respect for FTC and anti-trust enforcement. Josh Hawley comes to mind as a model of this. I think lots of current RW reactionaries would be happy to see more law enforcement tools in that arena, provided that they were going to be operating solely against Blackrock et al, but they would be afraid that when Dems got back into power it would be used to persecute the pillow guy and a handful of random Trump-loving boat dealers in Alabama.
RW reactionaries basically discount massive amounts of corporate crime (Trump talking bribes, Elon gaming regulations, etc.) even as they obsess over one white girl dying in a car driven by someone without documents. I don't think your description holds up.
In theory, yes, there's some right-wing populism that complains about banks and tech companies and monopolies. In practice, I don't see Republicans doing anything substantial about it, at least not right now. Trump will throw away any principle at all to "make a deal" with anyone who sucks up to him enough, which in practice means that those companies will continue to get away with whatever they want as long as they pay him off.
But even putting Trump specifically aside, Republicans' proposed "pro-American-worker" policies are largely beneficial to those same corporations. Tariffs, repeals on environmental regulations or labor laws. These things have *some* benefit to workers, but the overall net effect is probably negative to regular people and positive to large companies. I don't see Republicans actually engaging in meaningful antitrust action either.
Most of the Reaganite/market above all conservatives are either very old or jumped ship to the Democrats a while ago. Trump is sort of a bridge between the old GOP establishment and the rising populist right that is more open to questioning the neoliberal economic order. I think future rightist politicians will be more amenable to at the very least anti-trust and economic protectionism.
First, maybe so, but that would need to be established. In a real world of limited resources, policies to address the one don't exist independently of policies to address the other. Second, this was a treatment of "crime rates," not a treatment of policy options.
And for how long has Minnesota been hosting the fraud?
And other states, since the more we search, the more we find?
Is the crime rate climbing if the corrupt cops have the mask pulled off?
Industrial espionage can lead to outright murder. It's been known to happen, and it isn't always reflected in the murder statistics.
Another reason to treat it separately is that it's very difficult to condense into a rate. You very quickly run into the "is this one crime with ten thousand victims or ten thousand crimes" problem.
+1
Murder, tax evasion, software piracy, and dumping toxic waste in the river are all bad things we should want less of, but they are very different bad things. Their impact and the sorts of things you might do to prevent them are all very different, and it is hard for me to see how glopping them all together helps us think clearly about anything. To decrease the murder rate, probably you want to arrest/keep in prison the most violent young men who are committing other crimes, make it hard for criminals to have guns, hire more police to patrol the streets, stuff like that. To decrease tax evasion, you probably want to hire more IRS agents, and maybe revise the tax code to be less complex. To decrease toxic waste dumping you want more environmental monitoring and more people working for EPA/state equivalents. And so on. The solutions to one of these problems have nothing to do with solutions to the others.
At best "corporate crime" is a way too broad category for this discussion.
Some of it increases morbidity and mortality, but the category also includes accounting irregularities and and uncountable other bureaucratic missteps.
Witness how all the fraud ate up money that autistic people *needed*!
> Corporate crime results directly in massive morbidity and mortality.
Maybe I'm dense, but how? Like the most direct example I can think of, health insurance companies killing sick children and grandmothers for profit by denying coverage / care, which they do to the tunes of tens of thousands of people every year, is legal.
Selling people junk and fast food so 80% of people are overweight (and have ~2x all cause mortality?), also legal.
What corporate crimes that are actually illegal directly result in massive morbidity and mortality?
Two words. Purdue Pharma.
There are tons of examples:
PG&E pleaded guilty to 84 counts of involuntary manslaughter for the 2018 Camp Fire.
CEO Don Blankenship was criminally convicted of conspiracy to violate mine‑safety laws related to an explosion that killed 29 miners.
C'mon, that's literally rounding error.
There's ~5k workplace deaths per year, about half the number that are murdered every year, and roughly 10% of the number that die in car accidents every year. And that's assuming all those deaths are due to corporate malfeasance, when it's probably something like 500 of the deaths.
If we focused on legislatively making it easier and faster to roll out self driving cars, we'd move the deaths needle 50 fold more.
If we passed out free Ozempic to everybody, it would halve mortality in the ~50% of the population who are obese and over 40, and we'd save 600k lives per year.
Way bigger interventions are available before going after that 500 people.
Would you consider examples like the Volkswagen Diesel-scandal? They knowingly and actively cloaked their car's real emissions from authorities. The death count is much higher, but also not directly traceable to their source on an individual level. But epidemiologically, it is undisputed that it increases mortality.
In other cases, there is a lot of "risk-skirting", where management encourages risky behavior that is not per se criminal, but can result in e.g. ecological catastrophes that outright kill people directly or indirectly. For example, Boeing and MCAS (Boeing pleaded guilty in a favorable deal), or various fossil fuel drilling and pipeline companies. Would it be fair to include this or not? And why?
Those are just examples and weren't meant to be anywhere close to comprehensive. And please, seriously, you're going to downplay the magnitude of the effects of the crimes committed in relation to the opioid crisis?
I mean, that goes into definitional issues: the people Scott was answering to (like the neoreactionaries) definitely don't mean corporate crime, they mean the kind of crime people experience directly and first hand. Their general argument is that the liberals are too soft and worried about looking racist to address the real causes of such crime.
I think it's important to point out the selectivity in how those people address the impact of crime. Just answering on their defined term, without pointing out the embedded biases, misses a large element in the societal impact of crime.
True, but also, these different layers affect different levels of survival instinct and Maslow hierarchy. If an exec embezzles money or plays recklessly with investor money, I can lose my savings or pension, which is awful and hurts my future, but in a long term perspective. If a mugger stabs me or a rapists assaults me (for a woman), I get immediate pain and danger, possibly death or permanent damage, and trauma that has been wired into our minds by evolution because it's pretty much THE definitional fight-or-flight kind of situation.
Basically I don't think it's weird that people weigh this kind of crime as something they perceive more dramatically close to home. The utilitarian perspective is "which does more harm at scale", and that's the rich finance guy. But the personal perspective is "which does more harm to ME if I am unlucky enough to fall victim to it, and which is scarier", and that's the petty crime. Another common moral consideration tends to be the fact that the rich guy has more moral choice - he's more educated and has less needs, he does crime only out of greed, whereas the thief or robber may be poor or mentally ill or from a background where he was never taught that crime is wrong. And that may also be true, but ultimately, getting hurt is getting hurt, I can only care so much about what specific excuses the culprit had.
I'm not saying it's "weird." It aligns with pretty standard cognitive biases like availability bias, temporal discounting, proximity bias, intentionality bias, etc. I'm arguing that it doesn't necessarily reflect the actual harm at the societal or even individual levels.
Oh boy, I appreciated the irony, even though I didn't think it was very subtle. But now apparently it wasn't ironic
What's the reason you think considering a form of crime that causes great amounts of societal harm shouldn't even be considered as part of a thorough and comprehensive discussion of crime rates?
Nothing at all, just not in response to an article that's literally about how if you say crime is down people will find some type of crime is not down and then generalize it to say crime is not down
Sure. I get that within that narrowed context of responding to particular claims, the distinction isn't directly on point. I still think it's important, however, to point out the larger biases embedded in claims people make about crime rates.
I laughed!
I'm more interested in limited liability partnership crimes.
Lol. Funny. But workplace injury resulting from corporate malfeasance are real. Environment and public health harms resulting from corporate malfeasance are real. Product safety harms from corporate malfeasance are real. And then there are indirect effects from financial crimes and fraud. It is a deliberate choice to exclude these factors when describing crime rates. What's your rationale for doing so?
Are all workplace injuries the result of corporate malfeasance?
Who says all? Not me. Put words in mouth much?
Bit of an ironic response. I just asked a question, I didn't put any words into your mouth! But implications do exist, don't they...
I was just trying to see where you were coming from. Whether it was a starting place of:
"We need to do dangerous jobs to make the items we all depend on. Sometimes people will get hurt. But, evil management can sometimes exacerbate things so that people get hurt at a much higher rate than is needed"
or
"Everything should be able to pop into existence, ready-made, with no danger to human life or limb at any step of the process. But evil management wants to make things in a life-threatening way because they are so evil"
I am glad you are *probably* starting from the first one, even though my clarifying question failed to induce a clarifying response.
"I just asked a question, I didn't put any words into your mouth!"
Not a fair question. It was loaded. It's implying he (might) think that.
What he said (in the first post) is implying there are other sources of "workplace injuries". (Which makes your question, as written, pointless.)
=============
It looks like the "place he's coming from" is "corporate malfeasance isn't some abstract thing. It can actually have an understandable affect on 'regular Joe' by sometimes resulting in workplace injuries".
I think most of these have clearly declined, and so are not worth arguing about. For example, consider this photo: https://en.wikipedia.org/wiki/Lunch_atop_a_Skyscraper
There's no argument to be made that workplace injuries occur at a higher rate now than in the past. Ditto for environmental impacts. Not sure about financial crimes.
So we should just assume they've declined rather than examine whether they have?
OSHA and the EPA will have data on both these topics. I would guess they all kinda generally go in the same direction, but if you think I'm wrong, by all means....
How about you go and examine, then get back to us?
Lol. Investigate it as you will. I haven't made a claim about trends either way. My point is that trends of more categories of crime should be addressed under a thorough and comprehensive treatment of crime rates.
Trump admin. has moved to end EIS/NEPA so “data” will have to have an asterisk going forward.
I mean, Musk is being granted a national wildlife refuge, home of listed species, that was initially purchased by the Nature Comservancy. This is incredible.
Never Trust the Government. I mean, you do realize that, right? They have a job to do, and it's not "preserve the ecosystem" except when DARPA says it is (and then, you're looking at the Department of Deer Warfare).
What has been the overall trajectory of workplace injuries over time? Is that offsetting the decline in assaults Scott blogged about?
I don't know.
"Worker injuries and illnesses are down—from 10.9 incidents per 100 workers in 1972 to 2.4 per 100 in 2023" https://www.osha.gov/data/commonstats
"Private industry employers reported 2.5 million nonfatal workplace injuries and illnesses in 2024, down 3.1 percent from 2023" https://www.bls.gov/news.release/pdf/osh.pdf
That's a huge drop. Interesting, as I wouldn't have expected that.
Violent crime creates larger second-order effects (fear-driven behavior changes, depression of property values, erosion of social trust, security costs, etc) than white collar crime. Most white collar crime seems avoidable because a) it doesn't generally target individuals and b) you have to first choose to do business with most white collar criminals. I have no fear that the next Bernie Madoff is going to bother me, for example. Violent criminals on the other hand will bust down the door to your home. That generates much more fear and avoidant behavior in people.
The other consideration is that white collar crime exists on a more-or-less continuous spectrum with normal net-social-positive business transactions. The distinction is often one of perspective and of somewhat arbitrary laws about what is and isn't legal: it's not like most white collar criminals stab people. That means that there's no clear way to crack down on white collar crime without exerting some nonzero chilling effect on regular business activity and so people shrug and accept most forms of white collar crime as the cost of doing business. Obviously there are exceptions like SBF and Madoff, but I suspect it's actually a rational attitude most of the time.
I also suspect that violent crime is more strongly auto-catalyzing. Its successes are public (a dead body, a robbed store, a smashed window) which signals to the marginal would-be criminal that maybe they could get away with it too. White collar crime is generally only apparent when someone gets caught and therefore has built-in anti-mimetic properties.
Do you think wage theft or embezzlement are what people are talking about when they say that cities are more disorderly?
It seems obtuse to put corporate crime in the same category as mugging in this context.
I'm not putting them "in the same category." But they're both crime, and they both impact societal safety, heath, and well-being.
Well, the article wasn't meant to address all things that impact societal safety, health and well--being. It was meant to addess the things like murder, vandalism and mugging which some people call "crime".
Would it be useful to include wage theft? Probably not. The people pointing to high "crime" rates usually are advocating for policies like increased police presence. The rate of wage theft would not be affected by that policy so isn't relevant to the discussion
Sure. But that seems circular to me. I think there are biases embedded in how people separate out different forms of crime, even forms that directly (and indirectly) affect safety and health on a massive scale. We make choices in how we spend limited resources to address societal harms. The choices we make in the public discussion of crime rates, where an entire category of crime isn't even included, arguably leads to a misallocation of resources when we consider rate of return.
What are your thoughts on the justice creep article?
https://www.astralcodexten.com/p/justice-creep
I think you're doing something similar by using the shared label of "crime" to complain that people focus more on mugging than wage theft.
It's reasonable to argue(like you are) that corporate malfeasance is more harmful than mugging. Just like it's reasonable to argue that things like agricultural policy (which isn't usually called a crime) are more harmful than both. But pointing out that we use "crime" to refer to both corporate malfeasance and mugging doesn't make your argument more compelling than a similar argument that agricultural policy causes more societal harm than mugging and should therefore get more focus.
Edit:
For example, the following things happened to me while I lived in North Seattle:
1. Someone broke into my apartment and stole things
2. My job did not allow me to take mandated 15 minute breaks
3. I got hit by a car while biking to work and was hospitalized
All of these harms might have been prevented by policies like sweeping the nearby encampment (where I found my discarded backpack) or designing streets better. We typically only refer to the break-in as "crime" in these discussions while advocating for policies like sweeps. Your argument is that we should also include the 2nd harm(but probably not the 3rd) when we talk about ccrime. I don't think that's particularly helpful.
When companies falsify safety data, cut corners on environmental protections, ignore workplace hazards, or market unsafe products, the result can be toxic exposures, injuries, chronic disease, and even premature death. These harms are every bit as real as those caused by street crime, they just show up through polluted air, unsafe working conditions, defective products, or contaminated water rather than a single dramatic incident.
Unless the car had right of way and you didn't, 3 should definitely be counted as a crime, if you ask me.
How do people spend their individual resources in response to crime? Economists look to that sort of thing when measuring the costs of crime and asking how much people would be willing to pay to avoid it.
"The people pointing to high "crime" rates usually are advocating for policies like increased police presence."
Really? I did not get that impression. What I mostly saw was people complaining about "Well now I stay in at night and lock my doors, so all the muggings I've avoided don't show up in statistics. We've traded high crime rates for a lower quality of life". They want to go out at night and not get mugged.
Similarly, they (presumably) want low crime without there being cops stationed on every corner.
Now, the commenters pointing to high "crime" rates have not been all that forthcoming with what kinds of policies they have in mind. (See my question to Steve Sailer way back in Scott's first post about crime rates like a week ago. Sailer never responded to it.) Personally I suspect they want a policy in which some kind of unpleasantness is continually meted out on members of their outgroup, who they blame the crime on.
Who would mete out this unpleasantness?
I just gave you my personal suspicion. It would be easy for anyone to deny that it is true and impossible for me to prove it, so to give you any more details is kinda pointless.
But I think when people fantasize in unarticulated ways about unpleasantness being meted out on members of their outgroup, even if it would be cops (or some other special law enforcement agency) doing the meting-out if those fantasies were brought to fruition, that doesn't look the same in people's minds as cops standing around deterring criminals with their mere presence but also reminding non-criminals of the nearby threat from criminals.
People are complicated and brimming with cognitive dissonance too.
Sorry, but I did reply:
You asked:
Nadav Zohar
6d
What's your policy proposal?
If Steve Sailer were king, what would he do about this?
I replied:
Steve Sailer
5d
We've had three periods in my lifetime in which liberals held the upper hand in criminal justice policy: the Warren Court Era, the Ferguson Effect, and the Floyd Effect. All three led to explosions in crime, especially black on black murder.
So, first thing: Let's learn from that unfortunate experience and Not Do It Again.
Second thing: We've seen New York City radically reduce its murder rate by focusing on point-of-use gun control to change the culture of criminals: make lowlifes more afraid of the cops than of each other, so they leave their illegal hand guns at home.
It shouldn't be that hard to learn from the world's most famous city.
Jeez, you're right, I'm sorry. Somehow I didn't get the notification. I retract what I said about you not replying to that question.
But I do see I was correct in a sense, that you want your outgroups punished in some way (in this case, not allowing liberals to hold the upper hand in criminal justice policy). I note also this is a different outcome than increased police presence.
I thought NYC had general restrictions on gun ownership with very few exceptions.
There's a very simple test about which "type of crime" is has "worse" impacts: which sort of criminals would you rather you and your family live near?
If you have to live on Street A, which is home to, in order, a rapist, a murderer, a carjacker, a serial duelist, and a burglar, or Street B, which is home to an embezzler, a wage thief, a tax fraud, a corrupt politician, and a digital pirate, on which street do you think your physical safety and your mental peace is more likely to come under direct adrenaline-spiking possibly-violent threat in the future?
I'd rather live next to the wage thief and tax fraudster, sure. But I'd rather be employed by the rapist and duelist.
I think women would be less keen on employment by the rapist.
What kind of duelist are we talking here? The Princess Bride kind or The Good, The Bad, and The Ugly kind?
The Blade Runner kind, where your enemy's final act of mercy proves that he, and perhaps you too, is more human than the man that made him.
That's a terrible test. I'd rather live next to Bernie Madoff than a carjacker, but Madoff has caused more harm.
I phrased the test without the right amount of delicacy; my quotation marks around "worse" were meant to communicate "more urgent and pressing impacts to your physical safety". I concur that in the aggregate most white collar criminals directly and indirectly cause more "harm" to society.
When people talk about "crime", they almost always mean "urgent and pressing impacts to my physical safety". White-collar crime indirectly threatens the prosperity and safety of your society in often difficult-to-quantify ways; "blue-collar" crime threatens to delete you tomorrow on your way to the corner store. Of course it's reasonable to be more finely attuned to the latter type.
Yes, and the other poster is arguing that that focus on the immediate is irrational. Your "test" doesn't address that.
Living next to someone is one vector for measuring the impact of crime. There are others. IOW, which society would you rather live in, one with a good environment and relatively little poverty and some property crime, or one with a degraded environment and high rates of poverty and a police state that prevents property crime?
I think living besides our prime minister (you probably heard his name, although our country is a smallish spot even on the map of Europe) might be quite OK-ish, still, his actions wreck the whole damn country.
Dictators with really bloody backgrounds can be quite charming in private, as well.
They don't impact physical safety. They're basically like taxes or bureaucracy. If I get paid less because someone in my company is embezzling a ton of money vs it's getting wasted on pointless bureaucracy, it makes no difference to me. But if someone mugs me, I'm worried about a lot more than money.
Corporate crimes, of course, impact physical safety. Illegal dumping, bribing law enforcement, etc.
What portion of corporate crimes do?
In your opinion, is solar f*cking roadways a corporate crime?
My opinion is that crimes defined legally as crimes, are crimes. Not really that complicated.
Something I find a bit annoying about this whole conversation is that personally, I rarely if ever hear people complain about "crime". "Crime" is not a terribly meaningful category.
I agree.
Next Door will have plenty of posts, if you care to look. Most people are complaining about a particular crime (porch pirates that are continual and not caught are a favorite)... but yes, people do complain about "crime in general."
And if they do, it's usually clear from context what subset of crimes they're talking about.
"I don't go downtown any more, there's too much crime" -- clearly they're talking about the sorts of crimes of which they are likely to be a victim if they go downtown.
That’s difficult to analyse. I would expect that: different size corporations make different size crime; distribution of sizes of corporations (including “number” of corporations) changed significantly over time; regulation changes significantly over time (and therefore the type of crime committed). For example, securities fraud today is different from 50 years ago. An analysis would need a very different technique than just looking at crime rates.
Because white collar crime is not what people are afraid of when they say they are afraid to go outside, it is not relevant in this context.
But that relates to cognitive biases. Yes, the threat of some forms of crime is more concrete and tangible. But that doesn't actually translate directly into it being less of a threat to health or safety. That's precisely the reason why corporate crime shouldn't just be excluded from a discussion of crime rates: to address the underlying biases. When we do so, we run the threat of irrationally addressing threat levels.
it is apples and oranges. this argument is about "if the apples are rotten", you are claiming that in fact "oranges are rotten", this being or not being the case has no bearing on the claims being made about apples, even if both are fruits.
Street crime and white collar crime generate differing types of threat, illicit different psychological reactions, and require completely different policies to deal with. they are only the same insofar as both cause social harm and break the law in some way.
Other people disagreeing with you does not necessarily make them "biased". Just saying that the label "crime" applies is what Scott pointed out as the non-central fallacy https://www.lesswrong.com/posts/yCWPkLi8wJvewPbEp/the-noncentral-fallacy-the-worst-argument-in-the-world
My point is that (imo) as a society, we direct our attention towards different categories of crime in a way that is disproportionate to the extent to which those different category of crime impact health and safety. That's what I mean by bias, as in other biases where attentional focus isn't purely "rational."
But I agree, disagreeing about the relative importance doesn't inherently mean "bias."
I think you've made this claim in many different words throughout this thread; that corporate crime should not be excluded from crime rate discussions. However, I don't feel you've really supported that claim with any data or evidence and have only supported it with "vibes".
It's an opinion, that when discussing rates of crime, the definition of crime should be inclusive of a form of crime that has massive societal impact. I'm not sure how one would provide "data" to support that opinion. Do you mean data that corporate crime has massive impact? Do you actually believe that it doesn't, and need to data to believe it so?
you started this with "Wait - why did you not address corporate crime at all? That's arguably an equally important dimension of crime and stating "crime is down" without adjusting for corporate crimes seems quite inadequate."
so argue that it is an equally important dimension of crime in way that is not just "because Joshua Brooks thinks so". It doesn't have to be data driven but at least a logical argument stating expected consequences by not addressing it as a dimension. or whatever it is that is driving your statement
Hopefully that makes my statement more clear.
When companies falsify safety data, cut corners on environmental protections, ignore workplace hazards, or market unsafe products, the result can be toxic exposures, injuries, chronic disease, and even premature death. These harms are every bit as real as those caused by street crime—they just show up through polluted air, unsafe working conditions, defective products, or contaminated water rather than a single dramatic incident.
What epidemic of corporate crime is supposedly going on? They got the Pinkertons down at the mine again?
Not sure why people keep putting words in my mouth. It's weird.
There was a huge explosion of low brow white collar crime during covid as the government did helicopter drops of money thru various programs such as paycheck protection loans and meals for kids, such as in Minneapolis. Lots of dubious NGOs got their hooves in a lot of troughs.
On the other hand, throwing all that cash around with little oversight may have headed off a recession.
How common is it for people to "vote with their feet" away from corporate crime?
is there anywhere to go?
Different polities have different regulations.
Different industries and companies have different rates of compliance with occupational safety and environmental laws, and you can choose whether to work for a company that cuts corners on safety or near a company that is known to get away with pollution. Different states have very different levels of enforcement of laws involving environmental damage, fraud, and the like, and people can move from one state to another.
I expect this happens all the time, but there probably aren't statistics on it.
Also cybercrime -- I looked up stats on that and, not surprisingly, that form of crime has grown enormously over the last couple decades.
Yes, 50 years ago you'd never get scam emails from Nigeria. We need to factor this into any discussion on crime and disorder.
Try the stats on spam. People targetted the spammers, and they're mostly dead now.
(Solar f*cking roadways isn't spam, exactly. It is a scam, but... a relatively legitimate scam.)
This place is a hive of billionaire defenders and guys who love capitalism much more than Adam Smith ever did, so I predict you will get a chilly reception.
Already have. Apparently my comment merits mockery.
You are talking to the wrong audience. They are far less likely to be victims of any type of crime, but right now they're far more likely to be victimized by thugs rather than corporations violating workplace safety laws. Why should they care what happens to those below them?
Astute.
Corporate crime is also less "sexy" to talk about, harder to see, and it isn't as clear whether the people to blame are in this readership's (and Scott's) outgroup, so you're less likely to see it discussed here.
Yes.
While we're at it, why is "wage theft" frequently defined as paying less than the minimum wage (when did this start?) while the less frequent but possibly more serious crime of employers not paying what they promised gets ignored?
Do we have data for the amount of crime committed by corporate entities going back to the Dutch East Indies Company?
You are a great illustration of the Streetlight Fallacy. Scott says crime is down. You reply by saying, "What about corporate crime? That matters too." You might as well have said, "But war in the Ukraine has increased." Nobody worries about corporte crime, whatever that is. Crime means burglary, robbery, larceny, ... So you are changing the issue.
Great point. Except the war in Ukraine isn't a crime. But outside of that, very insightful.
I wish there was a LIKE, so I could use it to say "Thanks for reading and replying nicely",but there isn't.
Depends. I mean when you're watching neighbors stringing up their fellow Ukrainians, and calling it "catching Russian Spy"... that's lynching* by any other name.
*description deliberately changed to a more American-centric referent.
No one worries about big companies screwing them over and getting away with it? Really?
The rule of law in the US disappears at a certain market capitalization, if this isn't a problem for you you're extremely lucky
Well, this readership happens to be filled with lucky people.
Because he wanted to investigate “why peoples’ intuitions clashed with the data.”
I do not think that the general public has strong intuitions about corporate crime rates.
This is a good point. Also very interesting to think about. I bet if you ask Average Joe On The Street if he thinks corporate crime is rampant and out of control, he'll say yes. But he'll say it in a shrugging "whaddaya gonna do" way, as if corporate crime is something that happens on another planet. Contrast with how people talk about street crime, you can almost see the torches and pitchforks ready to come out--even if they have never been the victim of it or even witnessed it.
Homicide and drug overdoses (in absolute numbers and rates) are way down.
I’m a life actuary and I’m not going by “crime stats” but by CDC numbers, because they’re not waiting for adjudication (and it’s tough hiding dead bodies… it can be done, it’s just a lot of work. Only a few overlooked/hidden bodies show up each year.
I’m not trying to make a point about crime, other than various cities trying to claim their great policing did it, because it happened everywhere over 2024 and 2025. (A little bit in 2024, and a lot more in 2025). I doubt all these cities had great policing improvements simultaneously.
>by CDC numbers, because they’re not waiting for adjudication
Does that mean those CDC numbers can't distinguish between murder and manslaughter and you are using "homicide" colloquially? I think that's an important distinction.
The distinction matters for an individual homicide, but unless you think the *mix* has changed a lot then for overall statistical trends it won't matter.
>unless you think the *mix* has changed a lot
That's what I don't know and why I'm asking for clarification.
I believe that the distinction between murder and manslaughter matters for the question of what effect improvements in healthcare have on the murder rate. Murder implies planning and forethought that includes the knowledge of the state of healthcare, meaning a would-be murderer is going to put in more effort if they know healthcare for their victim is good and available. All else being equal, I'd expect improving healthcare to reduce manslaughter rates more than murder rates, shifting the mix towards murder.
Only first degree murder requires planning and forethought. Murder generally just requires that the killing wasn't an accident.
Manslaughter is stuff like a drunk driver killing someone in a crash.
(Exact categories vary by state, but that's the gist of it.)
While this is true, I don't expect competent murderers to show up in the statistics. There's enough poisons out there...
Here are your choices for “external causes of death”:
-caused by someone else, not accidental (I simplify this to “homicide”)
- self-caused, not accidental (suicide)
- Accidents (this is a huge category, including drug overdoses, motor vehicle accidents, falls, drowning, etc)
The data are aggregated from the state level, based on death certificates and subsequent determination (external causes are censored from the public for 6 months in CDC data).
CDC death data are separate from FBI, etc, reporting, which is useful for checking.
Is there some sort of official technical definition of "homicide" that excludes manslaughter? The definitions I'm finding on Google (both formal and informal) all seem to include it.
If someone kills me dead, I wouldn't care what their exact mens rea was. Super from the fact that I'd be physically unable to, the consequence it's the same.
Drug overdoses are absolutely not "down" in a meaningful historical sense. Vs. the all-time highs a few years back? Sure they're down a little. But vs. overdose deaths in the 70s/80s? Still way, way higher: https://www.cdc.gov/mmwr/preview/mmwrhtml/mm6346a19.htm
and also
https://www.cdc.gov/nchs/products/databriefs/db526.htm
Notably Scott did not mention overdose deaths at all, even though they are a pretty reasonable proxy for drug use (in the same way murder is a reasonable proxy for other violent crime). I don't think it is "malicious" per se but this difference stands in contrast with the other metrics.
They are down year-over-year after the huge spike up in 2020 and continuing rise in 2021
Yes, the long-term hideous trend of decades is pretty bad, but we barely had any turnaround, especially once fentanyl was in the mix.
Now, again, there was some noise made about widespread availability of Narcan helping with the recent reversal from the OD spikes of 2020/2021… again, don’t think that’s what is going on.
I think the illicit fentanyl supply has been reduced by quite a lot (starting in 2024), and that probably has a lot of knock-on effects. Perhaps some other stuff has been going on as well, but again, I don’t think it’s been due to great policing or drug abuse treatment programs. Not for such a large effect.
The Chinese used to manufacture fentanyl and just mail it to the USA.
The Chinese government cracked down on that a number of years ago, so Chinese manufacturers then made the raw ingredients for fentanyl and sold them to Mexicans, who mixed the final products and smuggled them into the USA.
Lately, it appears that the Chinese government has cracked down on this recent two stage method, so until some criminal perfects a work-around like last time, fentanyl deaths seem down.
Mary Pat and I like talking about the numbers, while most commenters want to talk about how to think rationally without dirtying their hands with any actual historical data.
Scott seems to be in the middle.
I'm not sure what Scott is trying to do.
I've been reading Steve for a really long time, so I know he's been going back to some of these external causes of death, even pre-pandemic effects. (aka, "Ferguson effect" stuff)
I'm a life actuary and have been following U.S. mortality trends for decades in my career, and I'm interested in multiple trends. When I saw the huge spike in homicides/drug ODs in 2020, on top of the COVID deaths, I was pissed. (Oh, and alcohol-related deaths also increased, as well as motor vehicle accident deaths... and some other stuff.) I follow lots of stuff... I work in the insurance industry, so yeah, I'm kinda invested in death.
I have a theory that some of the fentanyl precursors could have been making trouble in China internally, not just through Mexico/Canada routes to the U.S. Something seems to have gotten shut down by the CCP, I think, and not necessarily due to anything the U.S. is doing.
People get so tied up in U.S. politics that they think that whatever they care about must be the driver of whatever they're seeing, as opposed to looking more broadly at potential cause-effect chains.
I've heard some wind about the term life insurance industry being endangered, post the covid19 vaccinations*. You heard anything, working in the field?
*significant increases in deaths of the younger varietal (not preterm).
Sooooooo this is my day job and I get paid for this. ;)
(And also, I'm involved with this group: https://www.insurancecollaborationtosavelives.org/ -- check us out)
But public info you can look at:
https://www.soa.org/programs/mortality-longevity/mortality-longevity-research/
https://www.soa.org/resources/research-reports/2025/qmmr-us-population/
That's U.S. general population mortality, monitored by the Society of Actuaries (I'm a member, and I've been involved in past mortality studies there as well) -- the life insurance experience studies can be more difficult to come by if you aren't in the industry.
I'm not convinced that drug overdose deaths are a reasonable proxy for drug use per se; isn't the crisis almost entirely related to the advent of much more dangerous drugs than were available before?
That said, insofar as drug overdose deaths are the biggest actual impact of drug use, they may be a good proxy for what we actually care about when we talk about "drug use".
Yeah I think that's another thing to keep in mind - when people say "drug use" in the context of crime/disorder they don't mean high schoolers smoking a bong in the basement when mom and dad are out of town. They mean someone in downtown SF with untreated mental illness passed out with a heroin needle in their hand.
Right. I too encourage people to look at CDC homicide numbers, which come from doctors and coroners, while the demographics come from funeral directors going over a quick checklist with loved ones. There is no political pressure on undertakers to fiddle with the race or ethnicity data of homicide or car crash victims.
There is more political pressure on FBI murder stats, which are collected from local law enforcement agencies. But, in general, CDC homicides and FBI murders correlate closely together, suggesting that dead bodies get counted pretty honestly in this country.
The police are inherently political.
But the funeral business is not.
Murders in the United States exploded during the last weekend of May, 2020 due to the triumph of the Black Lives Matter movement over the police following the death of George Floyd on May 25, 2020. For example, the most murders in the storied history of Chicago murder was May 31, 2020 with 18.
In June 2020, 44% more blacks across the U.S. died by homicide than in June 20219.
Traffic fatalities exploded during the month of June 2020 as drivers noticed that the cops weren't pulling people over as much, much less searching them for illegal handguns. In June 2020, 53% more blacks died by motor vehicle accident than in June 2019.
Eventually, American elites started to figure out that Black Lives Matter was losing popularity -- I started to notice less propaganda in the second half of 2020. Murders and car crashes started to fade consistently and murders dropped like a rock after Trump's election in November 2024 signaled the end of the second BLM era.
When people talk about whether crime is up or down, what they typically want to do is make an argument over whether the BLM Era of 2014-2023 was a good idea or not. People who emphasize that crime is down in February 2026 vs. June 2020 or 1991 or some other cherrypicked starting date typically want to say: "Don't Blame BLM. It Was a Good Idea. We Should Do It Again In a While."
People who say crime is up are typically arguing that American elites made a huge blunder by falling for BLM.
I, for one, am of the opinion that the 2020 were stupid and counterproductive in a lot of ways, but they probably did not cause the murder increase. That was a general increase in homicide that started months before George Floyd died and was clearly downstream of COVID somehow, and then, yes, an actual "Floyd Effect" in the Summer but that was almost certainly a "viral video" effect, people reacting to the video in ways that made them more likely to commit murder.
The most obvious tell is the shot spotter data, which yes, shows a big spike in murders immediately after Floyd's death - immediately as in a few *days* after. It's absurd to think that police pulled back and criminals noticed and starting murdering people in large numbers in only a few days. Even your example of May 31, only 6 days after Floyd's death, is incredibly implausible as any sort of "protest effect" if you put any serious thought into it.
Which, of course, easily explains why we don't see any such effects in Los Angles homicide rates after the 1992 riots. Internet video sharing as we know it did not exist in 1992. And easily explains the lack of a similar international effect, we wouldn't expect a bunch of, for example, Chavs or Yardies in the UK to particularly care about what happened to some Yank in Minneapolis.
Regardless, I still am in favor of saying that crime is down. Because crime really is down, and I care about the truth.
Many poisons don't actually show up in a toxicology report. I'm not sure I'd say you're getting good data from the CDC on anything except blunt trauma, bullets and knives. The Darwin Awards are often a good example of "what people will do if you point a gun at them" -- "the easy way or the entertaining way?"
Boring as it may be, very few murder victims die from the kinds of murders we like to read about in murder mysteries.
Well, yes. Why create a mystery, when you can simply have the coroner say "died of natural causes" (or stupidity, see Darwin Awards).
Again, these are claimed murders, and if you aren't counting them...
OK ...
I don't really see the point of this post.
People subjectively feel that crime is up, falsely, as you have illustrated (and as they basically always have when surveyed on it, as best I understand it). And they also feel that disorder is up.
You argued (accurately) that crime is way down. Some people refuse to agree. Others shift to "well, disorder is up". Has anybody even established that the latter is true (thus making this a motte and bailey) or is this just retreating to an unfalsifiable claim? Why should I accept that people aren't just as wrong about this?
One of my favorite comments from that discussion was the guy in Honolulu saying one of people's main complaints is crime. Honolulu is the safest city (looking at the homicide rate specifically) in the 69 most populous US cities (i.e., over 300k people). Their homicide rate is less than half the national average. Crime more broadly is also low there. I went there a few years ago and at least my best guess is that "disorder" is also much lower than the major city I was living in at the time. And yet they are (allegedly) seized with the same panic as everywhere else.
[EDIT:] Every one of my follow-up comments invariably cites https://news.gallup.com/poll/1603/crime.aspx
People think crime is pretty bad/getting worse, just not near them, and they're not particularly afraid of being the victim of something.
"Why should I accept that people aren't just as wrong about this?"
I think it is reasonable to just believe that people are wrong about this. I do that.
But I think there is also value in trying to figure out WHY people are wrong about this. They may have something in mind that they have a difficult time articulating. For example, I believe this is true for people who prefer the "warmth" of analog records vs digital music such as streaming and CDs. No, the vinyl record isn't qualitatively better in any measurable (for most people) way. But there is something to the preference.
If we care about what might be bugging people who think crime is up then it makes sense try to figure out what it is that is bothering them. We may then decide we don't care ("I'm upset because the *other* team is winning politically ...") but maybe we will care at least a little bit.
I think that using an easily measured and understood phenomenon as a proxy for something else that seems psychologically related is pretty widespread. People still think that inflation is out of control. It isn't crime or disorder, it's people feeling defensive and territorial because the future seems uncertain and other people seem to be getting aggressive about competing for scare resources. But that's complicated and messy to think about. Instead, most people are just saying to themselves (not necessarily fully consciously) "I'm feeling socially anxious, let's look around and find a reason why. Oh, crime, of course." I think we all do this. I could cite research on it if anyone wanted me to.
Please cite research! I'd love to have data on this.
There's something called the "Two Factor Theory of Emotion" which states that emotional reactions start as a physiological response (the heartrate increases, bloodpressure rises, muscles tighten, etc.) and then the mind starts looking around for a reason in the environment. We have a strong tendency to attribute our reactions to the first thing that comes into view. Someone once tested this theory by placing an attractive female survey taker in the middle of two different bridges: one rickety suspension bridge, and another that was solid cement. She stopped male hikers, and them answer some fake questions, and they gave them her phone number, telling them to call if they had any questions about the survey (of course, they actually called the researchers). The hikers were far more likely to call her if they had been on the rickety bridge--the only way to explain that is if they misattributed their physiological reaction to her instead of to the bridge.
See:
Schachter, S., & Singer, J. (1962). Cognitive, social, and physiological determinants of emotional state. Psychological Review, 69(5), 379–399. https://doi.org/10.1037/h0046234
Dutton, D. G., & Aron, A. P. (1974). Some evidence for heightened sexual attraction under conditions of high anxiety. Journal of Personality and Social Psychology, 30(4), 510–517. https://doi.org/10.1037/h0037031.
I don't think it's that deep. Bad-faith political actors on the internet have an incentive to spread around false narratives that serve their agenda. For some of them, one such narrative is that crime is increasing. Because most people form their beliefs not by rationally thinking about the evidence but by social diffusion from the people they trust, those who share a political tribe with the aforementioned bad-faith actors become convinced by them. So people don't believe the narrative because they're searching for some easy-to-understand proxy for their general sense of malaise. They believe it because their favorite political influencers told them so.
I think the Internet is a factor currently, but also people have been erroneously thinking crime is up for long enough that it can't be the main reason.
https://news.gallup.com/poll/1603/crime.aspx
And it's usually the same except-in-my-neighborhood thing that makes me think it's not the "disorder" story. If people thought their lived experience of graffiti and crazy people = crime then they wouldn't think "crime is really bad, but not around here".
Note that the "but not in my neighborhood" has applied to how good the public schools are for a long time. It is common for folks to rank public schools in general lower than the public school their children are attending.
[Though this may have changed during Covid.]
Yeah, I think this trend is really common. Also fits the "Vibecession" data pattern where people think the economy has gone to sh** even as they acknowledge their personal situation has improved.
[EDIT:] E.g., Figure 7 at https://www.federalreserve.gov/publications/2024-economic-well-being-of-us-households-in-2023-overall-financial-well-being.htm
Agreed, the general problem is that politics is the mind-killer and how exactly that manifests differs from time to time. In our time, the internet is the primary vector.
Well politics + online is certainly the mind-killer we are most likely to run into here.
I haven't been able to find the poll, but I believe this phenomenon is true of economics as well: There's a big gap between what people say their financial situation is and what the think the overall economic health is, and the gap has grown over time.
Figure 7 at https://www.federalreserve.gov/publications/2024-economic-well-being-of-us-households-in-2023-overall-financial-well-being.htm
Consider that two things can be true at the same time:
1) I don't see any crime or disorder around where I live.
2) I see a lot more crime and disorder when I go into The City to work/shop/whatever.
Even if I assume people would answer the questions that way in that situation, I don't think that fits the questions where people also report rarely or never being concerned about being the victim of any particular type of crime.
A poll by Pew Research or whatever would compress people's conflicting sentiments about their varied experiences (e.g. between work and home, or home and travel) down into one thing, so it's interesting to consider which thing they would choose to express to a pollster and why they would choose that thing rather than the other thing.
But here, for example, where we can write basically as much as we want, people still tend to minimize one and focus on the other, and I think it's for exactly the reason(s) noted in the comments immediately above.
For the crime thing, though, it can also be that stuff is locked up in stores that wasn't locked up ten years ago (personal observation. I don't have data and thus this is an anecdote ... maybe I'm wrong about the stuff being locked up.).
"Crime is down a lot. That is why we have to lock up merchandise so that it won't be stolen," sounds like gaslighting to a lot of people. Who are you going to believe? Me (with my crime stats) or your lyin' eyes?
There are many explanations for why stores appear to be locking up certain items and not all of them involve increasing crime. The explanation that one buys into depends heavily on what their trusted pundits are telling them.
Moreover, the choice to focus one's attention on a particular piece of evidence, such as this one about stores, depends on the narratives that they believe or want to believe. For any hypothesis, there are always pieces of evidence to point to and say, "See!" and a clever pundit can pick and choose the evidence to make virtually any case seem convincing.
Because of this, I don't think it's so easy to disentangle "I feel that crime is up because I notice that items are locked up in stores" from "I notice that items are locked up in stores because I feel that crime is up."
>why stores appear to be locking up certain items and not all of them involve increasing crime
What alternatives are there for stores to increase their costs of doing business, other than it being better than the alternative of crime?
Does the manager enjoy annoying customers or forcing them to socialize with the staff to get stuff unlocked?
You say "stores appear to be locking up certain items". The word "appear" is a tell that you are using skepticism to avid the obvious conclusion. They are locking up certain items. I can see it. So can everyone else. I asked the store manager, and she told me it was a corporate policy. Do you think I hallucinated, and just so happened to have the exact same hallucination as everyone else, including the people working at the store?
> There are many explanations for why stores appear to be locking up certain items and not all of them involve increasing crime.
There are many possible explanations for everything. That doesn't mean they are all equally likely, or that we can't get any knowledge of what is happening in the world. Are you this skeptical of all possible conclusions? Or, do you selectively choose to apply total skepticism of all knowledge when you don't like the conclusion you would otherwise reach?
When I go to the drug store, deodorant is locked in a plastic case. I need to ask an employee to get a key to open a box and watch me as I take an item from the shelf.
Am I a bad faith political actor for noticing this? Has CVS been tricked by bad faith political actors into thinking this was necessary?
See my response to Mark Roulo further down this thread.
John M, you put it perfectly.
Economically, there isn't any reason to prefer the change in price levels over the last week, the last year, or the last decade as the measure of inflation. Well, ideally we'd use the instantaneous rate of change but that's not measurable. The government, and hence the media, reports inflation in yearly rates by long convention but there's nothing fundamentally special about years here.
People have a hard time adjusting when a unit of measure changes, just they would if they were trying to figure out if they could lift something when the definition of a "pound" had recently changed.
Which isn't to say that restoring the price level to 2020 like people want wouldn't be a much worse cure than the disease. But I don't feel like the complaint is actually unreasonable.
>Well, ideally we'd use the instantaneous rate of change
This is trivially false. If we imagine the price of everything suddenly doubled in one day and inflation is 0 for the next four years, the negative impact of inflation would harm people throughout those entire four years.
The actual impact of inflation on an individual is the amount of purchasing power they lose as a result of it. However much money you have in the bank, you effectively lose the inflation percentage of that (100k in the bank *4% yearly inflation costs you $4k per year). However much your income is, you lose that much income (if not adjusted for in your raise). If you would have gotten a 2% raise, but your boss gives you a 3% raise instead to compensate for inflation, but inflation is actually 4%, you've effectively got a 1% pay cut, and 3% pay cut compared to the counterfactual case where you could have had a 2% raise with no inflation.
For each individual, if you had perfect information of what their income would have been in the counterfactual case there was no inflation, you could compute the total harm to them from inflation by integrating it across their lifespan. They were be permanently that much poorer. You can't just say "well, you lost $3k this year, what are you complaining about? Nevermind that you lost $50k last year, that was last year." The harm remains, but isn't realized until their savings (which have the same numerical value) don't go as far as they buy things.
This sounds similar to what I said on the vibecession post: https://www.astralcodexten.com/p/vibecession-much-more-than-you-wanted/comment/184600156
People can easily tell they feel uncertain/anxious or like something is wrong. It's not always easy to figure out why. But inflation and crime get headlines, which means it's likely that people can grab on to them and then try to justify it as an explanation afterward. Human memory is unreliable and can even be rewritten to justify current beliefs.
I often think about inflation "being out of control", and the statistics that get used to disprove it
i imagine a scenario where the government prints a quintillion dollars in a single day, and for that year inflation is like +100,000% or whatever... but then maybe next year they print nothing, and there's no inflation
if i lived in that society, it would probably take me the rest of my life before I stopped claiming "inflation is out of control", and i would feel justified in making that claim even if there'd been 0% inflation for decades prior
I think I would not feel unjustified in saying that, unless prices actually decreased back down to the level they would have been at if inflation had been more like 5% annual
Technically, your concern isn't about inflation, per se, it's more about price stability. To the average consumer, that may be a distinction without much difference, but to an economist...
I don't mean to dogpile, but the inflation discourse frustrates me. It feels like it often goes like this:
"Inflation is out of control!"
"No it's not. High inflation happened years ago and it's back to reasonable levels."
"But prices increased massively and my wages didn't keep up! Why would I care about the current rate? The problem is affordability!"
"Yeah but the affordability problem happened years ago and hasn't had any sharp increases since then."
"So? My wages *still* haven't caught up to the price increases!"
"Oh, so you want deflation? That would actually be bad for the economy because blah blah blah."
The person pooh-poohing the problem isn't saying anything incorrect there (other than attributing a desire for deflation to the other speaker) but they're completely dismissing the correct things the other person is saying, just to feel smug. The affordability crisis isn't over. Perhaps there is nothing else any government can do about it now, but dismissing the problem out of hand because inflation is back to reasonable levels is crazy-making to the people still feeling the squeeze.
It isn't just to feel smug, though. Your two actors are using different definitions of "inflation" (or "affordability", or whatever). The second is referring to a population phenomenon: economic change is something that happens to populations of people, not single individuals. Meanwhile, the first person is using something like "Inflation is what happens to me!"
There is a sense in which inflation can happen to a subset of people within a larger population, but identifying exactly who those people are has proven surprisingly hard (even harder is convincing them that they are a minority).
> But I think there is also value in trying to figure out WHY people are wrong about this.
Sure, but I think this is well explained by a combination of sensationalism and polarization in traditional media, and general amplification of emotional content on social media.
As has been discussed here before, people mostly don't believe things based on careful study of the truth. They reason socially. If they hear about crime all the time, they think it's up. The news loves to broadcast about crime. Politicians love to attack their opponents over crime. You hear about dramatic stuff like murders even though murder is one of the rarest crimes.
You still have to account for changes over time. Fear of crime isn't a constant, it rises and declines. Those changes can't fully be accounted for by constants in human nature.
I'm not really sure what you're trying to say here. My explanation doesn't suggest people will always think the exact same thing, so much as that their thoughts are mostly driven by trends in media reports.
I guess I'm assuming that the media reporting on crime is the same as it always was. The news has always had a disproportionate interest in crime.
A lot of more-partisan sources will be inconsistent over time in how they highlight or downplay crime, based on who is in power or what policies are being discussed, though.
Isn't it obvious why people think crime is up? There are fewer white people now and more black and brown people.
It's a predictable consequence of a big demographic shift: different racial groups don't trust each other as much and don't feel as safe around one another. And there are existing stereotypes anout black and brown skinned folks being more associated with crime that even non-whites often absorb (eg. through television and media)
I'm skeptical of this explanation, based on the data I keep citing too far down the thread to be helpful:
https://news.gallup.com/poll/1603/crime.aspx
People are mostly reporting that crime is going up in the US but not near them and that they feel pretty safe. If they thought this based on how scary minorities are, presumably they would be more scared for themselves.
That's an interesting point, although I do notice the defintion they give of "near you" is "less than a mile from where you live". Imagine how parents who live in the suburbs of a big city might answer this question. They have the impression that the cities have gotten more dangerous and disorderly, and maybe that's even the reason they moved to the suburbs (perhaps even looking for the whitest neighborhood unconsciously) but they don't see this as affecting the area within a mile of where they live.
This survey might also be an indication of the role that politics and journalism have played in amplifying this effect. Perhaps originally it was a marginal effect, whether real or imagined (that crime was going up). But once that happened, politicians capitalized on that and began using the "crime is up" narrative in their platforms. It became part of the party line, and that provided a financial incentive for Rupert Murdoch and other right-leaning media owners to encourage more coverage of crime anecdotes. Covering that makes more money at that point, because it's telling people what they want to hear, validating their fears and re-enforcing their beliefs. Soon enough, people are watching (and resharing) a rising number of crime stories across the nation, but most of them aren't happening near where they live.
I commented in the Disorder thread on what I believe it is that is bothering people: That crime hasn't gone down sufficiently as a result of widespread anti-crime technologies, and in particular that police ignore crime even when there is ample evidence to convict (for example, you have video evidence of shoplifting, but the government declines to do anything about it). Historically, police not doing anything about shoplifting might be acceptable because they didn't have enough evidence, and there was a societal notion that criminals would eventually slip up and get caught; however, we live in an era in which people perceive criminals constantly slipping up, and those responsible for catching them simply decline to do so.
Yes, I accept that London is much safer than it used to be, but I am absolutely furious when the police refuse to investigate crimes where the perpetrator has been literally recorded doing it.
That's a very good observation. There are cameras on every pole and in every pocket, there are trackers in every other device, so there's a reasonable expectation that this will produce swift and inevitable punishment.
However, I think this ubiquitous surveillance is also preventing the efficient elimination of disorder. Some decades ago, a cop or a deputy could simply trash a hobo's tent and dump the owner on the road halfway to the next town, Will Teasle style. The same cop or deputy could also detain a small-time dope peddler or a petty vandal and give them a wallop without booking them. Nobody would really see how the sausage is made, but everybody would appreciate the more orderly streets.
But in 2026 the LEO is wearing a body cam, the perp is filming them, his friends are filming them, random bystanders are filming them. The existing sentencing process is not calibrated to handle this torrent of evidence. It implicitly assumes that the cops have some leeway in how they enforce the law and if any evidence of them abusing this power comes up during the trial, it's because it was egregious enough.
The perception of crime is one of those things that confounds public opinion research. In 23 of 27 Gallup surveys since 1993, at least 60% of U.S. adults said there was more crime nationally than the year before.
It's worth noting that public perceptions on most other social tends are *not* totally unmoored from reality. Generally speaking, peoples' perceptions of which trends are getting better or worse are correlated with real numbers. Given that background, it's somewhat of a mystery why peoples' perceptions of crime trends are so distorted.
Tangential: the most salient explanation I've heard for the preference of records is that operating a record player has significantly more friction. You don't just have the music available at the press of a button, so if you're not familiar with the concept of active listening, putting on a record is more likely to put you in the right space to do just that, as the instantaneous availability of digital music (and being served on platforms full of distractions) makes it easier to drop and do something else during.
Funnily I just very recently read an article on this blog that touches on this https://www.astralcodexten.com/p/conspiracies-of-cognition-conspiracies
> But I think there is also value in trying to figure out WHY people are wrong about this.
I think that's for neuroscientists to figure out, not some random web blog. Either way, it doesn't make sense for Scott to get mad about this. How is he still getting surprised about how people function?
Two major confounders:
"Overall crime rate" is not what most people care about. "Crime rate that might affect me" is the common concern. Murder can be down dramatically overall with more than all the decrease being among active participants in the drug trade - everyone who wasn't involved in the trade's "chance I might be murdered" is up even though murder rate is down.
"Crime avoided by costly preventive measures" goes into most people's perception of crime, but not into crime statistics. Even if the rate of theft is the same, a town where no one locks their doors and you can just walk into the stores feels a lot safer than one with bars on every window and armed guards in every store.
> "Overall crime rate" is not what most people care about. "Crime rate that might affect me" is the common concern.
True, although as I noted elsewhere people pretty consistently report that crime is not that big of a problem in THEIR neighborhood while thinking it's bad in the US overall.
> Even if the rate of theft is the same, a town where no one locks their doors and you can just walk into the stores feels a lot safer than one with bars on every window and armed guards in every store.
Yeah, agreed here. Which means you can make people think crime is up when maybe it's that you're not willing to pay for more retail employees so you have to do these things instead.
Maybe a pithy way to put it - instead I’d doing The expensive thing they used to do to mitigate shoplifting (hiring more employees) since crime is down they’re doing the cheap thing instead (locking shelves) but that makes customers feel like they’re doing the expensive thing, so they feel crime is up.
I doubt that's the motivating factor, though - companies seem almost (or maybe literally) pathologically committed to reducing labor spending regardless of whether it makes any sense.
""Overall crime rate" is not what most people care about. "Crime rate that might affect me" is the common concern. Murder can be down dramatically overall with more than all the decrease being among active participants in the drug trade - everyone who wasn't involved in the trade's "chance I might be murdered" is up even though murder rate is down."
I lived in Washington when in it was the murder capital of the US. The well off "civilians" were still very worried about crime even though the ratio was something absurd like 6 out of 400+ murders occurred west of 8th street NW.
I understand it's mostly the same sort of localization in Chicago, which also has a bad reputation for being out of control now.
I noticed when I used to be on next-door that people posted videos all the time that were for the most part just making everybody unhappy. I don’t think anybody was ever apprehended as a result of any of these grainy videos. But each morning here they’d come: I camera caught the car that was going by and emptying the mailboxes last night, here’s my camera showing a guy walking up into my garage and taking the beer out of my fridge, here’s a snapshot of the homeless guy who was banging on my door at 3 AM, here’s a picture of people who rifled through cars on our street last night.
Gadget lovers were over the moon about this.
But nobody ever got any interest from the police. At least not individually. Once in a while, you know a male theft gang might be arrested, so maybe it’s the case that they were collecting geographic data and then when it reached a tipping point, acting on it.
Stolen cars were often found, often in a sort of Bermuda triangle for gang crime but there wasn’t generally anyone to prosecute.
In one memorable case, where a homeless guy was yelling every night and just disturbing the peace in one location and frightening people by being at their doors and windows, the person who managed to get the police out, was for some reason actually cited for harassing the homeless person.
The post uses this specific example (responses 6o a post about whether crime is up) to make a more general point about online discourse.
A lot of online writing (including Scott's crime post) is debunking what other people think. The issuenis that we lump together similar but different claims together under the same general umbrella. This means that a debunker, like Scott, has to focus on specific claims to falsify. This leads to people with similar, but different claims, to complain that he only debunked some noncentral claim. But if Scott doesn't address the false claim, he is not adressing something that others feel like is the central claim.
Does this blog need to be "debunking" people? Isn't it simply enough to find a discussion, do your own research on the subject, and then publish it? People are free to disbelieve the findings, that's their problem.
I think the point is that one article cannot cover everything. Like, when you write an article about X, someone will probably say "but why didn't you mention Y?" and probably suggest that it's because you have an agenda. But of course, if you mentioned Y, the same objection could be made about Z.
And this becomes more tricky when X and Y *are* actually related... but so is Y and Z... and also Z and Q... because things are connected, in general. So wherever you draw a line, the very fact that you drew a line can be used to attack you.
Here, X, Y, Z, Q could refer to things like "national statistics of violent crime", "national statistics of white-collar crime", "national statistics of general nuisance", "statistics of San Francisco specifically", "things that people do to protect themselves from potential crime they expect" etc. Each of them a perfectly valid topic. But it is annoying when you choose one of them, do the research, write the article... and the comments are mostly about why didn't you talk about some other topic. (Especially if you already preparing a separate article on that other topic.)
> Like, when you write an article about X, someone will probably say "but why didn't you mention Y?"
"Because this is an article about X. Fuck off." See? It's that easy.
Sort of. I think Scott has one of the best reputations for being earnest and sincere and unbiased in his search for truth, so there is certainly no need for him to "both sides" every issue so that every aspect is evenly and fairly covered. But lying by omission is a real thing. It's one of the primary reasons people have grown to distrust a lot of media. You can publish a lot of technically true articles that give a heavily distorted view of a topic.
When I read about nutrition - one of the few things I know about more than 99.9% of people - everything I read is wrong.
When I read about … I dunno something I know more than a vast majority of people about but not at the 99.9% level, and it’s all still seemingly very completely wrong all the time.
So then when everyone tells me crime is down - but all the Walmarts and cvs has everything on lock … and my own experience mentioned last week where my own big box retailer reports basically 0% of shoplifting … what am I to believe? Them, or my own lying eyes.
Crime is probably down in big ass cities since the 80’s or something.
Murder is down.
But crime? Crime? Like vagrants and assholes doing slightly illegal things and people stealing petty amounts of shit? It’s not down.
For some reason it’s called disorder now just like homeless are called unhoused.
> So then when everyone tells me crime is down - but all the Walmarts and cvs has everything on lock … and my own experience mentioned last week where my own big box retailer reports basically 0% of shoplifting … what am I to believe? Them, or my own lying eyes.
Said this elsewhere, but this could also be a function of retailers employing fewer people that would otherwise discourage shoplifting.
> But crime? Crime? Like vagrants and assholes doing slightly illegal things and people stealing petty amounts of shit? It’s not down.
How can you be confident about this? Your views on nutricion are presumably not from your own personal experience of watching people eat.
Staffing is horrifically down amongst all retailers compared to even ten years ago. But that seems like a ‘ she was asking for it ‘ argument.
I’m confident because I see it! We literally see it in our stores and neighborhoods.
Life wasn’t an unlivable shithole 30 years ago. And it’s not now. But it’s not drastically better.
We see crime not being taken care of, we see it not being reported … I understand it’s anecdotal, but like Scott touched on (somewhere) everyone is lying about everything all the time.
> But that seems like a ‘ she was asking for it ‘ argument.
Eh, I'm not saying anybody deserves to get robbed. I'm just saying this could explain all of the locked shelves, rather than some independent sweeping crime trend.
> I’m confident because I see it! We literally see it in our stores and neighborhoods.
https://news.gallup.com/poll/1603/crime.aspx
People mostly DON'T report seeing crime in their neighborhood. They report that crime in general kinda bad but their personal experience of it isn't.
Shoplifting can be down and shoplifting in Walmart and CVS could be up. Shoplifting inequality could easily be decreasing which would increase exposure of the average shop to shoplifting.
Huh where are you seeing everyone saying crime is down? I’m seeing everyone talking about crime being up.
Isn't the simple explanation "crime is down *because* shops lock down a lot of things, so they are now much harder to steal"?
The illegal things you are gesturing towards haven’t shifted to being called other things such as disorder, you can just look up shoplifting statistics, or assault statistics, or burglary statistics, or lighter types of crime statistics, they all have specific legal definitions.
A lot of people just refused to put in effort to engage and it looks like Scott felt frustrated.
I'm one of those people who think "crime is up, cops are hiding it" but since Scott addressed this *in* *his* *title* I would have needed to put in a lot more effort than I cared to in order to treat his argument fairly. Instead of just saying "no, I don't care what you prove" I said nothing, and feel happy with that decision.
> People subjectively feel that crime is up, falsely, as you have illustrated (and as they basically always have when surveyed on it, as best I understand it). And they also feel that disorder is up.
Agree this is what it comes down to. People always feel like crime is up. These posts from Scott have amounted to a back-and-forth between people claiming crime, or whatever crime-adjacent thing, is *really*, by some definition, up, and then someone pulling some stats to test it and finding that it's not up, leading people to come up with a *new* definition for the thing they care about and why *that* is up.
Usually ends with something sufficiently not quantifiable that it can't be proved wrong.
I'll just say - whatever time period people think are the good ol' days when crime was low - look at what people said at that time. They probably thought crime was higher at that time than before, back in *their* good ol' days, when crime was *really* low!
The peak age of low crime when I was old enough that I remember it, but young enough that I didn't have the attention span to watch the news.
"But the inverse evil trick is saying something 'directionally correct', ie slightly stronger than the truth can support. If your enemy committed assault, say he committed murder. "
I'll add one variation (?) on this, which is the logical fallacy where a term gets used to mean two separate things.
One example that I see a lot are taking "casualties" counts from the Ukraine war and treating them as "fatalities" where casualties also includes wounded.
Another is when discussing education overhead and turning "administration and other non-teaching staff", which includes IT staff, librarians, janitors, etc. into simply "administration".
I think a lot of how to deal with this depends on whether we think the folks engaging are searching for truth/correctness or just want to win.
This sounds sorta like the non-central fallacy that Scott coined years ago.
Similar but different
I think that's the Equivocation Fallacy:
https://en.wikipedia.org/wiki/Equivocation
Thank you!
Yes, that's it.
There's atrocity sorta fallacy too like after Oct 7, lots of claims of what Hamas did was going around, and there was the baking, and beheading babies one, Noah smith actually blocked people for saying we don't have evidence of baked babies yet as them "denying atrocity", another dude blocked me when I said after two years it seems Hamas did not indeed target children and the kids killed seem to have been fron stray reckless killings. To preempt any "Hamas apologist" thing, i stated clearly at the start that Hamas is evil, oct 7 was bad and all, didn't stop anything. Person literally stated I was lying about atrocities while admitting i was right that there was no 40 babies killed, it was a discussion about how it's fine to sometimes ask for evidence of some of the worst crimes about Oct 7, and I mentioned how the 40 babies ended up being a lie and he was lie "liar we know Hamas intentionally killed children even if it's not 40", this was crazy on three dimensions, he shifted from babies to kids, admitted the 40 numbers is incorrect and his actual evidence didn't show intent too! Same with mass systemic rape, tjis is actually something i believed for like a year after Oct 7, only to find that evidence basically shows maybe very few like 2 or 3 rapes happen and we cant actually prove those either and nothing to indicate mass or systemic rape.
Oct 7 is a great example here because you can't question anything, they could say Hanas dropped poisonous manner that melted people skin off and questioning it is basically considered implicit support for the atrocity. Trying to say the baked and beheaded babies thing is without evidence, and it's why point this out, you are hamas supporting because why would you try to draw the distinction at all. It's like why would you care about the distinction being drawn at all too if it's indeed still bad.
The Noam Chomsky Khmer rouge denial too, is basically "i think this number is unlikely to be true cause they are extreme" during the height of the event, and after a while, he admitted the numbers then were correct but it's still "Genocide denial" for him.
Noah Smith blocked me years ago for referencing Matthew Yglesias correcting him on Tsarist economic growth. Every once in a while I want to "like" one of his tweets and remember I can't.
Yeah this is interesting. I'm sympathetic to the point you're trying to make and even I get this feeling while reading your post of: "why does this matter so much to you?" I sense this weird pull at the back of my brain that I think these other people calling you a Hamas supporter must also have. Arguing about the facts here feels in bad taste somehow, but I agree with what you're saying that this doesn't make any sense. It's worthwhile to be accurate about what happened on October 7.
If I'd have to guess why this happens, I think it's because it's very hard for our minds to see the difference between you disputing some of the facts, and you disputing all of the facts. Or perhaps in arguing against some of the facts, it sounds like you're making a moral argument pro hamas. It seems like one of these system 1/system 2 heuristic type of bias. I have to force myself into system 2 cause otherwise I get a visceral reaction against your text.
Yeah, most of this are not actually me but observations on the discourse. Like stuff I see, the Noah thing was not one i directly participated in, but I was following him and had his notifs on then so i had front row to the entire thing.
The main one that was mine was actually a meta discussion specifically about conduct on the Oct 7 discourse, and to make myself clear i stated to start with Hamas was evil, terrorist, oct 7 was a war crime and massacre and unjustified, etc. didn't stop much. That discourse basically was like i said i think Oct 7 can be so sensitive people that are shocked by some of it might not be so quick to believe all the claims and that some of this reluctance seem to have been good in hindsight and I listed ones like the baked babies and Hamas beheaded 40 babies, which he agreed tgat indeed Hamas did not do that to 40 babies but that I was evil for denying Hamas targeted kids even if it's not beheaded 40 kids, it's the directionally correct thing Scott mentioned. And he attached a picture to a source which I hadn't even seen before and he would have known if he didn't just reply to my one comment and block. But also the supposed evidence didn't even prove it but like it would be pointless to try to argue indeed I wasn't arguing that Hamas didn't kill kids on Oct 7, just that they didn't seem to have deliberately targeted them as claimed.
This was actually over a year after Oct 7 cause for the most part, I didnt enter the discourse about that event for a year cause we didn't have evidence, that is true of why I thought a lot of the claims are being made without evidence but it's also why I didn't comment because I didn't have evidence to the contrary and people didn't seem in the mood then.
Thing here is the stuff is a big event, it's actually important we know the details of such a pivotal event. That was important, it's true it's a huge crime against humanity but there's a reasons claims od baked babies took off, it adds different stuff to the discourse. Like did Hamas truly did it just to kill jews or because they thought they could destroy israel or did they indeed do mass rapes and go around baking and beheading babies. This is important, it's the reason we study Holocaust and stuff.
Unfortunately, relying simply on the facts, while the best long term strategy, often makes things worse in the short term, as various factions realize that their agenda may be weakened. Since humans compete on an in-group basis by instinct, there really isn't any workaround, except persistence.
Eh, 'Lies, Damned Lies, and Statistics' is a saying for a reason.
Sticking to 'facts' is not universally safe, it's easy to mislead people and manipulate conversations while only saying facts.
The alternative to good statistics is not no statistics, but bad statistics, which is what people complaining about statistics tend to resort to.
"It is easy to lie with statistics; it is easier to lie without them."
Sure.
All I'm saying is that facts are a necessary but not sufficient prerequisite for truth.
But is it as convincing? Rhetoric can still be assisted by cherry-picked facts to lend it some authenticity.
Things that take more effort to fake and are easier to disprove if false SHOULD be more convincing, because they are costly signals. That's a feature.
Though I wouldn't dare to claim that humans have an accurately-calibrated sense of the appropriate degree of extra convincingness.
But nothing they say is untrue. The facts are real. They just aren't relevant.
You don't get to know WHICH dastardly trick is currently being used against you before you choose which heuristics you're going to deploy to detect dastardly tricks.
Arguments with supporting data are (and should be) somewhat more credible than arguments without supporting data. This heuristic will occasionally cause you to give more credence to an argument that turns out to be wrong, but it's still better overall than not having that heuristic.
Also, cherry-picking still usually involves some risk of getting caught, though it is typically a smaller risk than with outright lying.
"Figures don't lie, but liars will figure"
You're doing a fine job. Yes, Type I and Type II errors both exist, but that doesn't preclude the existence of truth that we're all striving after.
I suppose ideally the crime-is-up people would clearly state how their claim can be measured and then we measure it. You can try to be clever in this setup and name a metric that's prohibitively expensive or borderline impossible to actually measure in the real world, but one can hope that it would be obvious that that's what you're doing.
How dare you use your platform to argue with people who are not me. For some reason, there are people on the Internet who are responding to people other than me rather than making content that only responds directly to my beliefs and interests. This is intellectually dishonest.
There is another difficulty, which is that to most observers, a good-faith attempt at precision looks identical to a malicious streetlight. The surface form is the same: “That’s not quite right, here are the numbers.” Distinguishing motive requires context most readers don’t have. This is what makes this protocol both effective and routine in politics.
More generally, opinion often comes first, with the supporting logic second. That "streetlight" is just as often shone by someone selectively illuminating the part of reality that happens to fit their intuition. Thus, the intent is not always malicious. We just live in a world of invisible gorillas.
And unfortunately the number of malicious dark arts users who actually are trying to manipulate people with these sorts of arguments outnumbers the number of good faith actors like Scott. Being suspicious of such arguments is good epistemic practice (though concluding that they're obviously bad without confirmation either way is bad practice).
It's good and useful to identify and call out rhetorical sleights of hand that are used to obfuscate the truth. Identifying them and giving them names is good for this.
But you don't want to get tangled up in the labels. They should not get in the way of the core mission, which is to find the facts most likely to be true using the best available evidence, and debunk 'facts' that have been contradicted by the evidence.
The oldest sophist trick (at least the oldest in my subjective experience; I think I noticed this being done to me when I was still had a single digit age), is to jump to an adjacent topic the moment you realize you can't support your position on the current topic. Then, if you get into trouble on the adjacent topic, you can jump back to the original one, and most people will forget the trouble you had with that one. To anyone who's not paying close attention, it will look like you've come up with new points and your counterparty will have to labor mightily to re-explain.
It appears the labels, useful in many contexts, can be used to assist this oldest trick. I think the solution is to just call it out explicitly when it happens. Force a concession before you allow a change in subject, or a change in rhetorical framing.
There's some tweet I forget exactly, but it goes something like: "to an average person, if their package is stolen, or if they stand around all day looking at their front porch / install a ring camera so they can grab the package the instant it arrives, those are all the same, even though only one would appear on crime statistics."
Basically, to a normal person, if they have to take extra steps to keep themselves safe from crime, that's the equivalent to crime being "up" even if it doesn't get statistically reported.
I don't think I understand how this relates to Scott's point, can you explain?
There's a difference between crime statistics and "can I trust the world around me?"
Where I grew up, we didn’t lock our car doors. Later, I moved to a “bad neighborhood” and still didn’t lock my car door. Somebody stole stuff from my car there. I didn’t report it. But after that, I always locked my car.
My impression is that crime in that neighborhood is worse than where I grew up.
Scott could not conclude that, based on his methods. There’s no police report; and me locking my car door shows up in no data. He has to conclude that I’ve been swept up in a moral panic, because the circumstances I’m actually responding to are invisible to him.
But he would have noticed if someone stole the car. It's awfully suspicious that only forms of crime that don't get reported go up.
For about a year, there’s been a tent set up in the woods in a park near me. Camping isn’t allowed there, it’s for people to walk their dogs and jog, and children to play soccer and baseball in the park’s fields. It closes at sunset; nobody is supposed to live there.
*I* know about this particular instance of disorder, but I haven’t called the police about it. So there’s no police report. And the victimization surveys don’t ask about homeless people you’ve seen. So he won’t show up on, say, the NCVS.
That tent is completely invisible to you and Scott’s hard-nosed empiricism.
Is it “awfully suspicious” that neither I nor any of my neighbors have reported this guy? We’re just minding our own business, like normal people. You think this tent was hallucinated by my moral-panic-fevered brain?
I guess the only way for my fellow citizens of a scientific frame of mind to believe me about the disorder I see around me is to lodge a formal complaint with my local police. Then it’ll be reflected in the statistics. I don’t like that, but I guess that’s where we are.
If it's just one type of disorder, then yes, that's not suspicious. But the problem is if every measurable type of crime and disorder seems to be going down, and all the immeasurable ones are going up. What about them being immeasurable makes it so they're the ones that happen more?
Do you have a good argument that the fractions of break-ins that are reported has declined over time? Or are you just pointing out that statistics can never prove anything with absolute certainty because they can always miss something?
The article in question addressed that by looking at homicides and car thefts, because they are almost always reported.
What I actually think the situation is:
- Crime (as in somebody is actually victimized and experiences harm) is probably about the same
- Disorder (graffiti, drug selling, homelessness, loud noises, hooliganism) is up but hard to measure and distinguish from cultural differences and bias
- Lawlessness (people who don't care about the law and will break it if it benefits them and they can get away with it) is up a lot. (A very small number of lawless people have a huge effect)
- Defensive measures taken by normal people to protect themselves from crime and disorder are up
- The US taxpayer spends a lot more money to address disorder than in the "good old days". Because of this, taxpayers would feel betrayed if disorder stayed at the same level -- because they're paying a lot for it to decrease
- A small increase in crime visibility due to social media.
- Conservative political operatives online think they can score points by making crime seem worse than it is, because liberals "own" disorder and lawlessness. This part of the "moral panic" narrative is totally right.
So the increase in complaining isn't necessarily due to actual crime victimization. But that doesn't mean it's not legitimate. It's partly illegitimate, partly justified.
Crime is probably about the same as when? 2000? 1960?
Thanks for clarifying your position.
Keep in mind I haven't read all the comments on this article and I'm only asking about the point made above about crimes not being reported. Scott attempted to address that concern in the article and made a convincing argument that it's not reports that are declining, but crime itself. Do you have a good reason to think a smaller fraction of crimes are reported now? Ideally, an argument stronger than you know what conclusion you want to prove and you're looking for any possible gap in Scott's argument. The problem with hunting for gaps is you're only showing it's possible, but not showing it's likely. Given that Scott showed murder and car theft are almost always reported and are down, and that there is no obvious reason to think there would be a smaller fraction of crimes reported today than 25 years ago (before cell phones and video recordings were ubiquitous), it seems unlikely that the fraction of crimes reported are down.
If the fraction isn't down, your argument about not reporting your car break-in doesn't affect Scott's arguments. If the same fraction of thefts were reported 25 years ago, then thefts really are down.
It seems to me that the major point of the parent post is that people would interpret "successful crime" as somewhat equal to "prevented crime" - events that can't show up even in perfect statistics because they didn't happen; but they didn't happen only because people took undesirable preventative actions; in both cases the "threat of crime" was real.
If a store locks up its goods and they don't get stolen, that's technically a reduction in shoplifting, but it's in no way an indication that people feel that the threat of shoplifting has decreased, quite likely the opposite - the fact that they didn't lock them up earlier but do now is some evidence that the threat might have increased.
If threat of stealing stuff from cars increases in some city, people will start habitually never ever leaving any bags in cars because their windows would get smashed otherwise (I've seen these habits change over time in both directions). Now once that habit gets ingrained, the number of thefts from cars would decrease, and the number of smashed windows would decrease - but that doesn't mean that the threat has decreased, it's just that there are fewer opportunities now. What people actually care about is the chance of coming back to a smashed window *if* they left their stuff inside, but what the crime stats show is this valuable metric multiplied by how often people leave stealable things in cars. What people want is to feel safe enough that they could intentionally just leave things in their car as the risk is insignificant, but that would increase total crime stats by providing more opportunities.
If someone decides that they don't feel safe anymore in the park and stop their evening jogging, this can only reduce the crime statistics (because they might have gotten assaulted in that park before, but now it's not going to happen), but it doesn't mean that the threat of crime went away - actually, the person *did* suffer a real harm from crime, namely, that their freedom to jog was meaningfully restricted by the threat of crime.
So crime statistics count "crimes which happened despite countermeasures", but people actually care about the general *threat* of potential crime and if we want to compare different places and times with meaningfully different level of countermeasures, we need to adjust for that; if we want an apples-to-apples comparison, we need to estimate a "what if" of what the risk of crime would be for people who take the exact same (strong or weak) level of precaution.
I think you're right that their point was the threat of crime and I misunderstood that.
There's a problem with the threat of crime theory as well that needs an explanation. People taking precautions against crime is a negative feedback loop. This should act as a multiplier. If criminals collectively are trying to commit a lot more crime, but people take protective measures, the actual crime rate only goes up a little bit. The potential crime is some multiple of actual crime.
But crime has been decreasing for decades, and by the same logic, that should mean the threat of crime has decreased even more than actual crime. E.g. as the city becomes safer, people go on their evening jogs again, and stores hire fewer security guards, so the actual crime rate decreases less as people reduce precautions.
A second question is what does it mean for the thieves if property theft is down due to precautions? If theft is down by half (I'm making up example numbers), but attempts at theft have doubled, thieves are getting half the reward for twice as much time and effort. You would expect that to discourage theft attempts. On the margins, thieves should give up. People generally spend less time on things as they become less worthwhile. So we would want some explanation for what's driving more crime attempts even as it becomes a worse option.
You didn’t give an example of crime going up, you gave an example of crime being more common in one place than in another place, and finding that out the hard way. Nobody said crime is equally low everywhere.
Yup. My example is meant to illustrate how disorder can be meaningful to people and yet invisible to Scott’s methods. I can’t say from personal experience that disorder has gotten worse in any one place, since I’ve moved around the country a few times.
The people where I live now definitely have a sense that disorder has gotten worse since the 90s. But I wasn’t here then, so I can’t compare.
Imagine 2000 people who don't lock their car doors live in your old neighborhood. 1000 move to your new one.
Of the people who moved, 900 of them never have anything stolen. 100 of them have some stuff stolen from their car, 90 of whom don't report it, and 10 of whom do.
Of the people who didn't move, 990 of them never have anything stolen. 10 of them have some stuff stolen from their car, 9 of whom don't report it, and 1 of whom do.
From this data, Scott concludes that there are 10x more thefts in the new neighborhood.
The real world will be messier than this. But most of the time, if the rates of people having stuff stolen and not reporting it rise, the rates of people having stuff stolen and reporting it would also rise. If we suspect this isn't the case for some reason, we can find other sources of data to test alternate hypotheses. Scott used data from the National Crime Victimization Survey, which asks random people if they were victims of crime, to address the very problem you're concerned by.
The point of my example wasn't the theft, but the door locking. Look at the first comment in the thread. People who feel like they have to watch their porch all day to prevent packages from being stolen are having a bad time, but the NCVS doesn't notice, because the package didn't get stolen.
If minor crimes and the defensive behavior that protects against minor crimes are in an equilibrium with each other, then you would only expect a rise in lawlessness to result in a rise in reportable crime when people's ability to defend against it starts to fail, and the equilibrium breaks. (Or in cases like my own, where I thought I was too cool to have to defend myself.) But you would expect people to complain more when lawlessness rises and they're having to take a lot of defensive measures.
It gets a tricky, because people who are in a moral panic _also_ take defensive measures, and they also complain. That's why corporations locking up their goods gets mentioned; presumably they're only sabotaging their business that way because they're losing more money in shoplifting than the defensive measures cost in lost sales and payroll. Cold calculation, not moral panic.
Realizing that with the typo "gets a tricky" I sound fake Italian, and with the sentence "Cold calculation, not moral panic" I sound like a darn AI.
You're exactly right! That's **cold calculation** — not a moral panic.
> If minor crimes and the defensive behavior that protects against minor crimes are in an equilibrium with each other, then you would only expect a rise in lawlessness to result in a rise in reportable crime when people's ability to defend against it starts to fail, and the equilibrium breaks.
Not necessarily. The thread-heading comment implies equilibrium has already broken, because crime is “only” down due to everything being locked up or surveilled. That could happen if the cost of crime has gone way up or the cost of surveillance has gone way down, regardless of the underlying lawlessness.
Your incident would show up in victimization surveys.
The theft would. The subsequent door-locking wouldn't. If you look up at the first comment in this thread, it's about people's defensive behavior changes in response to crime not being reflected in the statistics. That's what I was trying to explain.
The crime discourse is increasingly sounding like the "God of the gaps" argument to me. When someone presents data showing that crime is not up, certain people insist that the crime must be where we do not yet have data for.
+1. Really does not feel like the pro-crime people have a grasp of basic Bayesian reasoning or truth-seeking, it should be obvious that crime is down way before reading Scott's post, never mind after.
You could consider it part of the cost of crime. If we increase the amount of time and money we spend trying to prevent crime (cameras, guards, locks, etc.), but see the actual crime rate be mostly unaffected, the cost that the perceived crime level has imposed on us has increased. I don't think society does a great job of considering those costs when we implement programs concerned with safety, so it results in a lot of extra resources going towards things like crime prevention, which may not actually help at all, thus increasing the true costs of crime, even if actual crime levels aren't changing.
I'd expect that you wouldn't spend enough time and money to bring it all the way back to the base crime rate.
It may take less than you'd think, if crime is actually relatively rare, and the costs of prevention are spread among everyone. According to https://www.safehome.org/safest-cities/ there was about $27B in property crime in 2023, which is a ton, but only breaks down to around $80/person. I can believe that the average American feels more than $80 in losses from crime prevention than they do from the actual crimes. This is even more true when you consider that there are only about 6.4 million property crimes per year, or one for every 50 people.
I'd definitely believe that people think there is more crime or disorder, or whatever you want to call it, just because our prevention costs increased, regardless of the actual impact it has had on crime.
Yeah, one of the big complaints I hear is about how a whole bunch of merchandise at the store is locked up when it didn't used to be. At least in my area, that is inarguably true.
Now, we can argue until we're blue in the face about why that's happening, whether it's necessary, whether these companies are behaving rationally, etc. etc. And that's all interesting. But at the end of the day I still have to spend 5 minutes waiting for an employee to come get a circuit breaker out of a locked case at Home Depot, which I didn't 5 years ago. And that's equally annoying regardless of why it's happening.
That is indeed very annoying but it's a corporate decision. You're not going to solve *that* problem by hiring more police or taking a more "tough on crime" approach to politics.
> You're not going to solve *that* problem by hiring more police or taking a more "tough on crime" approach to politics.
This is only true if the corporate decision is completely independent of the rate of shoplifting in the area.
Not sure about that. About 5 or 6 years ago there was a noticeable increase in shoplifting at local stores, and today, there's a noticeable increase in the amount of goods that are locked up and require employee help to access. I realize technically this is "post hoc ergo propter hoc" or something in Latin, but to me it seems very plausible that the former caused the latter.
The former may have caused the latter but that doesn’t mean that a reduction in the former will uncause it.
I don't think the reduction will be immediate. Even if shoplifting immediately dropped to zero it would likely still be at least a few years before all the locks were removed. But over the long term, yes, I do think a reduction in crime rates will lead to a reduction in anti-crime measures.
If we're only arguing about the object level issue of whether crime is currently lower than it was 5 years ago, then to be clear I do think you're right about that so maybe we're wasting our time here. My point here is that my life, and a lot of people's lives, are noticeably worse now in certain regards than they were in the 2010's, and that part of this worsening traces back to crime even if the crime spike itself is mostly over.
This is exactly wrong. The problem was solved. Many places used to be tough on crime, and the prior state of unlocked circuit breakers was perhaps even a manifestation of that policy that everyone took for granted. If we lock up enough criminals for long enough, we can reverse it.
For circuit breakers in particular, part of the issue is that there are newer styles of breakers with added safety features that are a lot more expensive, so there's more incentive to steal them and hurt the store more when they are stolen. Old-school basic circuit breakers look like they cost around $7 at my local Home Depot, while ACFI breakers cost $60+. A bit over a decade ago, US electrical codes started requiring ACFI breakers for circuits feeding bedrooms.
That said, a bunch of other stuff that doesn't have this problem also seems to be locked up these days.
This is technically true, but only if 'hiring more police and taking a more "tough on crime" approach to politics' does not actually result in an environment reducing the incentive to crime. Which might be the case! Frequently "do X" politics does not result in X being done. More police hired might go to places with low crime and high influence, and thereby do nothing for the overall crime rate.
Surely we can agree, though, that if shoplifting is no longer a risk and all else remains constant(!), those locks and plexiglass panels will at least eventually not be included in new stores and the percentage of products locked up in stores will decline?
No, I don’t think that’s necessarily the case. The TSA was put in place in response to 9/11. It’s been 25 years since 9/11 and terrist attacks involving planes have been very rare in the US since then, but we still have the TSA despite it being very inconvenient. If shoplifting goes down after the introduction of locks on merchandise, then corporate is going to assume that the locks are doing the trick and keep them in place regardless of whether they or the police are what has stopped it.
Perhaps, but Walgreens has its incentives lined up in a way to make sensible decisions here, and the US Federal Government does not.
I agree with Melvin that the incentive structures are lined up differently for private companies vs the federal government. If the locks were actually doing the trick, what I would expect to see is cycles of fewer locks, shoplifting rising, and more locks. There's always pressure to remove a cost, celebrate the additional profit, and get a promotion or polish that resume and look for a job that's a step up. Things that are effective but costly have to prove themselves effective on an ongoing basis in such a corporate culture. In the federal government, DOGE notwithstanding, the existence of a workforce is its own reason for existing and being funded.
I think it’s helpful to consider exactly how corporate is going to make its decision. The problem is the visibility of the cost vs the benefit. As far as I can tell, there are basically two costs to locking up your mechandise: the cost of materials and installation, and the cost resulting from the inconvenience of having to unlock the merchandise for every customer who wants to make a purchase. Of the two, only the first one is going to show up on a balance sheet and it’s a sunk cost. Removing the locked cases is an additional cost so from a balance sheet perspective removing the locked cases might look like they cost more than keeping them even if that’s not the case.
Keep in mind that corporate managers aren’t stupid, they understand that there is a cost to keeping locked cases in place, but they can only really make decisions based on information available to them. Inventory loss due to theft and other causes is easy to measure, but the cost of keeping merchandise behind glass because that’s mostly going to be lost sales and that’s going to be noise against all the other factors that impact sales. When you have a measurable cost vs an unmeasurable cost the measurable cost is usually going to win. I used to work in business intelligence and I used to see this sort of thing all the time.
If it's just "venting annoyance at having to wait 5 minutes" then it doesn't really matter why it's happening. But the majority of these discussions are either explicitly or implicitly arguing for *policy change*. And the policies that would address the annoyance are dramatically different if the cause is "shoplifting is up and stores must protect themselves" vs "stores are employing fewer people, and so are less able to watch for shoplifters and have installed countermeasures even though shoplifting isn't any more prevalent" vs "nothing materially has changed, but stores are being pressured by higher ups to look like they're Tough On Shoplifting" vs "there has been a surge in funding for anti-shoplifting measures despite no surge happening so the money is being spent because it's there". It's irresponsible at best to advocate for a specific fix while not caring about the difference between those (or other) causes.
I don't know why more people don't talk about the locked up merchandise as a consequence of reduced staffing? It used to be the case that my local CVS had between 4-6 staff members on the floor at all times, stocking, working the registers etc... They could keep an eye on the merchandise. Now there are between 2-3, and instead we have self check out and locks. I imagine this is cheaper than hiring staff.
Because the locked up merchandise only exists in high crime areas?
(Or at least the minimum dollar value of the locked up merchandise varies a lot depending on the local crime rate.)
You would think you'd need more staff to handle needing to unlock a goddamn cabinet every time a customer wants to buy a toothbrush.
Yes, and people adjust their behavior to create a thermostatic effect on crime. Part of why car thefts and carjacking are down, for example, is widespread adoption of habits like not leaving things in your car, using a wheel lock if you have a Kia/Hyundai (visible everywhere in major US cities), and not idling in your car. But from the POV of the car owner, it feels like things are perpetually getting worse, since "back in the day" you used to be able to just leave your stuff in your car, idle outside your friend's house and scroll on your phone while waiting for him to get home, not bother with a big awkward metal bar on your steering wheel, etc. The loss of those things is a real quality of life effect that won't show up in crime statistics but will show up in "do you feel safe" and "do you think crime is up"?
Matt Yglesias has a post where he says that voters' definition of "corruption" is basically "politicians doing things I don't like," even if that's just voting for a bill I wouldn't have supported. I think crime is similar: "crime" means "stuff other people do, or might do, that threatens my quality of life," whether or not it actually results in a Bad Thing happening.
ACX titles are getting out of hand...
Thank your for writing that post as thoroughly as you did. And of course people who want to be upset will continue to find ways to be upset. Please don't stress out about it. The title of your post was very clear, and your post dealt with the title topic.
If you want to write another post about disorder, please do so! If you don't want to that's fine also :-)
My 2 cents, this is why your “much more than you needed to know” posts are the gold standard. If you edited both of those posts into one post, it makes it harder to argue against you. I recognize this is more work for you and the algorithm favors multiple shorter posts, but it’s probably more persuasive to anyone that actually cares enough to read in depth on the issue.
I agree with all your observations here, but as Scott said I think 'part 2 on Disorder coming next week' probably optimizes across the different trade-offs here.
Or at least, it's worth trying that format a few times to see how it works.
That and explicitly saying "Hey, I'm rebutting against this specific argument from this source, but I acknowledge there are other dimensions to the broader topic which people will feel I've missed".
I am also in favor of Scott's comprehensive posts, against the people complaining that it's too long to fit in a tweet https://x.com/TeaGeeGeePea/status/2018814285984694323
My favourite Immanuel Kant quote is "if I had more time, I would've written a shorter letter"
I thought that was Mark Twain.
After looking it up, we're both wrong; it was Blaise Pascal!
"I never said half the things I said" - Yogi Berra, MLB player and stealer of pick-a-nic baskets.
Seems like a combination whataboutism and goalpost moving, but where you can choose to move only one inch in some direction and dodge the argument.
But in reality lots of people have lots of different ideas so maybe nobody does either of those. You land in one argument spot and there will always be people 10 inches to the right left or some other dimension that can reasonably say "you didn't hit me!"
In reality I liked your last few posts and changed my mind on crime rates, but yes, one single mega post would be better.
This is why I think your "more than you wanted to know about X", such as guns, posts are so good. They cover a large enough space to really prevent that issue.
>I don’t know how to get around this.
Trust in data was a strange and fragile thing, incredibly historically anomalous, and enough entities responsible for trust and data screwed their Mandate of Heaven so hard that Uriel now has severe PTSD and a bad limp. Now we are in the time of ~~monsters~~ epistemic agnosticism bordering on nihilism.
I'd like to say enjoy it, the water is fine and the cookies are good, but I can't really be sure of that, either. Tend your own garden and try not to think what happens when we cease to understand the world.
>no, crime rates really are down
A big flashing header that you're only talking about the US might dissuade at least a few comments about Europe's issues, where the data collection problem runs screaming past Poe's Law and various forms of disinteresting biases. Or for a *slightly* less CW example than crime rates, Ed West's quite reasonable article about the many ways the UK population estimates range by millions: https://spectator.com/article/we-have-no-idea-how-many-people-are-living-in-britain/ Illegal immigration? Bad data collection? Fraud? All that and more!
Related to the crime question, the occasional mask-off moment of a leftist actually saying they think of crime like weather probably doesn't do anyone any favors when trying to convince anyone to Scott's right that crime is down: https://x.com/hecubian_devil/status/2025731262132920820 One idiot like this fuels a million confirmation biases that "the left" (writ too broadly) really doesn't care and so why trust any numbers from "the left" (again, writ too broadly) at all?
For me it seems you are in the business of finding truth and reporting it regardless of its political direction. It's a commendable position and rare. I sometimes disagree with your posts but I don't believe they are ever bad faith. What's my point?
Do you have listen to us?
I don't want you to get disheartened about what you are doing and stop or even slow down. I guess all writers have to listen to their audience at some point to hone their craft but are you at the point where you can just trust your own instincts and ignore your audience forever more? I think I'd vote for that - even if your audience was just me! Or when writing about correcting misconceptions do you have to listen to the feedback to get more insight into the misconceptions?
> I don’t know how to get around this. On the one hand, it’s a problem if people are saying false things, and nobody can correct them without getting mobbed by a bunch of people accusing them of committing malicious streetlight fallacy, muddying the debate, using Dark Data Journalism to steamroll over lived experience.
> On the other hand, it’s a problem if malicious streetlight fallacy can never be challenged, because perpetrators can always defend themselves by appealing to some hypothetical group of people who think Mexican immigration is worse than Central American immigration and are lying to convince people that it’s Mexican immigrants specifically.
This is precisely what cruxing was invented to address, yes? Your interlocutor has to be able to point to something tangible - and preferably predictable - that's the honest source of disagreement, so that addressing it honestly *resolves* the disagreement. If they know what the point of contention is but refuse to offer it, that's the essence of bad-faith argument. If they don't know what the point of contention is, they have work to do in making their beliefs pay rent. And if they're yet another in the endless string of anons with bespoke niche disagreements, then sometimes you have to pick your battles.
And ofc, lionizing ideological consensus or illegibility for their own sake rather give the game away at the beginning.
Right, but that's a method for a dialogue between two people.
When you are writing a persuasive essay to a large audience, every member of that audience may have a different crux.
And since the article seems to be written to persuade them, if it doesn't address one person's *personal* crux, then it *feels* like it's intentionally misrepresenting them.
Addressing a broad audience is a difficult proposition.
Good point. Expressing disagreement isn't productive if people aren't willing to specify what, exactly, they disagree about.
People experience things relative to expectations, not absolutely. People may be angrier about disorder than the absolute amount of disorder merits, if they expect less disorder, for some reason.
One reason I can think of is that, relative to the “good old days” of the early post-war era, the US taxpayer has paid several trillions of dollars, at least, to fix disorder.
If it’s the 1950s and there are no expensive programs to address disorder, you’re not as mad about it as if it’s the 2020s and they’ve been spending tons of our money on fixing disorder for decades and the best they can say is that it’s no worse than before.
If you hire a plumber and pay them a lot and after a lot of work, the pipes are still clogged? It only makes you angrier for the plumber to say, “Well ackshully, the pipes aren’t any worse than when I started, so what are you complaining about?”
I've really been enjoying your clarifying posts on crime in the US, for what it's worth, and thank you very much for your diligence in writing them!
I wanted to say this on the crime posts, but never got round to it.
Those people who were arguing against you on the crime posts should have been welcomed! But welcomed as only one datapoint among many.
You can't demand that everyone believe the truth, even when it's true. Imagine if crime has fallen, as you suggested. It still must be true that for some fraction of people, crime has risen. Their viewpoint is necessary and valid! Your post was arguing that they are in the minority - and that was amply borne out in the comments. Lots of people offered contrasting viewpoints to the "crime up" commenters.
But the crime up commenters could definitely have been right about their own experiences. I saw nothing in the comments to disprove them.
The crime up commenters were wrong to think that their personal experience disproved the larger statistical trends that you were describing in the posts. But they were still right for them, and I think everyone should have been more chill towards them.
I wonder about crime perceptions skewing because of distribution of crime, to be honest. Like, I lived in ATL for ages and it feels noticeably less safe in the good neighborhoods than it did 10 years ago, but if crime went down substantially overall but had a minor increase in randomness, then the bad neighborhoods would dominate the stats in having less violent crime but the median commenter-who-never-went-to-those-neighborhoods-anyways might be more worried about random crime. Might be hard to fully map out to test though. But feeling less in control of one's ability to avoid crime is much more about where crime happens than about total amount of crime happening.
Is 'amount of crime with white victims' a sufficiently precise metric to crux this question for you?
Based on the stats Scott cited, I'd be very surprised if that metric wasn't also down, but it's not impossible and it's something we could check.
That'd probably work as a proxy in a lot of places, though might vary regionally. I might try to dig up the appropriate ones if I can.
I'm just trying to put sense to the subjective difference in how things feel on the ground. But it also is tied into a just sort of weird antisocial vibes aspect I think, where folks aren't as convivial with other folks on the street as they were pre-covid, and the mental health crisis homeless are significantly more visible in cities like ATL (though they're aggressively trying to shunt them out for world cup prep, so there's also a humans-being-shifted-more-than-normal aspect at play too). But I know that there's a lot of places I'd have walked alone a decade ago that I'd def feel more nervous about, and I'm not inclined to be overly precious about such. People on the street feel less predictable in interactions than before.
I'm looking forward to the post about the disorder!
If you're not joking, it's already out: https://www.astralcodexten.com/p/crime-as-proxy-for-disorder
I hadn't been joking. Thanks.
I wonder how much this relates to the phenomena of "Summer of the Shark", where what gets reported can make people anxious about something that hasn't even changed (or is improving). In general, I think it's a mismatch of scale, where people expect the frequency of hearing about bad thing X is proportional to how likely they are to experience it themselves. The US is huge with a population of ~350 million though, which means that you only need a base rate probability of about 1 in a million chance of occurring per person per year to have a daily example of said thing to harp on about. If the probability is higher than that, then even if it's decreasing, the news can continue to harp on about it. If the news itself is going from a low rate of reporting to a high one, it can even reverse the apparent direction of effect by making it seem like it's increasing, because you are hearing about it more often, even as the actual rate is decreasing. Social media only makes all of this worse because you can get into some bubble where the only thing people talk about is 'horrible thing X just happened again'. This constant information input is providing you with the integral of bad thing X happening over time and space, instead of the local rate, but it's getting treated as local rate.
Yes, I expect this is mostly driven by media ecosystems, and in particular a backlash against progressive politics (inasmuch as 'crime is high, we need conservatives to fight it' is a long-term stable meme in our culture).
I'm brave enough to admit I'm a work-from-home computer jockey who rarely goes outside, my priors on how much disorder there is in the country is almost entirely driven by a murky cross between my media exposure and my ideological commitments.
Given the audience for this blog, I have a strong suspicion that this is a source of a *lot* of commenter's priors on the amount of disorder, but I haven't seen many others admit it.
I could always be wrong, of course.
I was looking at shark attack stats and holy shit are bites several orders of magnitude lower than I thought. Like, 10ish fatalities per year iirc?
Still, how many people used to swim in those shark-infested waters compared to before? If people are still afraid enough that those waters are still emptier than before, than there's still worries that needs to be addressed. And before the waters are full again with visitors and without shark attacks, no statistic about shark attacks would matter.
On a meta level this is an issue I have with discourse generally, and people who specifically make "caring about the truth" a core part of their identity specifically.
One of my favorite pieces of rationalist writing is Zvi's series on simulacra. He basically posits four ways of communicating, each further removed from caring about object-level truth:
1) "There's a tiger across the river" - there is a tiger, the jungle cat, on the other side of the river
2) "There's a tiger across the river" - there might be a tiger across the river, that's irrelevant to me. I don't want to cross the river.
3) "There's a tiger across the river" - people in my group are saying there's a tiger across the river so I'll be popular if I also say there's a tiger across the river
4) "There's a tiger across the river" - I think I can curry favor with groups that help my goals if I advocate for crossing the river.
While this framework was formative in how I think about communication, it's extremely flawed. Because nobody says "there's a tiger across the river" to say an interesting thing about a tiger and a river. Nobody. That's not real. Otherwise why didn't they say, "Ducks don't have thumbs" or "My resting heart rate is 82 BPM?"
Statements are never made absent the context of some goal. And so people who think their goal is "truth seeking" without doing a lot of qualification are actually just unaware of their goals.
Which isn't to say that the truth is irrelevant. It's more relevant than ever. And I'd rather people try than not, so good on the rationalist community for doing so.
But good luck coming to a consensus. And not because of malice. Because we are creatures in the dark shining tiny flashlight beams on the smallest possible part of an elephant, thinking the rest of the elephant is probably irrelevant.
I came to the comments to say the same thing.
Behind every communication there is an intent. When you write a post about crime statistics, it might feel as though your intent is "debunking fallacies", but there's a reason you're debunking fallacies, too, so there's an intent to that. You're trying to influence somebody, somehow. Maybe it's out of an autistic truthseeking for its own sake, sure, but I doubt it. The tone of the discourse suggests something more like: assertions that crime is high are being used to justify stances which you find threatening; therefore disproving those assertions ought to weaken others' interest in those stances; therefore you try to publicly weaken them. There are some ancillary reasons as well, of course: probably a bit of autistic truthseeking; plus 'blogging is your job' so you gotta post something; plus I imagine there are a lot of downstream points and cases you want to make that rely on the facts being established vis-a-vis crime rates so it's a building block in a longer-term project. But ultimately, it is a *goal* of yours to changes others' opinions on this matter.
Meanwhile the people replying to you have a goal also. They feel threatened by the state of crime they feel they live in, and they wish to debunk claims that this state is not real, because if it is not real then their feeling of anxiety and unsafety is justifiably ignorable, and nobody ever (EVER) wants society to be able to justify ignoring their anxiety. Some of these peoples' fear is caused by, well, their fear being actively fomented by e.g. fox news / twitter panics / doomscrolling.
But some of it is real also, and real *despite* the arguments he made. For example it is essentially a fact, not a misconception, that "bikes left out of in San Francisco will almost certainly be stolen within a day" (note: my anecdote here is a couple years out of date since I do not life in SF; it is just an illustration anyway), whereas that is not the case in the recent past or in many other cities. If you say property crime is down, yet it feels like my example fact is true, then they simply see you as mistaken in a fundamental sense. And it is easy, trivial really, to extrapolate a reason why you are mistaken: in this case, because bike theft got so prevalent that everybody stopped leaving their bikes out, and also because the bike theft is so unhandled by the police that people stopped even reporting the, so via multiple mechanisms the lower rate in the statistics is DUE TO the greater prevalence of crime in actuality. And of course there are a million other similar arguments for all the other crimes. Nobody cares about the *actual* rate of crime when they're calculating what to be mad about; they care about the one that they experience in their life.
So anyway we have to look at the goals of communication. Scott's goal is to establish that crime rates are lower (and implicitly: tell people whose experience of unsafety is greater to get over it). Their goal is to have their feeling of unsafety acknowledged (and implicitly: signal that he can choke on his statistics). It's probably true that he is more 'right' in a scientifically-certifiable sense, but nobody cares, because communication is about purposes and his purpose includes negating their fears.
> I don’t know how to get around this.
If you actually care to get a different reaction, stop negating anyone's anxieties. Only synthesizing gets a positive response. If you don't care then, fine, but don't get aggrieved by the result of not caring, because you signed up for it.
> " Scott's goal is to establish that crime rates are lower (and implicitly: tell people whose experience of unsafety is greater to get over it). Their goal is to have their feeling of unsafety acknowledged (and implicitly: signal that he can choke on his statistics). It's probably true that he is more 'right' in a scientifically-certifiable sense, but nobody cares, because communication is about purposes and his purpose includes negating their fears."
This is *BY FAR* the smartest and most useful thing anyone has said on this series of posts, including Scott himself.
Because as a living-above-a-notorious-homeless-encampment-and-working-downtown-and-thus-seeing-crime-every-literal-day-Seattleite, crime is very fucking definitely not "down" ENOUGH. Not enough for me to use the bus system to commute to work (a stop on my route was closed for years due to crime! That route had OSHA-violating levels of airborne fentanyl!), not enough for me or anyone with dogs or kids to use the park neighboring my home, and certainly not enough to feel "safe."
So why the everloving fuck is Scott - who says himself he avoids going into San Francisco - spending so much time and energy attempting to convince everyone to believe "crime is down?"
What can possibly be the benefit of convincing people it's "down," when there's still much too much of it?
> What can possibly be the benefit of convincing people it's "down,"
Even if the destination is very far off, it is still important to be clear about whether we are walking towards it or away; if we are walking towards it, we want to walk faster, not turn around and reverse all our progress to date; cries of “this isn’t working! we are nowhere near where we want to be, we need to radically change direction” are counterproductive, and we need to push back on that and call for walking faster instead.
I will double down on my point: that you have to parse purposes, not facts, to understand communication. If Scott wants a different reaction he needs to synthesize his purpose for the original post with their purpose in responding. Which is to say, 'listen'.
This is tangential to your point, but it reminds me how insane it drives me when I read about polls saying that 72% of Americans feel that the country is heading in the wrong direction.
But we are not going in the right direction. The reductions in crime are seemingly unrelated to policy, as they've happened worldwide. If people really want crime to be all but eliminated, then this centuries-long movement towards policy and culture that provides mercy for the downtrodden needs to be reversed first.
If you argue that crime is never affected by policy, you will then need to do way more work to explain why your proposed policy will affect crime.
I never said they're unaffected. You were implying that social progress was responsible for the decrease in crime rates, when the crime rates are decreasing worldwide regardless of how regressive cultural attitudes are locally. Crime rates improving is simply an inevitability with the changing world, but that doesn't mean changes in policy were responsible for it. There are countries like Japan who have much less sympathy towards liabilities and minorities, and they have far less crime than we do despite having less wealth to work with. Radically changing direction from the current liberal policies towards such people could offer massive improvements.
I think you're hedging here by half-admitting that maybe Scott is right about the overall crime rate but then railing against him for even bringing it up when crime is still too high, and (I think wrongly) impugning his motives for bringing it up.
Suggesting that Scott's secret motive is to "tell people whose experience of unsafety is greater to get over it" is totally crazy and off the mark in my view. Correcting misconceptions is itself a common human goal. If everyone thought that beluga whales only dove down 500 meters deep, and you found evidence they can dive over 1000 meters, then of course that is interesting information to tell people. There's no hidden leftwing/rightwing motive. People being wrong about something is interesting to us all on its own.
I don't think that Scott 'feels' that as his motive. He feels like there is a factual misunderstanding and also he feels people debating him on it and he wants to win. But *functionally* that result feels like ignoring/negating those people's experiences. The fact that he does experience his personal motive as equivalent to its functional result is the reason he's getting so much pushback--because other people do. The point is that progress will be made when understands how this looks at the level of purposes and does something about it.
>He feels like there is a factual misunderstanding and also he feels people debating him on it and he wants to win.
Which is different from the Beluga whale example... how?
I mentioned this in my other reply: the thing to analyze in the whale example is not "why the person thinks they spoke" (to correct factual misunderstandings) but "why the person spoke", that is, why the person wants to correct factual misunderstandings. The two reasons may have basically nothing to do with each other, but people's responses are based on the truth of the action, not what the speaker believes about the action.
Perhaps the whale-fact-speaker thinks they're correcting people but it is part of a general pattern of trying to sound smart and superior in unwelcome ways. The audience's reply (say, rolling their eyes, or feeling admiration) will be to the reality of the action, not the speaker's belief. Of course the audience might incorporate what the speaker believes they are doing into their response (e.g. "I understand you're trying to correct facts but it feels like you're just trying to sound smart.") That depends on what their motive in responding is.
Also let me add. The 'purpose' aspect of your whale example is: what is the speakers' purpose in saying something about the whale at all? Why are they trying to share interesting facts or accurate knowledge in the first place? For attention? Out of boring? Out of a mild compulsion for correctness? Because it is normal to do so? Etc. There must be some reason because they said it instead of not saying it. If you only analyze it at the level of facts then you have entirely elided the important part of the interaction.
> Why are they trying to share interesting facts or accurate knowledge in the first place? For attention? Out of boring? Out of a mild compulsion for correctness?
Yes, those are potential reasons and that was my point about that. A desire to correct misconceptions is a common motive on its own. You could (for example) fairly describe someone with a mild compulsion for correctness as seeking truth.
The point that there's some reason Scott thinks this is worth saying is fine. The problem is you aren't just doing that; you're making assumptions that he has particular left-wing, soft-on-crime motives, and in particular that "people whose experience of unsafety is greater" should "get over it".
You are very much misunderstanding/pattern matching to the wrong thing.
I don't know where you got anything about me ascribing left wing motives, I think no such thing. I am saying that functionally when you say 'no crime is going down' you are invalidating the concerns of people who feel otherwise and that's why they react so strongly. This is just a simple and predictable mechanism; there is no political angle to it. I'm not claiming that he understands or intends that he's trying to invalidate them ; I'm saying that he *is*, because that's what refuting people's anger does. He believes (perhaps) that he is refuting misconceptions, but by choosing to invalidate them (instead of entertain them them) he is clearly stating how he feels about the emotional stances he is replying to, which is that they don't matter to him. If he had the opposite purpose he could easily write an equally factual article that supports them instead. He seems to basically not understand how any of this works, or he would not be confused about the reaction.
Let's refresh on Scott's stated motive for writing this series:
>"My goal isn’t to deny anyone’s lived experience, nor to discount the importance of solving these problems (I support the death penalty for boom box carriers). It’s to push back against a sort of Revolt Of The Public-esque sense that everything is worse than it’s ever been before and society is collapsing and maybe we should take the authoritarian bargain to stop it. On an emotional level, I feel this too - I can’t go downtown without feeling it (one of many reasons I rarely go to SF). But I don’t like feeling omnipresent despair at the impending collapse of everything. Having specific thoughts like “house prices are up since the pandemic, so it’s no surprise that there are more homeless people, and more of the usual bad things downstream of homeless people”, rather than vague ones like “R.I.P. civilization, 4000 BC - 2026 AD” isn’t just more grounded in the evidence. It’s also more compatible with living a normal life. I’m not a pragmatist who thinks you should be allowed to lie or do a biased survey of the evidence in order to live a normal life and escape despair. But I’m also not some kind of weird anti-pragmatist who makes a virtue out of ignoring evidence in order to keep despairing. "
Scott actively avoids going into downtown San Francisco because of his feeling that downtown San Francisco is worse and collapsing, a feeling presumably based on literally *seeing* how San Francisco is worse than it was 10+ years ago and sure looks like it's collapsing.
He states that he doesn't like the feeling of despair at the impending collapse of everything, so he decides to literally avoid San Francisco, a place where despair is a completely appropriate response to the visible environment, and write about crime going down *nationwide*, so that he doesn't have to"feel the despair."
And shares it with his readers so that they, too, can - if you will - "get over" said despair.
My personal opinion is that this is a misfire and very much not a good use of his attention or time, and upon reflection, is particularly irritating if he's correct about the other subject on which he's trying to persuade folks, upcoming advancements in AI.
If AI is destined to advance as much as he believes, then there actually *is* an impending collapse of everything for the many (most?) people whose livelihoods will be replaced by a bot without any alternative. Right now the plan for their survival is being drawn up by the Underpants Gnomes, so the aggressively cheery articles about "crime is down, get over your despair like I did" aren't sitting well with some of us.
Why should Scott have to write about San Francisco instead of the US? The national crime rate is a perfectly valid topic and it's more relevant to all his readers outside of SF.
> And shares it with his readers so that they, too, can - if you will - "get over" said despair.
He shares it with his readers so they can have an accurate and informed view of the national trend.
If crime is worse where you live, by all means, you should be upset about it. But it's not fair to criticize people for talking about a trend because the trend doesn't hold somewhere. No trend will be true for everyone.
This is silly, and both you and Scott should know better.
National crime trend statistics are irrelevant to an individual's risk, which is why, e.g. car insurance companies use extremely granular information about your location when assessing a price to make you whole should your car be stolen, broken into, or crashed into by an at-fault person committing the crime of not carrying insurance themselves. They don't care about the national averages when it comes to assessing an individual's risk of being victimized in an auto-related crime, and neither Scott nor anyone living in places with crime and blight above the national average should be comforted by a dropping national average of crime if *their* *city* *is* *so* *awful* *they* *actively* *avoid* *being* *in* *it*.
Perhaps Scott should write a separate article on Seattle. Like, how various types of crime evolve in Seattle over time.
He can write one on San Francisco, as that's more or less the same thing; he's apparently aware that the city is effectively unusable enough to openly state he avoids going there himself.
It's nice to have the resources to do that!
Well, I would definitely be interested in both.
(San Francisco is more relevant for the rationality community, but I feel like I heard a disproportional number of horror stories from Seattle considering how few people I know who live there.)
Vancouver, BC, Seattle, Portland, San Francisco, and LA are all looking and feeling pretty dire, especially compared to what they were like 10+ years ago. (My personal observation is informing that statement. More vacant storefronts and buildings, more litter and graffiti, more open drug use and illegal camping, more human waste, etc. Like, sorry, I don't care what a study says, you can literally *see* it, check historical Google street views, it's *different.*)
Official crime statistics about these places should be viewed with a bit of skepticism, as there's a strong ideological motive by city leaders to downplay the volume and impact of crime, disorder, and blight. Seattle has a whole division of street team workers dedicated to cleaning up especially egregious messes made by trespassers/loiterers/indigents, as well as trying to move them along when they bother businesses too much. They're quietly literally cleaning up crimes, which prevents those crimes from being officially reported as the crimes they are. Lies, damned lies, and statistics, etc.
It also seems unclear that the studies are controlling for all the new measures being taken to thwart crime - if shoplifting is steady despite a recent trend of stores spending a lot of money locking up items and hiring more security and using increasingly advanced camera systems (Target is basically the CIA now), obviously crime would be *much* higher without those strong *new* deterrents in place. Criminals haven't gotten less crime-y, victims are putting in greater effort to minimize the harm done to them.
That's not a victory.
This argument frankly sounds like a schizophrenic arguing why the CIA is actually stalking them. Sure, the data might not agree, sure there is no tangible evidence of this, but I saw a broken twig on my way to work which was clearly a mistake by one of the agents stalking me. Also, why does that schoolbus drive by my house every morning? If the CIA isn't stalking me, why have I seen that Walmart employee look at me twice in the past month?
If your "lived experience" trumps data, why even bother collecting data in the first place? After all, it'll only be used to justify your position in retrospect.
I think "staying attached to reality" and "practicing good heuristics" is Scotts main motivation for the crime posts, not anything political. I guess what one person considers good epistemic practice is another persons autistic truthseeking.
> This argument frankly sounds like a schizophrenic arguing why the CIA is actually stalking them.
this is too stupid and antagonistic to bother replying to
Wow, you really signaled your superiority with that one.
+1
I think the issue is that disorder invites the potential for serious crime in a way that makes the true crime rates irrelevant if you're living with the disorder.
I saw a tweet thread once that pointed out that a common anti-social behavior, listening to loud music on public transit, carries the implicit threat of death if you confront the listener. Everyone has seen the many different stories of unstable or psychotically antisocial people who stab the person who asks them to turn down the music; and so when you're around someone breaking social norms like that, you're not just "dealing with the hassle of having your focused disturbed", you're dealing with an implicit death threat; or the threat of getting punched, or at the very least being screamed at and menaced.
And the people who violate social norms know this, and it's part of why they do it: it's fun to dominate the crowd. Will you turn your music down if someone asks you, or will you fistfight them, or will you kill them? Only you know the truth, and they don't, and so you live in comfort, with everyone treating you like a potential murderer, as if you're a warlord and they're peasants.
Even if "real crime" is "down", "situations that invite the threat of real crime" does not seem to be "down".
Someone else pointed out that, say, sharing a train with a loudly raving, clearly deranged person, triggers feelings of physical insecurity in people (and rightly so), even if that person eventually does not stab or shoot anyone.
IDK why Scott does not address such comments. He treats perceptions of crime as misunderstanding of hard statistics instead of something that is reciprocal to perceptions of personal safety. Which may be satisfactory for very exact minds, but it is also societally and politically tone-deaf and unproductive.
How would you address these comments? Are there any stats for people raving on the subway? Maybe it's also less common. Maybe it's more common. Maybe it's the same
The argument is that these unmeasured crimes likely track the measured crimes that he does discuss. This is part of the “crime as proxy for disorder “ post.
If your complaint is that he addressed some forms of disorder, but not the specific one you wanted, that's the point of this post.
In some cities here in Europe, there are attempts to make "perception maps", where people can say "this and this corner or park feels extra unsafe" etc. Subjective, but better than nothing. Over a longer period, this could actually translate in something that can be analyzed. Then we would see whether these really track each other or not.
Because he’s pretty obviously making a data based argument which contradicts vibes and the comments keep trying to push it into places where there isn’t data to look at but only vibes that could be used to support whatever the speaker wants them to, even while the original argument showed us that the vibes are way off the reality where the data does exist.
Why don't people who play loud music on transit have to live with the same threat of death that an unstable or psychotically antisocial person may become annoyed with them for it and stab them, scream at them, or menace them over it? Are they all just real brave?
Typically, people who transgress social norms don't spend a lot of time thinking about what the people around them might be thinking.
> Why don't people who play loud music on transit have to live with the same threat of death that an unstable or psychotically antisocial person may become annoyed with them for it and stab them, scream at them, or menace them over it?
Because the cluster of "people who listen to loud music on the subway" also clusters with being more likely to be unstable, antisocial, violent, etc in a way that isn't symmetrical, and similarly, the potential outcomes are assymetrical for both sides.
Anyone with a career is unlikely to listen to loud music on the subway AND unlikely to confront the loud music listener, not just because the person may be a psycho, but because even if they weren't, if the police were called, it's more likely to hurt the career-haver than the music listener.
The police can basically raze the career-havers life to the ground arbitrarily if they "assaulted" the poor music-listener by asking them to turn the music down and it turning into an altercation. The music-listener is immune to that threat, they have nothing to lose.
So career-havers can't do anything for two reasons - 1) the guy may be a psycho, and 2) the police aren't on your side and are more likely to ruin YOUR life than to improve the situation if they get involved.
The fact that this is true is a large part of the current problem. No good actor can act on behalf of the commons or career-havers, and any bad actor can act with impunity, and the police are literally worse than useless - they won't fix any problems, and are more likely to ruin any career-haver's life who tries to do anything about it.
Because the violence correlates with other behavior.
Public transit users are such a tiny proportion of Americans (Census says 3-5%) that it makes me suspicious that this is always the go to example for public disorder.
If public disorder were extremely common surely people would be using more common examples that actually affect the other 95% of Americans.
Insofar as some fraction of that 95% doesn't use public transportation *because* of the disorder on it but otherwise would, it does affect them
I think there's a common perception that the character of crime has changed, independent of its quantity. And like, rationalists are going to hate this because it's probably impossible to quantify. I worked various crappy retail jobs from 2015 to 2022 or so. I couldn't tell you if the dollar value lost to theft went up, down, or stayed steady during that time. But the behavior of shoplifters changed, a lot. There was a distinct sense of... I guess you could say shamelessness that crept into the whole thing. "Person stuffs a few items in a purse and furtively sneaks out" and "Man sweeps whole shelf of merchandise into a cart and walks out while making unbroken eye contact with the clerk, knowing damn well he can't even confront him" are both shoplifting. But, call it irrational if you want, they *feel* very different to be around. Even if the dollar value is the same.
People have been sneaking onto public transportation for as long as it's been a thing, and I don't know if fare evasion rates are up, down, or sideways. But watching a group of teenagers jump the turnstiles *while a transit cop stands there watching them and does nothing* (true story) has me thinking thoughts of societal collapse and degeneration of public morals, in a way that wouldn't be true if they were doing it sneakily.
This shift in brazenness has probably been influenced by social media. When clips of people shoplifting started going viral, it spread awareness of what had always been the case: shoplifting is easy to do and difficult to stop. It's not necessarily the case that many more people were willing to shoplift, it just might be that many more people know it's easy to get away with it, because they've seen videos of people doing it.
I agree with the form your non-apology takes here, people mad at you for defining a topic you want to talk about and then talking about it are getting over their skis in assuming malicious intent or incompetence.
What I think you might *instead* consider from this experience is that, if people thought you were spotlighting when you weren't and you found it hard to navigate between these two failure modes in a public space, maybe have a little more charity and sympathy for the 'dark journalists' and others who look to you like they're employing these tactics.
We all know about the fundamental attribution error... you did something that looked bad to outside observers, but did it for good reasons, and it wasn't actually that bad. But when you observe someone else who looks like they're doing that same bad thing, they must just be a bad person doing the bad thing intentionally and maliciously.
No doubt there *are* bad people in the world trying to do evil, but we could all stand to take a bit more time considering possible motives and context *before* jumping to that conclusion. If your own innocent and well-intentioned actions can end up being perceived this way by outside observers, then maybe others who look like they're employing these 'dark techniques' are also just doing their best in a complicated world.
I am *always* this guy, given my job, but I really think it's important.
Crime is ALWAYS a very technical thing, and it's ALWAYS going to be subject to "well, actually...", that is literally the point of why the justice system is set up the way it is: we punish specific violations of statues, where very specific "elements" all must be proven beyond a reasonable doubt to the satisfaction of a jury, precisely because we don't want to punish based on "vibes."
One can literally see someone gunned down on a street corner in broad daylight and correctly "well, actually" that act. After all, it doesn't take too much imagination to see that the act of shooting someone MIGHT NOT BE A CRIME. Kyle Rittenhouse can tell you a bit more.
Labeling a particular act, or category of acts, as a "crime" has very specific consequences for a debater. They're now playing by the rules of law, where pedantry is the name of the game by design. Imagine someone turns to me and says "we should punish all criminals more harshly, and in particular, people who smoke cigarettes." The natural pushback is "wait, smoking cigarettes isn't a crime." What I'm saying there IS rather pedantic: I'm saying that within the jurisdiction I'm in, the general act of lighting up a cigarette does not technically meet all the elements of a specific statutory crime. That's precisely that it means to say an act "Is or is not a crime."
I realize I seem like a pedant when I do this, but the whole point of crime is that it's not categorical. Most general acts (tackling someone, screaming really loudly, sending an email, uttering the words "I am literally going to kill you, no joke", firing a gun) are criminal in some circumstances and not criminal in other circumstances, and we generally need to point to specific laws and findings by juries to tell the difference. I hope we don't need much creative thought to imagine circumstances when all the acts in my parenthetical are very serious crimes, and other circumstances in which they're perfectly legal.
BTW I'm going to have a blog post responding to your triad of blog posts on this sometime in the next few days.
Rittenhouse was acquitted because the court deemed that he was defending himself from attackers, and those attackers were thus indeed committing crimes.
that's not necessarily true. One can have justified self defense even if one's attackers are NOT committing a crime. As a toy example: let's imagine two people are both told that there is a deranged shooter in the neighborhood with a gun, going door to door killing people. To face this threat, they both grab guns, and step outside, see one another, and shoot. In one universe, A misses and B hits, B kills A. In another universe, B misses and A hits, A kills B.
In one universe, a jury for person "A" could easily acquit them, based on the defense "I thought the other guy was the deranged shooter I heard about." A jury for person B, one universe over, might likewise acquit on exactly the same basis. These sorts of self-defense scenarios occur, a little less cleanly, all the time, and critically as well, they're jury dependent. They necessarily hang on what a jury would conclude based on the facts presented to them.
I don't mean to say that in Rittenhouse's case, his attackers WEREN'T committing a crime...its very possible they were. But the entire encounter is illustrative of how hard it is to talk about "unprosecuted, uncharged, and unreported crime", how many crimes was that anyway? one assault for every person who struck him? moved toward him? Threatened him in any way?
also important point of further pedantry: the court didn't "deem" anything in the rittenhouse case. The jury gave a verdict of not guilty, and juries don't need to explain their verdicts. I agree with you that the jury PROBABLY thought he was properly defending himself, but we don't know that by any means, and it's a little hard to "thusly" conclude anything about his attackers as a matter of law or fact
Very well said. This is important to keep in mind.
There's certainly a lot of edge cases when it comes to crime. And it's even possible that when you add up all the edge cases there's more of them than non edge cases.
But surely this can't mean that we can pedant ourselves out of being able to meaningfully say things like "Tijuana has more crime than Tokyo".
I think even greatly uncontroversial cases dwarf all other possible crime on a scale that's difficult to articulate. I'm writing a blog post on it now. FTX defrauded a million people. Was that a million separate crimes? There are only about 100,000 crimes reported in Minnesota every year.
We end up in the land of pedantry due to neo-reactionary assumptions of the sort scott is addressing. We can talk about how much "crime there is in Tokyo" with how many crimes are CONVICTED or CHARGED or REPORTED since these are objective criteria and often discernable even without reference to statute (you can say how many "assaults" were reported without worrying too much about the statutory definition of assault in Tokyo)
But we end up in the land of pedantry when we assume bad faith on the part of these counters, and instead claim to be able to measure how much crime there "actually" is. I have to ask for the metric then. How does one count the frauds that are unreported? how does one know about every drug deal that transpires in the shadows? how does one partition the thefts? Most every murder is an uncharged assault. Most every robbery is an uncharged theft. Most fraud is a ton of little frauds, and even a lot of theft is a bunch of smaller thefts.
I wonder how much of crime perception comes from parents not talking about crime to their kids. Then when they become adults and start reading the news, they get the feeling: "Oh my God! Crime has been so much worse lately!"
I vividly remember in middleschool the first time I became aware that our country was 'at war' and what that actually materially meant, and I thought it was some type of extraordinary alarming state of affairs that I needed to be hugely concerned and scared about all the time.
Of course, the US has been 'at war' for over 90% of its existence. Social studies class eventually taught me that this was not a state of exception, but rather extremely common and normal. But I don't think anyone gets explicitly taught this about crime.
I worry about how Kids These Days learn about the news.
In my pre-internet childhood I started learning about the news because the TV was tuned to the news at 6pm every night, and then when I got slightly older I learned more about the news from reading the newspapers that were always strewn around the house (because I was bored, and there wasn't a magical box constantly feeding me the things I was mathematically most likely to engage with).
My kids don't get any ambient exposure to the news, so I have no idea how they're eventually going to have to learn about it all at once, probably from someone with a specific agenda.
One thing I cannot find reference to in the commets on both posts is the possibility of changing laws / redefining the entire concept of a crime? Famously, it is one lever that governments have accessible, and since they tend to be judged on the published numbers (up until some threshold where the divorce between opfficial numbers and reality is noticeable to the naked eye) there's a strong incentive to use it.
This should be something we can check, also: have there been redefinitions of felonies into misdemeanors, and misdemeanors that become legal, etc over time? Could this be visible by discontinuity analysis in the data or is the resolution too small and it's been death by a thousand cuts over time?
I asked opus 4.6 (deep research). intro quote:
"" California has reclassified, downgraded, or decriminalized dozens of criminal offenses since the 1960s, with Proposition 47 (2014) representing the single largest mass reclassification in state history — converting roughly 40,000 annual felony convictions to misdemeanors overnight. These reclassifications have created a minefield of statistical discontinuities that make longitudinal crime analysis treacherous. The primary distortion affects arrest data (felony vs. misdemeanor counts), not UCR Part I offense counts, because the FBI defines offenses independently of state severity classifications — a theft is still a theft regardless of whether California calls it a felony or misdemeanor. Yet indirect effects on policing behavior, reporting incentives, and offense categorization ripple through every dataset. Neither the FBI nor the California DOJ publishes adjusted time series; researchers must navigate at least seven major structural breaks spanning 1976 to 2024, with the NIBRS transition adding yet another layer of complexity."" "
https://claude.ai/public/artifacts/b9e6b517-f83e-4fe7-b996-f8f465794005
matching crimes across jurisdictions is a famously vexing problem in the law. I generally think it's best to think of crime as a purely jurisdictional artifact, rather than some general proposition. There is no such thing (in america...with a few very weird exceptions...) as a "common law crime"
so, what is "murder" in one jurisdiction is not the same as what is "murder" in another. There are certain acts that might be "murder" in Minnesota that might not be "Murder" in Arkansas...because we have different murder laws. Crimes must be specifically articulated in a statute...and every jurisdiction writes its own statutes.
now, for differences in murder these are weird edge cases. If I say "imagine a murder", the thing you envision in your head is probably murder everywhere, but as we work down to the sort of ambient crime that is everywhere, this gets complicated. Some things are felonies in one state that aren't even crimes in another (Marijuana comes to mind). Other things are misdemeanors in one state, and felonies in another. Some jurisdictions have a thing called "Fraud" which is different than "theft", but in my jurisdiction (minnesota) they're literally the same crime. So in minnesota "theft of the shoplifting kind" is statutorily the same crime as "theft of the kind where you pad your hours in an attempt to defraud a subcontractor for some mega-corporation."
confusingly, not every jurisdiction even has the same definition of "felony" and "misdemeanor." In minnesota we even have a whole separate kind of crime, the "gross misdemeanor", sorta halfway between the two.
True about matching exactly - but luckily we could just estimate a dampening effect based on the downscaled acts numbers? For CA, it means at the federal level and local laws. And if someone was _very naughty_, maybe check for retroactive coding corrections in the databases - that'd send a spanner in the analysis...
Problem seems equivalent to a motte-and-bailey doctrine.
This is a failure in analysis I think: This assumes mistake theory, when actually the situation:
Is mistake theory up to the point where someone would have to challenge what they want, then it was conflict theory the whole time.
See: The right wing free speech absolutists who were all about your right to the freedom to say whatever you want, in whatever way you want, on whatever platform you have access to, until EG. people started making jokes about CK being truly, completely committed to the right to bear arms, upon which suddenly it wasn't about that after all.
This is another POSIWID situation, where judging from what people say vs. judging from what they do produced radically different goals.
What is "EG"? Who is "CK"?
Charlie Kirk?
Yup.
EG stands for exemplie grattia I think, traditional abbreviation for for example.
CK is charlie kirk, but it was pleasing to use CK after EG, it gave the run on sentence a pleasing symmetry for me.
Just an observation that nobody is fully mistake maxing or conflict pilled; people flow form one state to another based on the circumstances and their feelings on the matter.
It should be either e.g. or, less formally, eg; EG is never correct.
But then it isn't symmetrical with CK, so I choose to reject the rules of grammar in exchange for aesthetics.
Sometimes the challenge to what I want is never going to be affected by what I say (see the number of elections where any seat is decided by less than 1,000 votes), and I just want to pretend to be a good person for a while. What do I do then? I have the same problem Scott’s described, and I knew something was wrong but I didn’t know what. I’m hoping someone finds an answer
> and I just want to pretend to be a good person for a while
I'm pretty sure that's what charity and volunteering are meant for. Personally, I just stopped caring about being a "good person". It's pretty liberating, you should try it out some time.
I’ve tried it, and it didn’t work for me.
Quit pretending and just do what you know is right all the time. It isn't hard once you work out your axioms, which is a fun thing to do in it's own right.
Scott, you're talking as if there's a clean and clear distinction between malicious streetlight effect and a legit attempt to address a confused issue.
But the distinction between these two always turns on prior beliefs about 1) the empirical facts, and 2) what one values in the situation. In other words, there is no "context free" way to distinguish these two.
You can call out errors of empirical claim, if you have sufficient arguments and data, but you can't dictate the larger import of your correction.
I think the reason you heard so much about disorder is that people like to complain, and if you put up evidence that something they like to complain about is not so bad they look for ways to legitimize complaining.
Seconded, with the addendum that it's not just complaining for complaining's sake, it's also indictment against an ideology/politics you dislike, and towards one you prefer.
I think the issue is that "decreasing over time correctly-chosen timescale" is not the same as "not so bad". Things can be decreasing and still bad!
It's perfectly legitimate to complain about things even when they're not as bad as they might have been at some point in the past, and that's true of left-coded complaints like "inequality" just as much as it's true of right-coded complaints.
The problem in these debates is that "actually this problem is not as bad as it was at some point in the past" is taken as meaning "stop complaining about this problem you rubes", which gets people understandably upset.
<The problem in these debates is that "actually this problem is not as bad as it was at some point in the past" is taken as meaning "stop complaining about this problem you rubes", which gets people understandably upset.
Yes, I agree that's the problem. But I would locate that problem entirely in the people who get upset and then discharge their upset in the form of posts spewing disgust and annoyance about problems that are not the one under discussion. Doing that is not a person's only option when a blog post that clearly is not saying the reader should just stop complaining sets off an irritating fantasy that it does. And it someone really is not able to do anything but that -- well, this isn't group therapy, you know?
I would say that your problem and the problem of guys like Cremieux and countless others are that you are mistaking statistics with knowledge. You feel entitled to speak about fields you did not study, research or got direct experience of. None whatsoever. Your contribution could charitably described giving your vibes about the validity of the statistic at hand. By chance your vibes invariably are pro-business and everything-is-fine-but-you-are-too-dumb-to-realize-it. You are just increasing the noise and drowning the signal. Stay in your lane and the world will be grateful
Is your lane knowing exactly where other peoples lanes are and correcting them? If not, perhaps you should take your own advice.
Taking the framing that communication usually takes one of the three forms of information-exchange, rapport-building, or doing-battle, talking about crime tends to fall under doing battle. When doing battle, the other party does not absorb and integrate facts. Rather, it attacks them. Thus, people will either question the source or the framing of the data, or, barring that, fall back into some sort of conspiracy about how data is being suppressed at the data collection level. I would love to see an example of someone saying "oh, I think I was wrong about this." You won't find many such instances, though, because this is one of those worldview-defining positions of the right-wing mindset. People that think that crime is up, because this is pumped into their heads 24/7 by Fox News, will never be convinced by facts contrary to those beliefs. The left seems less wedded to their view on this (that crime is going down) and seems mostly to just take the opposite view out of principle. However, I believe it's always important to put the facts out there and to try to reach the minds that might listen, even if it is mostly futile. So thanks for doing the work, Scott.
"My plan was to publish a post one day on crime, and then the next on disorder, but I got so many negative comments the first day for talking about crime without mentioning disorder that I guess in the future I’ll include in the post that disorder is a separate topic and I’ll talk about it later. I don’t know a better way to thread this needle."
Is there a reason why you did them as two separate articles instead of one "Much more than you wanted to know"? It seems like they would have been stronger together and avoided much of this issue.
Like, the first essay on crime was ~2k words and the second disorder essay was ~2.5k words, while your "Prison and Crime: Much more than you wanted to know" essay clocked in at ~13k words (1). It seems like you could have combined both of those essays into one bigger essay, plus some more, and those have been better received in the past and avoided this issue.
(1) https://www.astralcodexten.com/p/prison-and-crime-much-more-than-you
Well he said he wanted an easy article to link to whenever someone said "crimes are way up!". I'm fine either way, mostly because the gap between posts is short, but I guess he'll be back to the MMTYWTK format know.
I appreciate the attempt to apply reason to this issue. The problem is that the government, the media, and academics are not trustworthy. It would not be shocking at this point to discover that officials were too lazy to even put their thumb on the scale tactfully and decided to fabricate the data outright. People trust their eyes and ears, and to some extent what they see online, because that's the only real information available. If the numbers contradict their general sense of the situation, then they will assume there is something wrong with the numbers or the analysis. If Pravda publishes numbers saying food production is up, but the shelves are in fact empty, would you trust the news? Would you even bother to quibble about the meaning of "up" vs "adequate"?
I'm old enough to have now been an adult through two different noticeable declines in violent+property crime in the US -- mid 2000s through mid 2010s, and now the faster/deeper decline starting in 2023 -- as well as the 1980s/90s crime levels which briefly recurred during 2020/21.
During all of those periods I resided in either a core city (one very large, one much smaller) or a suburb of same. And then in addition for the past 15 years my work has taken me regularly into rural parts of the US having population densities drastically lower than anyplace I've ever personally resided in, which has been eye- and mind-opening on some topics including this one.
Fully granting that this is simply one person's anecdata, here's what comes to mind from pondering the above life experiences and this topic:
(a) OF COURSE the rate of general public disorder tracks with the rates of non-trivial crimes. That sure as hell has happened where I live now, as it did (in each direction) during various previous decades when I lived in various parts of the US. I asked a few people in my age bracket about this and to a person they first waited for the punchline to my joke setup before realizing that I was asking a serious question. [I may, possibly, be somewhat known for such joke setups.] Their collective actual answer can be summed up as, "Um, _yes_....so hey is everything okay with you?"
(b) The obvious reality that those two things correlate does not _prove_ that they are linked nor if so, how (which comes first). I've never yet seen a really persuasive case on the specifics nor the direction of such a linkage. That said, having at this point seen it come and seen it go multiple times I would if in policymaking office take it as just common sense that those two things do go together somehow.
(c) Also OF COURSE, none of this is going to boost Democrats in any national elections during at least the remainder of my lifetime. Two reasons for that.
(c1) One is that American progressives and fellow-travelers have well-earned our comprehensive loss of public standing and trust on this specific very-high-salience topic. Even if we were all now reckoning with our errors and omissions on this, which way too many of us are very much not, the hole that we spent a generation digging for ourselves is too deep be filled back in quickly.
(c2) The other is that many people, not limited to Americans and not limited to social conservatives, are simply hard-wired to be sure that crime and disorder is higher than it used to be and/or than it reasonably should be. That's a big part of "very-high-salience". This is not a new fact, is not because Internet nor because "if it bleeds it leads" in the MSM [I used to believe that one] nor because social media; etc. Why it is true I'm not at all sure. But it is true, and will remain so long after I'm no longer around to compose grumpy-old-guy comments online.
> OF COURSE the rate of general public disorder tracks with the rates of non-trivial crimes.
Just an idea: Maybe this is true locally, but not globally. It could simultaneously be true that:
1) In any given city, the general public disorder is positively correlated with non-trivial crime. The streets containing more homeless people and drug junkies are better to avoid, because you are more likely to get attacked there, etc.
2) Over time, the ratio between the general public disorder and non-trivial crime changes globally. There is less non-trivial crime these days, globally.
To put it simply, a street with homeless and junkies is *more* dangerous than a clean street; and it always was. But a 2026 street with homeless and junkies is *less* dangerous than a 1976 street with homeless and junkies.
So one side goes like "of course, homeless and junkies = danger, and I see many homeless and junkies around me, what do you mean that the crime is low?"
And the other side goes like "yeah, but you are less likely to be stabbed by them today, than you would be in the analogical situation 50 years ago, that is what we mean by the crime being low".
Today you may see stolen goods sold openly on the street, but 50 years ago someone would have robbed you instead. You see the bad part, you don't see the alternative.
Well sure, that seems simply obvious as a matter of simple fact.
Seems though pretty much like just another useless well-actually type point. Irrelevant, at best, to real-world public/political response. “People used to get drunk and kill each other in their cars way more than they do now” does not impress anybody who’s furious that in the year 2026 some yahoo with a DUI on his record was able to retain his drivers license and now has obliterated a member of that person’s family.
Of course. No matter how much things improve, people want them to improve *more*, and I think that's okay.
It's just that when people take "things are extremely bad today, the worst ever" as a fact, and start looking for a scapegoat -- because that is the typical human reaction -- when I wish more of them took the more abstract perspective. (Ahem, make something something great again, etc.)
I have thought more about this since I wrote the comment you quoted and read your next post (and please edit to use my name and link to the full comment, and also for any time you quote me in the future). My conclusion is that I don't actually think it is a matter of the actual crime rate or the sense of disorder, but rather the tolerance of crime and disorder. When I see the police walk by people selling stolen goods on the sidewalk and do nothing it fills me with a blinding rage that seeing 100 thieves arrested will not. When I see videos of people stealing from CVS with impunity, and the security stands by and does nothing because if they do they will be punished by the law while the criminal goes free, it stirs a hatred in me that no level of crime in the statistics ever could. This is theft of my tax money. This is treason by the government that is supposed to represent me. Not only do they do nothing about the criminals, but actively protect them.
But then are you actually mad at the police and dedicated to firing most of them and reforming the system from the ground up?
Because most people who talk about crime and disorder seem to be pro-police, and in favor of giving them more funding and less oversight.
More funding and more oversight would seem like a good combination.
No, I don't want to waste money on that idealistic crap. I used to be of the mind that law enforcement was overdone and should be done with a lighter touch, but boy was I wrong. These people only speak one language, so I just want to see the batons come out for the people bipping cars and stealing from stores.
This may be about second-order effects.
As a toy example, imagine that there are only two types of crime, let's call them "big" and "small". Every year, there are 100 big crimes and 100 small crimes. The police only has capacity to investigate 100 crimes. Investigating a big crime or a small crime requires the same amount of time.
Would you prefer that the police investigate 100 randomly selected crimes, or 100 big crimes?
The response that considers only the first-order effects would be like: Of course, if you can't investigate all crimes, start with the big ones, they are more serious!
But if you consider the second-order effects, you can see that if you make "only investigate the big crimes, ignore the small crimes" the policy, it will have a huge psychological effect on both the small criminals and their victims. You have practically made the small crime legal! So you can expect it to grow, but even if it somehow doesn't, it will still be psychologically different for both the criminals and the victims.
So maybe a better policy would be something like "investigate 75 big crimes and 25 small crimes", to avoid the effect of de facto legalizing the small crimes.
I am sympathetic to resource constraints and suboptimal allocation issues. Getting that right is extremely hard. But that isn't what is happening in cities like SF. They are overflowing with revenue. And even in the most resource constrained city and worst crime wave imaginable I can't accept that the police will go after people defending their property while letting the criminals go.
Then why not name your articles "assault and murder are down"?
Just "crime" would include every illegal staying in country and fraud in paperwork. People also mentally measure financial "crime" and drug crimes including scams in their dollar amounts and not number of cases.
First I don't see how this maps onto the streetlight effect. Second and more importantly I wonder why most discussions have to be so adversarial. Why can't we talk about things more civilly, like we're sitting next to each other in a bar. I sometimes feel like the evil overlord is pushing everyone's buttons to make them angrier (or something). Don't listen to the evil overlord! :^)
> Why can't we talk about things more civilly, like we're sitting next to each other in a bar.
Because attacking people who make your life worse is an effective strategy for getting what you want? Sure seems to be working out for the right. People are rallying around hatred, and finally taking effective action against the people and systems that denied them their rightful position in society. Hard to argue that it's not working.
>>Because attacking people who make your life worse is an effective strategy for getting what you want?<<
Is it really though? Doesn't this just cause more anger and hate?
>>Sure seems to be working out for the right. People are rallying around hatred, and finally taking effective action against the people and systems that denied them their rightful position in society. <<
Is it the right that is rallying around hatred and such, or is it both sides? It doesn't seem to me that anyone is getting what they want. At least that's my take locally (I live in rural America). I woke up this morning feeling like grace was close at hand and that the right path is through love and God and the cosmic all.* And I hope to keep that throughout the day and pass on that love to those around me.
Be at peace my friend.
*I've been reading some fairly woo filled books, "A Course in Miracles" at the moment. Back to the good words of the book.
> Is it really though? Doesn't this just cause more anger and hate?
Not when you finish the job. That's why the left could never win. All that power, and they just never did anything meaningful with it. Their very ideology prevented them from imposing their will through force. The right isn't going to repeat their mistakes.
Sorry that you got so many negative comments. Thanks for writing these posts this week. I enjoyed reading them as usual.
Many of the people you're dealing with here are simply liars, to themselves and to others. You cannot "thread this needle" because this needle is not willing to be threaded.
I think a steelman (of sorts) of the "it's still the reporting bias" position is that the police are getting better at lying. In the past, it was possible to obtain some independent crime stats, and so the police wouldn't lie too much; but now, essentially no stats can be trusted at all, and the only data one can obtain is via personal experience.
(I've talked to some global-warming deniers who advanced the same argument: that true climate data is essentially unknowable, due to political pressures.)
I appreciate your willingness to seriously engage critics. It's probably very core to your thinking style, and it's a big part of your distinctive appeal as a writer.
At the same time, I think you might be subjectively better off if you had a little more condescension towards people who refuse to engage in good faith. Or even those who try, but are incapable of meaningfully understanding.
The whole purpose of that post was to respond to people who say "But what about reporting bias?" in a way that took their objections seriously. If someone responds, with no further clarification, like:
"Yeah, but have you considered that maybe the numbers are wrong?"
That person can safely be thrown into the "not worth considering" bucket. It seems like these kinds of responses bother you, that you interpret them resulting from some failure on your part instead of theirs. IMO, it would be better for you to just feel a little jolt of derision and then forget about them.
It's a bit risky, if you do too much of that you end up being pretentious + wrong + uncorrectable, but in small enough doses condescension contributes wonderfully to mental well-being.
> I don’t know how to get around this.
I’m not sure this is something you can do on your own. It takes both participants in a dialogue to keep the dialogue aimed towards truth.
In general, people need to try harder to engage with what their actual interlocutors actually say instead of things they pattern match to, things other people with similar political views say elsewhere, things they could have said that there is a ready answer for etc etc. This is really hard, especially when you are emotionally attached to the subject, and takes continuous deliberate effort. I, too, am too often guilty of responding to others with rhetoric instead of logic. We as the commentariat need to do our part.
Time to go reread the sequences again…
I like how you go about categorizing fallacies. The "directional" fallacy of exaggeration and fallback, though, is the motte-and-bailey fallacy, not a new one. The Streetlight Fallacy is good and new (for me) though.
There is something like the motte-and-bailey fallacy which is, I think, legit. We might call it the Trump Error Tactic. Someone says there are 1,000 rapes by immigrants. Opponents then say, "No, there are ony 900 rapes by immigrants, you liar!" The tactic has worked, because most people think there are 50 rapes by immigrants because the opponents control the media and are careful never to mention the numbers. Trump uses this all the time.
I think this is the right observation. Most people don't actually care that crime is up or down, rather they care if it is high or low. 4 murders per 100,000 people might seem "high" to a person in 2026, even though it's a near-historical low, because their frame of reference isn't "back in 19XX", it's "compared to X neighborhood/city/country". People travel more, are more mobile, and of course the internet exists now so people regularly talk to others all over the world and learn how exceptionally high the murder rate in the US is.
In this context, the "crime is down" motte for the "crime is low" bailey in the soft-on-crime vs hard-on-crime debate is almost cliched at this point -- pretty sure I've seen it as a punchline in some dystopian fiction before, though I fail to remember any specific examples. I think Scott is right that people responded based on what they *thought* Scott was doing (apologizing for crime and disorder as part of a familiar refrain they've become conditioned to react poorly to) as opposed to what he was actually doing (making a very narrow argument against another very narrow argument).
Out of all the pixels used to discuss crime over the past few weeks here, one thing I haven't seen discussed is whether people are just programmed to worry about it a lot, in the same way a gazelle is probably programmed to worry about big cats, and that maybe this leads people to perceive crime in our locales as generally bad, because otherwise, why would we be so worried about it? From an evolutionary standpoint, you can imagine why this is advantageous: you're better off as a gazelle over-reacting to that rustle in the grass because the costs of under-reacting if it turns out to be a leopard are really really high. Likewise, people are better off over-categorizing individuals around them potential muggers and rapists and so on because the costs of under-categorizing are potentially catastrophic. This leads us to perceive urban areas in general as more dangerous than they actually are and to engage in a bout of what Tyler Cowen calls "mood affiliation," where we search for arguments to support an emotional response we've had, rather than try to form beliefs based on FACTS or LOGIC, and that emotional response is fear or worry.
The other thing I would point out is the following: the major increase in crime and disorder that occurred in the US in the '60's and '70's was at least in the minds of some people enabled by the social upheaval that coincided with it (not saying it was all the hippies' fault or anything, just that the Boomer-fueled social revolution of the time had some negative consequences, one of which was higher rates of crime and disorder).
For better or worse, that social regime is still in place. The cultural norms and assumptions that emerged in the '60's and '70's are still around and largely unchallenged. We never had a Thermidorean Reaction or a Bonapartist Coup that turned back the clock. The Jacobin Club is still open and Robespierre's head is still on his shoulders (I mean, Reagan won a couple elections, but he was no Bonaparte, obviously). I think maybe there's a kind of unconscious association in people's mind that says "because the cultural attitudes of the '60's never went away, the associated crime it enabled never *really* did, either. It's just been temporarily suppressed by technology like Ring cams or modern trauma surgeries" or what have you.
Regards financial crime:
Didn't you know, "crime is legal" 😆
As long as you mumble something about Blockchain, and maybe make a few political donations 😉
I'm a big enjoyer of Coffeezilla on these topics.
Although I liked both posts, I am still a bit perplexed at splitting them up. Yes, there's value in clear standalone reference posts, and maybe there was some Speed Premium concern behind the scenes...but it seems quite predictable, even without d20/d20 hindsight, that there was gonna be a ton of Well, Ackshually conflation of low-level crime/disorder with capital-C Crime. Integrating both posts into a single cohesive piece, maybe a light rebrand as Falling Crime Rates: Somewhat More Than You Wanted To Know - there, preemptive disarmament of common retort. And much less likely to be screenshotted out of context by the Cade Metzes of the world.
Can't do anything about the "still asking questions already answered by t-shirt" phenomenon though, sadly...
You article was fine. Doing them separately was fine. This is a situation where you should ignore your critics.
You were not doing a "malicious streetlighting". There were specific arguments that people were making, and you were narrow and specific in what you argued against. The example you give of the bad thing you're accused of is helpful in highlighting the differences. If the claim is "immigration on the Southern US border is up", and the retort is "There are fewer Mexican immigrants", anyone can see that these are different claims. You responded directly to arguments about crime being up by citing crime statistics and arguing about crime. That OTHER people wanted to have a DIFFERENT conversation about crime-adjacent things and chose to excoriate you for not having the conversation they wanted you to have is them behaving badly.
You identify an 'on one hand/on the other hand' conflict in the end of your post. It's misplaced. There is no conflict to resolve. You were right, and the people who wanted to talk about something related but different were off-topic.
Agreed. I see this as more of a "blowing steam" kind of thing. Even I was frustrated while reading comments from the original post.
Agreed. It’s worth adding that people like myself, whose response to the post was, “an utterly unsurprising conclusion, but it’s nice to see it rigorously documented,” have little incentive to comment on the post. Since crime is down on average, people whose “lived experience” is rising crime are going to be a minority of the population, and quite possibly a minority of Scott’s readers.
"No, it is the commenters who are wrong."
...but unironically.
Framed as a game of Motte & Bailey Chess where a certain strong claim in the bailey is false and a certain weak claim in the motte is true:
Directional Correctness: 1. White is in the bailey and makes the strong claim. Black meets White in the bailey, and refutes the strong claim. 2. White retreats to the motte and makes the weak claim. Black points out White's retreat and refusal to concede the strong claim. 3. White accuses Black of Malicious Streetlighting. Black to move.
Malicious Streetlighting: 1. White is in the motte and makes the weak claim. Black pushes White into the bailey, and refutes the strong claim. 2. White points out Black's push and refusal to address the weak claim. Black accuses White of Directional Correctness. 3. White to move.
Scott, you have now written multiple posts in which you define "crime" and "disorder" as two separate and distinct things and use this distinction to try to tell us that everyone who thinks crime is up is an idiot who doesn't understand the facts, but of course we can probably all agree that disorder is up, that's completely different and irrelevant.
You define "disorder" as follows:
> Disorder takes many forms, but its symptoms include litter, graffiti, shoplifting, tent cities, weird homeless people wandering about muttering to themselves, and people walking around with giant boom boxes shamelessly playing music at 200 decibels on a main street where people are trying to engage in normal activities.
Just a super quick rundown:
1. "litter": Littering is a crime.
2. "graffiti": Vandalism is a crime.
3. "shoplifting": Shoplifting is a crime.
4. "tent cities": Camping on public property is a crime. So is trespassing on private property.
5. "weird homeless people wandering about muttering to themselves": While most people find this annoying, the muttering isn't really what makes most people nervous in this scenario. What makes them nervous is when some Jordan Neely wanders within earshot mumbling about how he's going to kill some white bitch today, because it makes them worry that this clearly violently psychotic individual is going to murder or assault them or their loved ones. Murder is a crime. Assault is a crime.
6. "people walking around with giant boom boxes shamelessly playing music at 200 decibels on a main street where people are trying to engage in normal activities": Violating noise ordinances is a crime.
I hope that helps you to understand why people didn't find your posts very helpful!
We do not agree that disorder is up.
For example I believe disorder is probably down, since measurable crime is down.
I feel what Scott is saying is "disorder is hard to measure, maybe it's down or maybe not but can we agree the measurable things are down please" and instead people are saying NONONONO CRIME IS UP AND ALSO DISORDER IS UP.
In 2010 the deodorant wasn't behind bars and there weren't hypodermic needles all over my kids' playground. Disorder is up where I live, relative to the pre-Woke days.
No doubt the disorder isn't evenly distributed. But are you saying that shit isn't more fucked up in your neck of the woods than it was 15 years ago, back before the Ferguson Effect and Defund the Police? Are you sure you're remembering correctly?
Sure, crime goes down when nobody leaves the house anymore. Want to look at those stats? They're grim. I work from home and don't go to movie theaters or restaurants anymore -- in this, I reflect broader trends that certainly contribute to lower rates of crime victimization. This is not because everything is hunky-dory; it's because movies aren't worth watching, restaurants have tripled their prices while service has gotten worse, and downtown has turned into a George Romero flick.
Get off my lawn.
Multiple people including myself made this observation in the comments of the previous post and I was really expecting to see it in a classic "Highlights from the Comments". I really thought I was misunderstanding something because I was sleepy, or maybe he'd phrased something poorly. The fact that we got this unprecedented refusal to engage is... concerning.
Anyone can have a bad day that leads them to write something that was dumb in hindsight, especially if they've moved to San Francisco and formed a progressive-heavy friend group. The doubling-down days later suggests an ideologically blinkered bubble that has corrupted his epistemics. The way they selectively change definitions to pretend that something isn't a problem is an example of the sort of thing he used to be against.
Yes. "Home invasions are actually DOWN since we turned into a Mad Max hellscape in which every family whose house is not protected by multiple machine gun nests is immediately raped to death by marauding biker gangs. Everyone without the machine guns got raped to death way back in 2023, but now in Q1 2026 home invasions are a real rarity! So the claim that crime is up gets five Pinocchios!"
Including disclaimers in the post about what common concerns you're not addressing can be a good idea, but shouldn't always be necessary. In my opinion, the problem is that a significant percent of blog readers expect your articles to address their specific perspective, and you simply can't write a single readable article that engages with a thousand different arguments at once.
"So you use FACTS and LOGIC to prove that something similar-sounding-but-slightly-different is definitely false. Then you act like you’ve debunked the complaint.
My “favorite” example, spotted during the 2016 election, was a response to some #BuildTheWall types saying that illegal immigration through the southern border was near record highs. Some data journalist got good statistics and proved that the number of Mexicans illegally entering the country was actually quite low."
The empirical world cannot provide proof. (Logic can in closed logical systems, e.g. proving mathematical theorems, but these are distinctly different epistemic domains than the empirical domain.) Not gonna stop harping on this because it's a basic stats 101 insight.
Are you saying something more obscure/interesting than 'probabilities cannot be 1 or 0'?
If so, I think I'm missing it and would like to hear more.
I feel like you're defining proof to only mean "mathematical proof" whereas the "proving something means providing overwhelming evidence that is convincing to most reasonable observers" seems just as valid to me? English isn't my first language though, maybe I'm wrong here ? 🤔
> this argument neutralizes a real and influential group of people trying to make the contrary argument that murder/crime rates are up, and to push policy based on that position
Isn't a big part of the problem here that we're trying to argue about numbers as a proxy for policy rather than arguing about policy directly? There's an assumption here, not quite stated and not quite fully believed, but smuggled in anyway, that if crime is "going down" over some timescale then we shouldn't do anything about it, and that we should only be dealing with problems that are currently going up.
This is of course silly; even if there's less crime than the recent spikes it's still far too much, far above the theoretical minimum, and even far above the practical minimum that could be reached with reasonable trade-offs.
Reasonable questions to ask when trying to decrease the crime rate include what tradeoffs would be reasonable (higher taxes? civil-liberties-violating random searches? draconian sentences for minor crimes?) but the question of whether crime rates are higher or lower than one, three, five, ten years ago are not that important; it can help us to gauge whether certain policies are on a right-ish or wrong-ish track, but there's a lot of other factors that go into the crime rate so it's unlikely to be that useful.
If we could dispense with the smuggled assumption that if crime is going down then we don't need to do anything about it, then we could discuss the numbers dispassionately and without trickery.
I think this just goes to show that there's no short cut for having the argument. You can write a post about X, people can argue that you're making fallacy Y, you can argue you're not (and maybe they're making fallacy Z), you go back and forth, clarify your positions and hopefully somebody somewhere changes their mind. There will almost never be an "argument" that is won after just one post, or one post followed by a knockdown accusation of fallacious reasoning.
The recent posts on crime and disorder are interesting and informative and they provoked some useful discussion (as well as, perhaps, some not-so-useful discussion). I don't see any need to reassess the way you approach issues like this - those posts were a success!
This is why the internet broke our politics. Not because of evil algorithms but because this sort of effect causes everyone to perceive the points made on the other side to be unconvincing strawmen.
We are always intensely aware of which subgroups we belong to but find it hard to make those distinctions about groups we aren't part of -- think parents mixing up goth, emo and punk or the like -- and we hate being mixed up like that.
This leads to an inevitable problem whereby some people broadly on the left [could be right, picked one for illustrative purposes] make some incorrect argument and to people on the right it seems to directionally capture what is problematic and mistaken about the left and they rip it to shreds. The problem is that to people on the left that wasn't even directionally their view, it was those crazy extreme greens while they are a sensible pro-growth lefty of whatever.
And honestly, I don't have a good solution. In the before times there was a clearer sense of what is within the mainstream of a given political ideology but now that we all have megaphones there is an unending drip of statements that are infuriatingly wrong coming from what you can reasonably label as 'that side.'
This creates a horrible trap where both sides can spend all day just telling you "can you believe people on the right/left said X" and each side will infuriate the other as it feels like they are attacking a strawman but they are real arguments out there that people are believing so....
just fuck fuck fuck
> And honestly, I don't have a good solution.
I just stopped associating with anything. You can't really get offended by groups or get offended by being associating with a group if you simply do not identify with anything in the first place. If people want to murder each other for stupid reasons, while that is my problem, it's one to be solved by figuring out how to navigate the situation without dying, not by getting attached to certain groups.
I meant I don't have a good solution for society. Yes not identifying too much or just being too contrarian to belong is great for Mr individually (well great in this regard...makes it hard not to annoy people)
The problem solves itself. The strongest group will eventually neutralize the weaker ones, and then you will have a period of relative cohesion and peace. Until entropy kicks in, of course.
I think people just absorb from an early age a verbal template for discussing collective problems in a Grown-Up way, and it includes saying that some statistics have changed a lot recently, which is bad. It's basically a news hook, like journalists typically put into op-eds and such. In reality, the feeling that a problem *exists* typically precedes any real knowledge of how the problem has changed over time. And there is a certain wisdom in this: if you're often accosted on the street by crazy violent guys, it isn't really that relevant to figure out whether the rate of accosting has gone up or down by 20% in the last year or two.
"So just say what you actually think is bad and stop pretending to care about quantitative window-dressing!" I agree. And I think it's good for people to rebut incorrect things that people actually say, even if those people don't really care about their own claims in the first place. But I think the realization that "problem is going up" is basically a trope, a literary device, helps to explain why narrow counterevidence rarely seems to chasten anyone.
> I guess in the future I’ll include in the post that disorder is a separate topic and I’ll talk about it later. I don’t know a better way to thread this needle
Please don’t beat yourself up about this - you are already more intellectually honest than 99.99% of writers, by posting follow-ups that surface your commenters’ objections and seriously engage with them. That’s why we are here!
Your existing process is exemplary, and here it is working as it should when you get more pushback than you expected.
Crime is always up over some arbitrary time span and always down over some other carefully selected time span.
For example, mainstream liberal publications are addicted to start counting crime in 1991. Why such a random-seeming year? It was the peak of The Wire Crack Wars. Once you start noticing this 1991 addiction, you can't unsee it.
Similarly, when people say that crime is up, what they often mean is that it went up recently, but not necessarily very recently. For example, the temporary triumph of the Black Lives Matter movement in the last week of May 2020 due to George Floyd's demise caused black deaths by homicide to increase from 903 in June 2019 to 1303 in June 2020, according to CDC WONDER mortality data. That was a 44% increase in the number of blacks dying by homicide.
Even more strikingly, black deaths by motor vehicle accident rose from 506 in June 2019 to 772 in June 2020, a 53% increase, as cops responded to the BLM hysteria by retreating to the donut shop, which encouraged people to drive more recklessly and pack their illegal hand guns as their fear of being stopped and searched declined.
Black deaths by homicide and car crash remained very high into 2022, but then the BLM fervor started abated and cops were encouraged to get back out there and start pulling people over and patting them down. And guess what? The end of the Floyd Effect saw black deaths by homicide and traffic fatality fall sharply, especially after Trump's re-election.
So, what lessons should be drawn from the Ferguson and Floyd Effects?
A common moral is that because crime is now down in 2026, we should completely forget about how elites drove black murders and traffic fatalities through the roof, twice, in the name of Black Lives Matter.
> Crime is always up over some arbitrary time span and always down over some other carefully selected time span.
Thank you for saying this succinctly.
Crime in the US may be down year-on-year, but it's way up on a rolling five years basis.
Personally, I find crime trends interesting, just as I find baseball stats interesting. But, then, I WOULD, wouldn't I?
Murders in the U.S. dropped rapidly in the mid-1990s after crack, flattened out for a decade, declined again after the economy crashed (murders, like traffic fatalities, tend to be Deaths of Exuberance as opposed to Case and Deaton's Deaths of Despair), and bottomed out in 2014.
So then elites got all worked up over Ferguson because the crime problem was, apparently, SOLVED. So murders went up in 2015-2016 due to the Ferguson Effect, then dropped in 2017 and especially 2018.
The murder rate seemed to start going up in mid-2019 as "New Jim Crow" thinking came back into elite fashion, then exploded during the last week of May 2020 due to the Floyd Effect. (Both the Ferguson Effect and the Floyd Effect were worst among blacks and were mirrored in traffic fatalities, due to to less proactive policing.)
Murders stayed high from May 25, 2020 (have I ever mentioned that date before?) through mid-2022, then started to drift downward as BLM fell out of fashion in the months before the 2022 midterms as Democrats started to worry that Tri-State Voters were blaming them for the increase in crime.
Murders continued to fall and dropped very hard after November 5, 2024.
It's almost as if murders and traffic fatalities go up or down depending upon whether American elites are telling the police (who, as their name suggests, tend to be highly political) to back off or go hard against crime.
Scott, I think you're doing great and appreciate what you're doing :)
I have nothing to add to the debate, but I have a nitpicky question: is there a difference between the "evil streetlight" thing in the post and good old strawman fallacy? "Defeating a weak version of an opponent's claim" seems to fit the streetlight description well here, no?
I do use the term 'directionally correct', albeit spelt out, for explaining things to my normal friends. Just the other day I said something like 'AI isn't as advanced as you might hear from CEOs, but you'd be more correct believing the CEOs than you would be believing the news', essentially a re-phrased version of directionally correct.
Still, I think that outside of low-time and low-information zones, such explanations ought to be avoided. But not everyone has the desire to spend all of their time studying esoteric things, so it shall remain used, at least by me.
The quality of American discourse on crime is very poor because crime in America is so fundamentally intertwined with questions of race.
For example, according to CDC WONDER data, in 2024, black males aged 15-34 died by gunshot homicides 22 times as often per capita as white males of the same age, 47 times as often as young Asian men, and, perhaps most interestingly, 5.8 times as often as young Hispanic men, who tend to be comparable in income and education, but not at all in terms of getting shot.
These numbers tend to make Americans very uncomfortable and drive many to switch to Shoot the Messenger Mode. For example, I've been asked countless times over the decades to explain what's WRONG with me that I'm not ignorant about basic statistical facts about my society, unlike all the good people who are empirically clueless.
Personally, my view is that knowledge is better than ignorance.
For example, tonight in the State of the Union Address, Donald Trump is expected to emphasize how cutting down on illegal immigration would reduce the crime rate. That's true, to some extent, but it's much less true for the worst crimes such as murder than most Republicans want to believe, because the murder rate is so incredibly high among Foundational Black Americans.
Of course, the liberal Establishment is at least as unrealistic, as we saw during the Ferguson and Floyd Effects, when elites got tens of thousands of incremental black lives murdered and splattered on the asphalt by making clear to the police to stop policing blacks as aggressively. But how many people know about not just the huge increase in murders but also the huge increase in car crash deaths among blacks when BLM was riding high in 2015-2017 and 2020-2022?
Okay, like, we're all in agreement here that a few ethnic cleansings would probably decrease the crime rate by some extent. You still have to convince people that ethnic cleansings are morally defensible and justifiable under the circumstances.
How is it morally defensible to demand less rule of law for blacks in the name of Black Lives Matter and then you immediately wind up with 44% more blacks dying by homicide and 39% more dying by motor vehicle accident in 2021 than in 2019?
If the crime is mostly about race, does the decrease of crime mean that the ethnic profile of America is changing in a good direction?
Or is it more like, if you believe that the ethnic profile of America is moving in a bad direction, and that the crime is mostly about race... then you are forced to believe that the violent crime is increasing, regardless of data?
As the American Establishment increasingly forgot in the early 21st Century how disastrously they'd screwed up during the soft-on-crime 1960s-1970s, they increasingly fell once again for the myth that the big problem facing African Americans was too much law and order.
So, from Ferguson onward in August 2014, the Establishment promoted the Black Lives Matter movement. That led to BLM triumphs over local police departments in the St. Louis area in the fall of 2014, in Baltimore during the Freddie Gray riots in April 2015, and in Chicago from November 2015 onward, all with disastrous local effects on the black-on-black murder rate (and similar if less spectacular effects on traffic deaths as fear of being pulled over declined).
After BLM terrorists murdered 8 cops in Dallas and Baton Rouge in July 2016, Trump got elected and BLM fell out of fashion. The national murder rate dropped into 2019, but started to inch up again.
Then came the events of the last days of May 2020, and the black-on-black murder and traffic fatality rates soared within days and remained high for several years.
My first suggestion: Let's not due that again. Let's allow ourselves to learn the data.
Ultimately, I believe it is incorrect to criticize someone for saying things that are true.
Let's say the crime rate goes up. And let's say that everyone takes heroic measures to stop crime. They lock up all the goods in the stores so that nobody can steal them. They raise taxes and spend the money on extra policing. They stay home at night and they avoid dangerous neighborhoods. They install expensive burglar alarms. These measures succeed in reducing crime, which is now back to its original level.
Is it fair to say that crime has gone up? It is literally true that in this scenario, the crime rate is no higher than before. But it's also true that this is the result of an equilibrium which leaves everyone worse off because of (what would otherwise be) an increase in crime.
This is why the malicious streetlight effect matters. "Low crime rates" is the similar-sounding but different concern. People are actually concerned with the effects of crime. High crime rates lead to bad effects from crime, of course, but so do "normal" crime rates that are only normal because people drastically reduced their quality of life to stop crime.
(Also, why didn't you take into consideration changes in age of population (older people commit less crime) and changes in racial balance? Those drastically affect crime rates.)
I'm not expressing confusion. I'm saying it reflects a problem, a bias, that corporate and white collar crime are so often excluded when people talk about crime.
And when people talk about "corporate crime" they should also specify "NGO crime," which appears to have become bigger problem in recent decades.
I don't see why Scott should give even a semi apology, and don't think it will help much with this problem for him to tell people in advance, "don't worry, I promise to post next about disorder." What's going on is that people are cranky and walking around with various resentments, and they feel a craving to express their feelings and related opinions. Then if somebody says something that appears to undercut the validity of some of their negative views, they feel a craving to argue. And because a lot of what's going on is mostly driven by irritation, they don't bother trying to argue well and fairly, but drag in stuff that's adjacent to the topic but not part of it -- for example, dragging in the Sidewalk Homeless Turd issue in a discussion of crime. Those doing this can distinguish between Sidewalk Homeless Turds and Crime with 100 IQ points left over. But they don't bother to do it because they are pissy.
Look, I have myself done the exact thing I am complaining about here. I get it. At the same time, I think that staying in emotion-driven mode in a situation where the task is to discuss and think fairly and clearly is by far the most common thinking error, and absolutely dwarfs all the rationalist-identified ones, and we should all try hard not to do it. It's not hard at all to recognize that you are in that mode: If someone's data immediately makes you angry, do some introspection. If the idea that their data might actually be *accurate* makes you feel sort of desperate and even more angry, then you know for sure that you have bonded with an opinion based on something other than good evidence. It sits well with you somehow. . You should probably abstain from discussion of the issue under dispute until you can figure out why you married Comforting Opinion and get at least a temporary legal separation.
In short, I think discussants, not Scott, should change. Don't type just to scratch your irritable itches. Jeez.
> Those doing this can distinguish between Sidewalk Homeless Turds and Crime with 100 IQ points left over. But they don't bother to do it because they are pissy.
Public defecation is generally illegal across the US. Whether it's a "crime" or a "civil infraction" depends on the exact jurisdiction; it could also be prosecuted as things such as public nuisance, disorderly conduct or public nudity, which are usually crimes.
While the exact classification as "crime" or "thing that is almost like a crime but officially not a crime" may differ from jurisdiction to jurisdiction, I don't think people need to get a law degree before they're allowed to complain about people around them doing things that are obviously both illegal and harmful.
Well Melvin, my point is that it just gums up the works to post about how bad disorder is in a discussion of crime, and that people posting here are capable of distinguishing between crime and disorder, and should not drag in problems better categorized as disorder. Perhaps I chose a bad example, or at least an imperfect one, if in fact producing a Sidewalk Homeless Turd is sometimes treated as a crime. In that, case, I should change my example to something like Being Disheveled and Stinky in a public place. Seems to me, though, that you must surely grasp my point, which is that there a lot of unpleasant things other people can do that fall under the general heading of disorder rather than into the crime category. If you are going to be a Turd Categorization Stickler about my example I will happily let you win regarding that sub-sub-issue.
Yes, Being Dishevelled And Stinky would have been a better example.
But I do think it's worth being aware that most of the examples put forward of "disorder, not crimes" are themselves still crimes, as argued by another posted elsewhere in the thread.
Maybe there really are circumstances where you really do see an increase in dishevelled and stinky people in the streets who are nonetheless not committing crimes. Maybe it's just an anime convention or something.
"I found that this was true - illegal immigration had shifted from Mexicans to Hondurans/Guatemalans/Salvadoreans etc entering through Mexico." My favorite part of this "favorite" thing of yours from 2016 is that - when presented with this clear evidence NOBODY jumped to the obvious conclusion that the US should pay MEXICO to secure their (much, MUCH smaller) Southern Border - serving both countries' security interests for a tiny fraction of the effort and expenditure in one go.
"Hey, a lot of people are crossing through and trespassing on your property and mine. But they're all coming though that gate you have. I'll happily by *us* a combination lock that you can put on there to secure the gate...how's that sound?"
Nope. Never heard anyone suggest that, even once.
Scott: this evening I wrote a blog post addressing your last few posts on the amount of "uncharged" or "unreported" crime that exists.
https://broodingomnipresence.substack.com/p/how-many-crimes-are-there-and-why
In this post Scott is using "directionally correct" in a way that I don't think is consistent with its 'original' usage or intent, or at least not the way i've seen it used everywhere else until now.
In this post he takes it to mean "slightly stronger than the truth can support. If your enemy committed assault, say he committed murder. If he committed sexual harassment, say he committed rape." By this definition the claimant is being hyperbolic and asserting the extreme version of the 'actual' situation. It's implied the claimant knows what is actually true and they are trying their luck at a juicier accusation in the hope it sticks. Isn't this just a motte and bailey, where the claimant reserves the right to retreat back to a claim the motte is 'in the same direction' as the bailey if the bailey gets undermined?
To my knowledge "directionally correct" was popularised by Peter Thiel in 2019 or a bit before. By contrast to Scott's usage, this was a way of saying e.g. "there may be exceptions, nuances and edge-cases my argument doesn't handle, but in general it's the best argument". In other words, the argument is good enough such that if we followed it to it's eventual policy conclusion, we would be heading in the right direction.
The massive difference in intent between the two usages, and the very high Pr(misuse of phrases by people posting online) is enough to make me much, much more reluctant to react in the way that Scott seems to be in this post.
No matter how Thiel (I guess?) meant it, the vulgar use of "directionally correct" that Scott uses here is the more common way the phrase is used these days.
I think the root cause of the disconnect is that complainants' concerns are mostly inchoate, but imprecisely rounded off to "crime" or "disorder".
Rebuttals to what is said fall flat because the actual problem isn't well-reflected by the specific terms used.
Everybody's talking past each other and, given the parlous epistemic environment of the 2020s, both sides dismiss each other as irrational, deceptive, confused, &c.
i feel for you
keep trying to thread needles please
Sounds like a difference between local and national stats, maybe?
Scott, you are a psychiatrist. You should get the hint that "crime is up" actually means "I feel unsafe" and then take a swing at why people feel so. Why one does not really see the 1960's style very carefree strolling down the street?
https://1.bp.blogspot.com/-X7tdUbPV6JY/YSOu2_XHOjI/AAAAAAAEL7c/9xKcx-GmfSUEDbKQ7jx6SnodgVpOJkUnQCLcBGAsYHQ/s1200/swinging-london-1.jpg
It's not lying, it is expressing an emotion in ways that are socially customary.
Can we say there is just generally more aggressivity in the air? Of course that can never be proven by the numbers-based approach you prefer. We might try measuring stress hormones tho.
Perhaps the correlation is the other way around. Perhaps 1960's people were hardened by actual crime and thus carefree, and today we got soft and anxious because nothing bad ever happened to us.
Or perhaps the psychological effect in play here is "I, personally, myself, am getting older and more frail, and less in connection to the youth culture so I cannot tell which parts of it represent real aggression and threats of violence and which are just playing around and fronting, so thus I personally feel unsafe regardless of whether this represents an actual rise in the amount of general unsafety".
Expressing your emotions by making false factual claims is in fact lying.
I think both "malicious streetlighting" and "directional correctness" are related to (maybe subsets of) a broader way in which the discourse is corrupted. Wrote about it here: https://meirbrooks.wordpress.com/2025/05/01/the-forest-and-the-trees/
Within echo chambers, people are bombarded by so many data points (whether true, misleading or outright fabricated) from "their side" that pointing out that a claim is inaccurate or even completely made up sounds like nitpicking or bad-faith argument. The claim you are knocking down is never a "pillar" of their overall view such that knocking it down would cause them to reconsider the whole structure; it's more like "I have 9,000 reasons to believe this, I'll give you two," in which case correcting the two just doesn't move the needle.
In fact it ends up sounding almost like a conspiracy theorist. Our usual response to "Why is the flag on the moon waving" or some anomaly with the video of the twin towers collapsing is not to take it seriously and investigate but rather to dismiss it due to our having been immersed in the overwhelming knowledge that a man did land on the moon and that 9/11 was a terrorist attack.
I think that's what's going on with "Their actual concerns were about disorder, open-air drug markets, tent encampments, and seeing people fencing stolen goods. They thought I was being deceptive in trying to trivialize these by saying that a similar-sounding-but-slightly-different concern, major crime like murder and assault, was down." I doubt it's that they're concerned about encampments rather than murder. Rather, they've taken in so many data points about crime and disorder generally that it's built up a "forest" of "Crime and disorder are shooting up everywhere because of [insert explanation here]" and then the point about murder rates decreasing sounds like trying to mislead with statistics.
I've had people of all ideological stripes tell me, when I showed them that a particular data point they had claimed was just flat-out wrong, that "you're missing the forest for the trees." I think that's the larger trend here, and it's easy to see how this is driven by internet-based news and social media.
Hate and angry sells. I think keeping us divided is part of the goal. My only solution is not to pay attention to the news. (I wish there was a better solution.)
Nothing to do with the above, but just because it's my personal question for people who worry about how divided we are-- do you try to talk to people on "the other side" (whatever that would mean in your case)? Have you had successes with substantive dialogue? I mean significant, ideally 1-on-1s.
I don't feel like I'm on any side. But yeah I talk to people on both sides. It's hard to have substantive dialogue unless there is a decent amount of trust from both people. But if so then sure I'll push back on what I think are wrong or misinformed views... but then again who am I and perhaps it's I who is misinformed. So listen and disagree and be friends.
Frankly there were a number of replies to your posts that clearly didn't read it, or understand it "did you read the post" that I would have considered handing out bans. At what point is "but you didn't consider X" when you have an explicit paragraph addressing X a violation of two if the gates (true, useful). If they aren't going to read your stuff, why should we read theirs. Your restraint is nothing but admirable.
The fair conclusion I've seen is that violent crime is down or steady, but only because people have taken costly steps to reduce their vulnerability to it that they wouldn't have needed to take some years ago. Going from ten robberies a year in a town where everyone leaves their doors unlocked to nine robberies a year in a town that would have one hundred robberies a year if people still left their doors unlocked isn't a reduction in crime by any meaningful understanding of the term.
I think there's some talking past each other on this issue.
Comments: "I experience an unacceptable level of crime and disorder in my daily life, and I'm upset about it!"
Post: "Statistics show crime is down. The things you are talking about are not really crimes."
Comments: The statistics must be lying.
Post: The commenters are being disingenuous and invoking logical fallacies.
The problem is people don't like being told crime is down because it appears to diminish the importance of an issue they feel is highly relevant.
In the past a higher level of crime may not have been as salient an issue as it is now because there were other issues of greater importance. In the present time, many other problems are not too bad so the highest priority issue that people encounter is "crime" even though there is on an absolute level less crime than in the past.
So, crime is down, but people are justifiably unhappy with the residual crime they encounter and want it to be a high priority public issue until the crime is further reduced.
<The problem is people don't like being told crime is down because it appears to diminish the importance of an issue they feel is highly relevant.
I agree that's the dynamic, but I think that both Scott and other commenters should call out people who are slipping into some gray non-rational zone where disorder is sorta a crime and they are infuriated by the dogshit in the park and anything that infuriating has to be a crime, therefore we should not be talking about why crime is down, but about dogshit, which is also a crime because they hate it, and also because DOGSHIT, QED
Here's Shankar calling out somebody for that kind of thinking in another context:
Person A: My opinion is that most cultures will think it [exploitive sexual misbehavior by the wealthy] is not a big deal
Person B:Considering there have already been real consequences for high-ranking figures like former Prince Andrew and members of the British government, that prediction has already resolved to false. Also it's not just about the sex, it's also about Epstein's Russia and intelligence agency connections.
Shankar Sivarajan: You're being disingenuous. The rest of the comment makes clear A’s talking about the differences in sexual mores.
I understand what you are getting at, but I think part of the disagreement is your example. You say dog shit in the park is not a crime. Which is technically true, but a distraction in that it is actually against the law to not pick up your dog's shit. I don't think it's really a satisfying argument to base your argument on whether the unlawfulness is handled by a civil or a criminal court.
I would like to see the law against dogshit enforced. I don't care whether the violators go to jail, get flogged, or pay a fine as long as *something* happens to them to dissuade them from this behavior.
The sexual stuff you mention seems to me an entirely different cultural issue. I don't accept the analogy.
Apologies if this point has been made already, but how would you (or anybody) know if crime is up or down if the reporting/recording entities are untrustworthy? How would we know the level of unreported crime? How would we know the level of reported crime that is unrecorded by the police? Why are you so sure the streetlight is operating at all?
This is why a lot of folks focus on homicides and auto theft as proxies for violent and non-violent crime trends. Both aren't likely to see a lot of unreported crimes.
This isn't without problems, but ...
Good point. Those two metrics are sure to be reported. But as you know, that leaves out so much of what constitutes what we mean by "crime".
I'm unsure what the relative levels are right now, nationally.
I find myself wondering if "Poor people are moving into my neighborhood" has anything to do with it.
Prioritizing which datapoints to use for an analysis *is* choosing which datapoints are directionally correct. Pointing out an analysis has conflicting data points is insufficient to determine if an analysis uses motivated reasoning, you have to show the conflicting data points are better proxies for the intended effect than the initial analysis.
My point is that it's significant that people talk broadly about crime rates and exclude types of crimes that cause significant amounts of societal harm.
both failure modes you describe are really motivated measurement: you choose what to count after you already know what conclusion you want. "directional correctness" is the more socially acceptable version because it lets you weaponize genuine victim status as a shield against precision. the fix probably isn't better fact-checking but declaring your priors before you pick your metric.
I don't use the term directionally correct this way and have not heard it used this way? I use it more in a context of high epistemic uncertainty or where there's a bunch of highly non-ideal options only. "Directionally correct" in this context means "least bad."
If I'm trying to get to New York from Houston and the only offers of rides I have are to A) Dallas, b) San Antonio, C) El Paso, none of those will get me to New York, but Dallas is the most directionally correct. It's at least going in the sorta correct direction to get to New York as opposed to basically the complete opposite way like the others. I will be a bit closer to New York if I take the ride to Dallas.
A strong rebuttal:
https://www.adorableandharmless.com/p/unsubscribe-from-the-church-of-graphs