Patrick Schneider (@patricksecon) profile picture

Macroeconomist, Assistant Professor at Imperial Business School Formerly @LSE, @bankofengland, @Uniofoxford, & @UNSW

London
Joined January 2017
Photo by Patrick Schneider (@patricksecon)Photo by Patrick Schneider (@patricksecon)Photo by Patrick Schneider (@patricksecon)Photo by Patrick Schneider (@patricksecon)Photo by Patrick Schneider (@patricksecon)Photo by Patrick Schneider (@patricksecon)Photo by Patrick Schneider (@patricksecon)Photo by Patrick Schneider (@patricksecon)Photo by Patrick Schneider (@patricksecon)Photo by Patrick Schneider (@patricksecon)
Turns out Claude was a MS365 product all this time, who knew?
2
345
Patrick Schneider retweeted
Thrilled to see my study on cash transfers and women's employment featuring in @vox_dev! In the paper, I show that Bolsa Família, Brazil's main transfer programme, increases mother's employment, and the effect is complemented by better local services👇 voxdev.org/topic/social-prot…
1
10
2
22
1,068
Patrick Schneider retweeted
Who pays for AI? This was a lot of fun to work on with @pj_lambert and @Nik_Datta. A couple of results I found particularly interesting... 1/N
Who pays for AI? In a new working paper, we study this using banking data covering billions of transactions to see who pays for tools like ChatGPT, Claude, Gemini, etc. We find that 1-in-20 UK adults now pay for AI, representing a $1.4 billion market annually. (1/N)
1
6
1
17
3,182
Election swings look better as weather maps than dartboards. e.g. here's 2024 Pres ->2026 House swing, relative to national average Check out the interactive map at patrickmschneider.github.io/…
137
Patrick Schneider retweeted
New paper from Peter, @patricksecon and myself. Check out the website here: whopaysfor.ai/
Who pays for AI? In a new working paper, we study this using banking data covering billions of transactions to see who pays for tools like ChatGPT, Claude, Gemini, etc. We find that 1-in-20 UK adults now pay for AI, representing a $1.4 billion market annually. (1/N)
1
8
367
Who pays for AI? This was a lot of fun to work on with @pj_lambert and @Nik_Datta. A couple of results I found particularly interesting... 1/N
Who pays for AI? In a new working paper, we study this using banking data covering billions of transactions to see who pays for tools like ChatGPT, Claude, Gemini, etc. We find that 1-in-20 UK adults now pay for AI, representing a $1.4 billion market annually. (1/N)
1
6
1
17
3,182
The gender gap is narrowing: others have observed men seem to be more into AI than women. While we do see this in our data, we also see that the trend is slowing down. The ratio of male to female first-time adopters dropped from over 3 in 2023 to 1 in this year 2/N
1
1
182
Occupational exposure could be the main thing driving consumer subscriptions. The cuts by age, income, gender, and region that we find get a lot weaker once controlling for occupation 3/N
1
142
Patrick Schneider retweeted
Who pays for AI? In a new working paper, we study this using banking data covering billions of transactions to see who pays for tools like ChatGPT, Claude, Gemini, etc. We find that 1-in-20 UK adults now pay for AI, representing a $1.4 billion market annually. (1/N)
9
75
17
286
29,762
Another exhibit of AI driving the cost of public goods below the private benefit to the maker
Built my dashboard for the evening ... Starting from my model predictions (based on polls) as priors, it will update and rerun the model as county level returns start coming in, based on a factor-model for polling error. Also reports updated betting market numbers all under one roof. [Did a simulation of the 2022 evening, seemed to work pretty well.] I'll post some updates over the course of the evening, provided it doesn't crash out.
1
202
Patrick Schneider retweeted
OK folks, here is my public dashboard. Those who want to follow it, here it is available for free! tinyurl.com/Election2026ADPu…
2
11
1
60
19,024
Patrick Schneider retweeted
1/ After three years processing 19 million planning applications and 120 terabytes of documents, @edoardo_badii, @AmritaKulka and I are releasing the first findings from our Housing Pipeline dataset. Fewer distinct proposed homes. A much stronger rebound in 2025. 🧵
1
1
10
559
It's pretty interesting how AI is reducing costs of this kind of thing - way more public goods become viable purely for their private benefit.
I asked Claude how much it would have taken to develop Econvitals pre-coding agents. It's answer: a team of about 50 people working for three years with a budget of $20 million. Is this correct or too high? I'm not sure. econvitals.org/about/before-…
1
3
390
For my own part, I've made dashboards to help me quickly contextualise UK fiscal and monetary news: - Fiscal space: patrickmschneider.com/fiscal… - Monetary space: patrickmschneider.com/moneta…
1
4
98
Are they useful though, or are we filibustering ourselves and each other? Does it matter? I like the things I've made, you might too 🤷‍♂️
34
Patrick Schneider retweeted
Turns out that if everyone was paid the same, saved the same amount, made the same on their investments, and worked & lived the same amount of time… … people 65+ would have 74% of the wealth. Surprised? Welcome to Equal Land. adamnash.blog/2026/10/25/why…
19
19
4
190
22,264
Patrick Schneider retweeted
It is an extraordinary failure that the BBC has the resources to do genuine public interest journalism that is much less viable commercially (from foreign correspondents to long-form local reporting) but instead we get relentless lowest common denominator nonsense
25
134
5
1,348
42,954
Ok but what if you got humans to guess two-thirds of the average of an LLM random number? What would two thirds of that guess be?
This is the least surprising graph in the world if you understand how gradient descent works.
83
This is the least surprising graph in the world if you understand how gradient descent works.
548
1,269
481
22,723
3,820,248