Crémieux Profile picture
Aug 29, 2025 19 tweets 8 min read Read on X
Crime is way down in D.C.

Is it because the National Guard is arresting tons of people, or something else?

While there have been a lot of arrests, crime is down too much for that to be all.

Let me tell you about one of my favorite crime papers. It's about police presence🧵Image
In 2010, the British government issued a report. The report held that there was far too much unnecessary spending going on in policing.

As a result, London's Metropolitan Police saw a 29% budget cut.

To save money, the city shut down 70% of its police stations. Image
The mayor's office worked to shut down police stations without reducing the number of frontline officers they employed.

They tried to make sure the remaining stations would be equally distributed around the city, so that police could plausibly still cover everything.Image
This change made it so that different areas of the city were different distances from a police station.

If you look at the distances before and after the shutdown, the change is oftentimes remarkable. Image
But c'est la vie.

Sacrificed had to be made to ensure the city's budget didn't run afoul of the law and its fiscal base.

When it comes to keeping frontline officers out and about, the mayor's office at least managed to do that. But they did cut down on admin! Image
So far so good?

Just as long as the police can still feasibly do their jobs, you should still get all the benefits of policing—or at least, that's what they thought.

Comparing census blocks where stations remained open to those where they closed, violent crime spiked overnight. Image
Violent crime went up ~11% in areas where stations were shut down.

In fact, the closer an area was to a police station that got shut down, the greater the increase in crime.

Those are the areas that now had the fewest police; criminals were responding to mere officer presence!Image
But remember how they didn't fire any frontline officers? That means they redistributed them to the remaining stations.

We can use that fact to see a direct corollary to what's happening in D.C. right now. At the remaining stations, crime went down a lot.Image
Unfortunately, with police being more distant, that means more crime and less effective policing when they finally do manage to make their way over to calls.

Clearance rates fell by about 0.7%, and they fell most in locations police disappeared from the most.Image
Though violent crime went up in total, and officers became less effective, was it worth it? There's a dollar—or pound sterling—value to criminal victimization, so we can do the calculations, and...

Not worth it.

Cost-benefit calculations suggest each £1 of saved cost £3-7. Image
There are, by now, tons of studies like this, and they tend to come to the same conclusions.

Namely, that police being in an area deters crime.

But generally America doesn't seem to get this. Compared to peer countries, America is extremely underpoliced.Image
America has far more prisoners per capita than nations like Australia, Portugal, and Germany, but it has a much lower number of police per capita.

If America caught up and police exerted the effects the literature suggests, America could have lower crime and fewer prisoners.Image
We're seeing a microcosm of that right now, in D.C.

Crime is down compared to the same time last year, and it's down so far that it's hard to come to any other conclusion.

We can provide further support for the idea the Guard are deterring crimes by a paper on D.C. Image
Terror alerts are not the sort of thing normal day-to-day criminals think about

In fact, most crimes are 'in the moment', and they occur without any sort of premeditation, just because criminals are so liable to blow up at a moment's notice

So, why does the terror level matter? Image
It matters because when the level is higher, more police are deployed in Washington D.C.

On days where the city is on high alert and more police are out, the number of crimes that happen falls considerably, by about 6.6%. Image
But high alert days don't distribute police evenly throughout the District.

They're primarily concentrated on District 1—the National Mall.

As it happens, that's where about half of the high alert crime reduction happens. Crime on the Mall falls by 15% during high alert! Image
It's clear why crime is way down in D.C., and why it will continue to be down compared to recent years, so long as the surge is ongoing.

It's because criminals fear police. They act a little less hastily and fewer people die as a result.

That's the power of police presence. Image
If you want lower crime, you don't need to go all-in; you don't need a bunch of National Guard running around.

You just need to Fund the Police: increase the number of cops!

Or do what @growing_daniel does and effectively increase cop numbers by cutting their paperwork in half.Image

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Crémieux

Crémieux Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @cremieuxrecueil

May 24
Are White women the primary beneficiaries of affirmative action?

That's a real claim that's commonly advanced by journalists, and the claim has gone so far that it's even made its way into academic publications and policy.

But the claim is completely false🧵 Image
This claim doesn't make a lot of sense. After all, shouldn't the primary beneficiaries of affirmative action be the people who the policies primarily target?

In America, that's African Americans and, among them, women get an added benefit. How could it be Whites? Image
To figure out where the claim comes from, I started reading supposed sources.

Often enough, journalists will just take the claim for granted without providing *any* source.

It's just tacit knowledge now, and that's not good!

Then, when you hit a source, it's not supportive: Image
Read 13 tweets
May 7
World War I devastated Britain and likely slowed down its technological progress🧵

The reason being, the youth are the engine of innovation.

Areas that saw more deaths saw larger declines in patenting in the years following the war. Image
To figure out the innovation effects of losing a large portion of a generation's young men who were just coming into the primes of their lives, the authors needed four pieces of data.

The first were the numbers and pre-war locations of soldiers who died. Image
The next components were the numbers and locations of patent filings.

If you look at both graphs, you see obvious total population effects. So, areas must be normalized. Image
Read 12 tweets
May 5
New Pangram validation!

You know how most books on Amazon are AI slop now? If you didn't, look at the publication numbers.

Compare those to the proportion Pangram flags as AI-generated. It's fully aligned with the implied numbers based on the rise over 2022 publication levels! Image
Similarly, the rise of pro se litigants has come with a rise in case filings detected as being AI-generated, and with virtually zero false-positives before AI was around.

You can also see the rise of AI-generated text and yet more evidence for Pangram's validity from looking at different journalists.

Large portions of the journalistic profession are lazy, so they cheat when they can.

For example, the Guardian's Bryan Graham = slop Image
Read 9 tweets
May 3
Pierre Guillaume Frédéric le Play argued that France's early fertility decline was driven by its inheritance reforms, where estates had to be split up equally to all of the kids, including the girls.

There's likely something to this!🧵 Image
For reference, the French Revolution ushered in a number of egalitarian laws.

A major example of these had to do with inheritance, and in particular with partibility.

In some areas of France, there was partible inheritance, and in others, it was impartible. Image
Partible inheritance refers to inheritance spread among all of a person's heirs, sometimes including girls, sometimes not.

Impartible inheritance on the other hands refers to the situation where the head of an estate can nominate a particular heir to get all or a select portion. Image
Read 11 tweets
May 1
In terms of their employment, religion, and sex, people who joined the Nazi party started off incredibly distinct from the people in their communities.

It's only near the end of WWII when they started resembling everyday Germans. Image
Early on, a lot of this dissimilarity is due to hysteresis.

Even as the party was growing, people were selectively recruited because they were often recruited by their out-of-place friends, and they were themselves out-of-place.

It took huge growth to break that. Image
And you can see the decline of fervor based on the decline of Nazi imagery in people's portraits.

And while this is observed by-and-large, it's not observed among the SS, who had a consistently higher rate of symbolic fanaticism. Image
Read 5 tweets
Apr 24
"Food deserts" are an example of social scientists getting causality backwards

They saw poor people eating unhealthy foods and blamed local supply

They should have blamed demand!

Using data from 13 years of supermarket entries, there's basically no effects on healthy eating🧵 Image
The significant effects are probably not meaningful. They're more likely under the null with this gigantic dataset (p's of 0.003 and 0.005 with a total sample size of ~2.9m)

Entry did affect sales for new stores, but not existing ones. It also affected more local places more. Image
When new supermarkets open up, they do nab a share of local grocery sales, but the effect on healthy eating in total, among low-income households, and in food deserts, just isn't there. Image
Read 8 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(