Share this page!

Enter URL or ID to Unroll

You can paste full URL like: https://x.com/threadreaderapp/status/1644127596119195649
or just the ID like: 1644127596119195649

How to get URL link on X (Twitter) App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
Shengwu Li Profile picture
Shengwu Li
@ShengwuLi
Economist. @Harvard. Mechanism design, behavioral theory.
3 subscribers
Feb 10 • 8 tweets • 3 min read
Ok here are my go-to papers for making small talk with smart non-econs.

(Yes, I am exactly as uncool as maintaining this list would imply.) 1. Budish-Cramton-Shim on high frequency trading and frequent batch auctions. Specifically, for this robustness check concerning special relativity. Image
Dec 12, 2023 • 9 tweets • 2 min read
Why unrealistic theories?

Most econ students go through a phase of thinking that more 'realistic' theories are automatically better. But here are some good reasons for unrealism. 1. the obvious: a more realistic theory would be intractable.

But that doesn't mean you should make theories as 'realistic' as possible subject to a tractability constraint.
Sep 6, 2023 • 11 tweets • 2 min read
Having angered the humanities-econs, let me now anger their complement.

I think that every PhD student should learn about the history of economic thought. Either by reading primary texts or by reading the work of relevant historians. Economics made a lot of stark methodological choices in the mid to late 20th century.

(Expected utility theory, preference utilitarianism, methodological individualism, a focus on equilibrium, non-cooperative rather than cooperative game theory...)
Sep 4, 2023 • 8 tweets • 2 min read
Let me explain this claim, as if to an undergrad, an aspiring economist who has not yet taken a math course with real proofs. 🧵 You are planning to join a discipline that has asked (and answered) foundational questions using mathematical models.

These questions touch every corner of economics. They include:
Mar 2, 2023 • 9 tweets • 2 min read
Many families have vindictive relatives. My relatives just happen to control a small authoritarian government. 🇸🇬 🧵 My grandfather, Lee Kuan Yew (LKY), was the first prime minister of Singapore. My uncle Lee Hsien Loong (LHL), the current prime minister, disagreed with his brother and sister about Lee Kuan Yew's will.
Sep 18, 2022 • 4 tweets • 1 min read
Is it true that “chess is just poker now”? When human players are guided by chess engines, it may be that the real strategy is deciding what to prepare. And the preparation phase may well have an equilibrium in mixed strategies. Image Article here: theatlantic.com/technology/arc…
Jul 10, 2022 • 4 tweets • 1 min read
It has been 5 years since I left home, because of a political prosecution by the Singapore government. Friends often ask me if it's safe to return. The court case is technically over. However, I assess that there's a substantial risk that my uncle, the Prime Minister, would find an excuse to imprison me were I to return to Singapore. He likes to relitigate old disputes.
May 13, 2022 • 15 tweets • 2 min read
The list of Basic Presentation Habits That Make Me Judge You.

Thanks @jmwooldridge and @KhoaVuUmn for getting the ball rolling!

🧵 1/219 1. Wall-of-text lit reviews meant only to ward off the evil eye.
Mar 12, 2022 • 8 tweets • 2 min read
Let me offer a defense of complex models for everyday economic situations.

Of course workers aren't literally taking the integral. They're behaving based on a mixture of heuristics, common sense, and folk wisdom. 🧵 The worker has to (roughly) solve the problem for particular parameter values. The analyst who is interested in counterfactuals must solve the problem for general parameter values. In that sense, the analyst's exercise is harder.
Oct 23, 2021 • 10 tweets • 3 min read
On reading the ‘classics’: in my own work, I have often come up with a seemingly new idea, only to discover that it is an old idea that resisted formalization. 🧵 E.g. I wrote a theory about why dynamic clock procedures made the 2017 FCC spectrum auction ‘obviously strategy-proof’. Then I found this passage from Kagel, Harstad, & Levin (1987) about mistakes in 2nd price auctions.
Oct 9, 2021 • 5 tweets • 1 min read
If it’s hard for grad students to transition from coursework to idea generation, what can we do about it?

One idea: faculty could hand students a starter project.

That is, a well-formed open question that can likely be answered via standard methods.🧵 How will students learn to ask questions?

Well, asking good questions needs you to know what it’s like to answer them. It’s a bootstrapping problem.

Not clear the start of a project is the best place to start learning.
Aug 3, 2021 • 7 tweets • 1 min read
Coauthorship thresholds are often a source of misunderstanding between econ and other fields. Too often there’s a simplistic critique that econ should be more ‘generous’ with coauthorship. But this ignores equilibrium effects!🧵 Some fields give coauthorship for RA work or even just helpful suggestions. For good or ill, this dilutes the signal from coauthorship.

(Let’s not pretend that people carefully parse contribution statements in fields that produce 10 publications a year.)
Jul 25, 2021 • 4 tweets • 1 min read
When you see other people’s research, you mostly see successful projects. To reduce survivorship bias, here are some projects I’ve dropped after more than two weeks working on them: 1. A theory of choice from mental tournaments.

Why: No clear testable predictions.

2. A lab experiment about buying commitment devices.

Why: Pilot studies indicated treatment effect likely small or zero.
Jul 16, 2021 • 11 tweets • 3 min read
Econ perspective: People make choices to satisfy their preferences.

Psych perspective: People adapt their preferences to justify their choices.

Here's a theory that combines them. (A new working paper joint with Erik Eyster and Sarah Ridout.) arxiv.org/abs/2107.07491🧵 First, a parable from @R_Thaler (1980): Bob pays $100 for a ticket to a basketball game to be played 60 miles from his home. On the day of the game there is a snowstorm. He decides to go anyway. If the ticket had been free-of-charge, he would have stayed home.
Apr 28, 2021 • 5 tweets • 1 min read
Alright #econtwitter, let's do this properly. Please rank these voting methods from best to worst.

Plurality rule
Ranked choice voting
Borda count
Condorcet method

surveymonkey.com/r/5P78VJT

(link because Twitter polls do not allow for full preference lists.) I will aggregate responses, and report what each method would have selected.
Apr 23, 2021 • 8 tweets • 2 min read
Thoughts on learning to "fail fast" in research. When you have a new idea, and the project is worthwhile only if P, then you need to check P first, not last. Examples: 1. You can only establish causality with a certain kind of data. 🧵 2. Your theory is only interesting if a certain conjecture holds.
3.Your experiment is only convincing if you can rule out a certain confound.

When the relevant condition can be cheaply checked, this has to be the first thing you do, not the last.
Feb 24, 2021 • 8 tweets • 2 min read
There’s an econ paper by a Harvard law professor that uses game theory to deny that the Japanese military held sex slaves during WW2. Yes, it literally makes that claim. chwe.net/ramseyer/ramse… #EconTwitter The consensus among historians is that the Imperial Japanese military held tens of thousands of sex slaves, sometimes called “comfort women”, from all over East Asia. The modern Japanese government holds that this happened. The UN holds that this happened.
Feb 23, 2021 • 6 tweets • 1 min read
As someone who occasionally read letters for PhD admission, I give less weight to letters from pre-doc employers compared to letters from undergrad or grad advisors. Why? Pre-doc employers have different incentives: if their pre-docs don’t consistently place well into PhD programs, then they are less able to recruit in future. Of course they will not make an explicit quid pro quo. But potential employees will look!
Feb 21, 2021 • 5 tweets • 1 min read
IMO: If you are an econ theorist, and are regularly attempting more than 6 hours of research work a day, you are not on the research-leisure Pareto frontier. Real focus is physically draining, and creativity can't be forced. 🧵 By research work, I mean work that requires full concentration and mental acuity, e.g. writing proofs, learning new technical tools, learning other papers. Answering routine e-mail not included. I make no claims about empirical research.
Dec 13, 2020 • 5 tweets • 2 min read
New working paper! Consider the decision problem, "For this choice rule, does there exist a strategy-proof mechanism?" This is an easy problem; by the revelation principle, we just need to check that the direct mechanism is IC, which takes linear time. (thread) #EconTwitter Image Now consider the problem, "For this choice rule, does there exist an obviously strategy-proof (OSP) mechanism?" OSP depends on the extensive form, so the class of candidate mechanisms is combinatorially complex.
Aug 31, 2020 • 4 tweets • 2 min read
A lightboard-style setup without the lightboard. All you need is an ipad, a stylus, and OBS. Thanks @rithvikra0 for the idea! #EconTwitter 1. connect iPad to mac via USB-C.
2. open sketch software with dark background.
3. add ipad as video source in OBS.
4. add crop and chromakey filters, choose custom color equal to background.
5. tweak to taste.
6. overlay on webcam video source.
7. write!