Muhammad Usman Naeem (@usmann1504) header
Muhammad Usman Naeem (@usmann1504) profile picture

2026–27 Economics Job Market PhD Candidate @TuftsEconomics & @FletcherSchool Environmental, Development, & Behavioural Economics

Boston, MA
Joined November 2021
Photo by Muhammad Usman Naeem (@usmann1504)Photo by Muhammad Usman Naeem (@usmann1504)Photo by Muhammad Usman Naeem (@usmann1504)
Muhammad Usman Naeem retweeted
Why do people fail to conserve when doing so saves them money? Meet Tufts job-market candidate Usman Naeem (@usmann1504). In a field experiment with 6,621 vehicles, personalised fuel-cost information cuts idling 9%—yet a real-time reminder raises it. usman-naeem.com/papers/Naeem…
5
1
6
492
Muhammad Usman Naeem retweeted
I was just walking my dog, John Nash, in our quad at the University of Chicago and I ran into a student. Nice chap. I asked him his major and he said “Physics.” I then, of course, asked if he ever thought about an econ major. He basically scoffed, and said "economists aren’t real scientists, they don’t even have laws.” I laughed, ended with a quip, and John Nash and I were on our way. In economics, I tend to think we have something harder: conditional laws. When I jump out of a tree, gravity doesn’t care which country or at what time of day I jump, it has a wonderful prediction of my velocity when I touch down. Our regularities, on the other hand, care a great deal about who’s doing the jumping and where and when that jump occurs. Take the endowment effect. This is about as famous as behavioral “laws” get, and was advanced by Danny Kahneman, Dick Thaler, Jack Knetsch, and many others. In my sportscard field experiments, I found that it more or less vanished for experienced traders. I viewed that as a law with really interesting fine print: it is true for some people, in certain situations, and at some doses. This brings us to the key point. There are two ways we experimentalists can respond to such fine print. We can sulk, and give up by despairing: “nothing generalizes, so why bother?” I prefer a different route. Make the conditions the object of study and be a scientist to understand them. As I discussed in my recent Nature paper, from the very beginning you should go beyond A/B testing and use “Option C Thinking”. For example, backward induct from where the idea will ultimately live. Conceptualize which people, which situations, and at what dosage. Then, generate the scientific evidence under those conditions from day one. I call this creating policy-based evidence. So, what was my quip as I left the Physics student? “Interesting point, your particles don’t read the newspaper, mine do. But that is why beyond measuring we social scientists must also read the room. In fact, mapping the conditions then becomes a key addition to our science. Because of that, I tend to think of economics as the harder science.” He just smiled and nodded. No idea if I convinced him, but the two papers (the endowment effect field experiment and the Nature piece) that I mentioned above are in the first comment if you are open to being convinced about economics as a (harder) science!
23
46
11
353
41,882
Muhammad Usman Naeem retweeted
Economists: Now that LLMs can do desk research, the relative importance of RCTs should go up. LLMs:
New working paper: Can AI forecast treatment effects in economics from pre-registration documents? Answer: Yes. 1/
3
24
1
206
13,805
Muhammad Usman Naeem retweeted
The economist job market is up and running!  Last year the economics job market was the toughest in recent memory, as JOE listings were down 20%.  This was below even Covid levels.   A bit of good news for PhD economists this year in that I am noticing some green shoots in my discussions with employers!  And, in the data, we have a bit of good news as well: JOE listings are up 5.5% over this time last year (see below). Under the hood is where it gets interesting, however: **Full-time US academic jobs: +14.4%. Liberal arts colleges +16.1%, universities with graduate programs +13.8%. The segment that took the biggest hit last year is leading the recovery. **Full-time academic jobs outside the US: -12%. The rebound so far seems to be a domestic story. **Nonacademic jobs: +7.5% overall.  Yet, this hides a ton of heterogeneity. Consulting/research +21.6%, Banking/finance/industry -25%, US federal government: -44.4%. A side note is that for my entire career, the federal agencies were a stabilizer of this market, the last few years we have seen this buffer disappear. At this point, a candidate might ask: how should I respond to all this news?  Considering the data and my chats with people around several markets, I would still turn to the science.  Back when I was on the market it was difficult too (at least for me…I had 1 interview from 150 applications!).  The next year we studied the search market (paper downloadable here: ideas.repec.org/p/feb/natura…). We compared a boom market to a bust market through the lens of optimal search across multiple markets. Candidates adjust on two margins: which markets they enter, and how many applications they send into each. In busts, we found graduates of elite programs broadened their search down the academic ladder, graduates of lower-ranked programs shifted toward the business sector, and employers everywhere raised their hiring standards. So far the data in 2026-27 is neither a boom nor the bust in our paper, it is both at once.  Today academia is loosening while government has effectively closed and finance has tightened.  Even with those changes, my advice is no different than it was back then: adjust both margins.  This means that you should re-enter markets you may have written off.  This is because the academic window is wider than many feared, especially at liberal arts colleges. In addition, perhaps redirect applications from DC toward consulting. In the end, the deepest lesson for me remains:  there is no “the market,” we live in the world of “many markets.”  Treat this like a portfolio problem, and thus the key question of the day is “how are YOUR markets doing?”
2
48
2
230
25,197
Muhammad Usman Naeem retweeted
Are we consuming too much groundwater? 💧 Dynamic models for most of the world’s 3,400 aquifers say 61–82% of aquifers are pumped faster than is optimal, costing trillions of $. New paper w BGreaney & KWagner joseph-s-shapiro.com/researc… 1/🧵
4
23
2
63
3,903
Muhammad Usman Naeem retweeted
… and the horror show is not that AI “solved DiD” Instead, NOT A SINGLE PAPER was deemed to be without errors:
3
4
2
31
6,485
Muhammad Usman Naeem retweeted
🚨 The tool now also covers Political Science, Accounting, Finance, Marketing, Management, Strategy, Organizations, Information, Operations, and Psychology. 🔗 scholar-tools.com/ If you know academics in those fields, please help spread the word. 📣
🚨 About to submit an Econ paper? I just updated the Econ Journal Matchmaker: 🔗 scholar-tools.com/ I incorporated some of the most requested features. #EconTwitter Please help spread the word by reposting. 🙏 More details below 👇
1
21
124
14,148
Muhammad Usman Naeem retweeted
Oof
Replying to @JohnHolbein1
Source: "Do Personal AI Conversations Reduce Loneliness? Evidence from a Pre-Registered Experiment" ifo.de/DocDL/cesifo1_wp12882…
5
12
99
14,077
Muhammad Usman Naeem retweeted
More economists should focus on building harnesses instead of estimation packages. e.g. there should be a diff-in-diff harness which estimates, looks at the results, identifies common errors after seeing the results, and uses the right econometric tools to fix those errors.
5
25
6
219
33,113
Muhammad Usman Naeem retweeted
A new analysis in Science of more than 116,000 bird songs has revealed that the astonishing variety of passerine birdsong is built from just eight basic acoustic motifs, like trills, whistles, and harmonic notes. The findings also reveal that local environments greatly shape how birds communicate around the world. Learn more: scim.ag/4wYoVXE
32
1,035
71
3,201
266,615
Muhammad Usman Naeem retweeted
Fascinating
New @Google & @GoogleDeepMind analysis of 5 million interactions across the Gemini App, Google AI Mode, and Gemini API to map AI usage to over 800 occupations, 4000 tasks, 300 household activities, 150 countries, and 140 languages. ai.google/static/documents/G…
2
1
9
5,692
Muhammad Usman Naeem retweeted
Project APE update! 🚨 "Verifying the Verifiers: Towards Autonomous Policy Evaluation", new paper with @OlafWillner Reminder: Project APE is an experiment in building a system that learns how to do policy evaluation autonomously. The potential upside is that we could then scale up our understanding of what policies to, cheaply and reliably in many countries. Premise for this paper: With *generation* of plausible-looking research becoming cheap, *verification* is a key bottleneck. But can we automate verification, cheaply and reliably? We need to verify the verifiers, somehow. A thread on what we did so far... 👇
6
32
11
120
29,468
Muhammad Usman Naeem retweeted
💧New Paper Water Works: Causes and Consequences of Safe Drinking Water in America 📉US drinking water pollution fell by HALF in 2003-2019 💰Safe Drinking Water Act $ helped ❤️Investments reduce mortality, benefit/cost ratio >20 w @DavidKeiserEcon BMazumder @davidmolitor 👇
3
22
4
52
6,495
Muhammad Usman Naeem retweeted
Petrol at Rs 30/litre in Pakistan sounds crazy. It is not. What is crazy is the policy failure that prevents it. Petrol is around Rs 300/litre today, excluding government levy, here's how it can effectively be Rs 30/litre. People do not consume petrol for its own sake; they use it to travel. The average Pakistani rides a motorbike. A fuel-efficient motorbike can travel about 60 km on one litre. An efficient electric scooter can travel about 30 km per kWh, so it needs only 2 kWh to cover the same 60 km. What should 2 kWh cost in Pakistan? Pakistan is one of the best places in the world for solar, with an all-in LCOE cost of around 5 cents per kWh. The electricity cost is 10 cents, or Rs 30/litre of distance travelled! The Rs 30/litre calculation remains the same for cars. The 300-versus-30 gap is the cost of bad policy. It reflects billions of dollars of saving that could instead finance EV infrastructure: charging, distribution, battery swapping, and smart pricing software etc. - boosting much-needed domestic investment. Since solar is highly modular. You do not need massive scale to get reasonable efficiency. That creates business and employment opportunities for small domestic power producers. Instead, Pakistan leaned into large fossil-fuel plants financed by dollar-denominated borrowing and guaranteed returns. Local firms face credit constraints, but solar creates a natural collateralizable cash flow through electricity sales to the grid. With the right regulatory framework, this could have unlocked large private domestic investment, and employment. Battery swapping is another area where small local businesses could have emerged and scaled. Electricity enables smart pricing. When solar supply is abundant, prices can fall, and poor households and firms can shift usage to cheaper hours - automatic demand stabilization Better air quality would mean longer, healthier lives and higher productivity. That is a growth multiplier Green technology industries could be developed domestically with the right industrial policy, easing balance-of-payments pressure while raising employment and investment. Instead, Pakistan chose imported-fuel power plants, protected a backward-looking domestic auto sector, and raised electricity prices by burdening them with the fixed costs of those plants and heavy taxation, slowing EV adoption. Then came the net-metering fiasco, all to keep zombie power plants alive. Pakistan’s energy policy may be the clearest example of a broken nervous system. I hope someone fixes it, because people are paying the price, 300-versus-30
Should petrol be 30 rupees a litre in Pakistan?
58%Yes
42%No
1,346 votes • Final results
168
879
71
2,899
666,433
Muhammad Usman Naeem retweeted
Advice for PhD students in economics about using AI, from the brilliant Isaiah Andrews. This should probably be circulated to all PhD cohorts economics.mit.edu/sites/defa…
19
465
39
2,035
555,879
Muhammad Usman Naeem retweeted
AI tools have made research incredibly exciting. J. C. R. Licklider called this period the "Man-Computer Symbiosis" and predicted they would be the most intellectually exciting in the history of mankind. But they won't last long... Here's the excerpt from his 1960 (!!!) essay. Sounds pretty familiar. h/t @glenweyl groups.csail.mit.edu/medg/pe…
2
10
1
63
4,808
Muhammad Usman Naeem retweeted
“Will there be a top scientific journal where the peer review will be done by an algorithm?” I asked this to (the always awesome) @m_sendhil back in 2024… made some excellent points worth keeping in mind as we move forward. (See convo towards the end, at 1h:22min. We also talked about the role of academia more broadly, as AI progresses, before that) youtu.be/z_svj3NP968?t=4932&…
2
8
36
7,422