Kris of Party at the Moontower on how to know if you're under-exploiting an edge. This is a great combination of a trading case study—in this case, making a market in oil options—with the general theory behind when and how you should hedge. As with other trading examples, it has broad applications: there are many domains where experiencing too few failures indicates insufficient risk, and where the optimal amount of variance involves a bit of pain here and there. (This post also drew my attention to The Laws of Trading, reviewed below, which is well worth reading.)
Quantpedia: What Can We Learn from Insider Trading in the 18th Century: well-informed insiders have often had an edge, though that edge leaks out a bit when they trade. (There's still value in looking at things like insider buying, and changes in management's incentive compensation.) It's notable that some traits of markets have persisted for literally centuries, though actually exploiting them is always a challenge.
Molly Fischer at NYMagprofiles economist Adam Tooze. Tooze's books are indeed excellent, and they clearly show a love of diving into the source material. There is sometimes a proof-of-work aspect to academic fame, where novel conclusions are much more powerful if they're bolstered by obscure footnotes. This does not just apply to economics and history; here's an example from number theory.
Auren Hoffman of Safegraph on the moral obligation to make data more accessible. Much of this post is on the mechanics of making data more broadly available while still maintaining user privacy—if you disagree with the headline, it's an especially worthwhile read! More widely available data is not a panacea, because more datapoints means more ways there to selectively draw trendlines through them, but data availability does make the right decisions easier to determine. So granting that privacy concerns can be addressed, the utility of more data availability hinges on the question of whether there's more demand for solutions or for reinforced narratives.
And on the topic of data: Jason Briggeman at Econ Journal Watchquestions the literature on long-run economic persistence, a topic that's come up in The Diff from time to time. Much of this piece addresses statistical questions of artificially selected and dropped datapoints, and arbitrary ways to set up controls. One of the broader questions it raises is how meaningful statistical theories of economic persistence are if they go back that far. Taking an example from the books below: Japan in the 1930s had a low standard of living, but fairly high economic output because military spending was so elevated. This is perfectly compatible with an inflection theory ("Japan was poor until disarmament, containerization, and industrial policy") and with a mean-reversion theory ("Japan has produced an exceptional amount of wealth considering its limited natural resource endowment throughout human history, and as global wealth and trade increased, this showed up in GDP per capita rather than in population per acre of arable land.")
Books
Why Japan Was Strong: Published in 1943, this book tells the story of the author's extended budget vacation through Japan and China in the 1930s. Perhaps due to wartime shortages, the publisher was unable to afford a sensitivity reader. It does have some very worthwhile observations. In particular, the author is obsessed with economics (as it turns out, he was a libertarian gadfly for decades in addition to many other things, which explains his periodic digressions on topics like his dislike of unions and disapproval of the Jones Act). The book is a remarkable portrait of an economy that has just risen above the subsistence level, where labor is incredibly cheap and everything else is quite expensive. Among other things, cheap labor means that the government can afford lots of spies—this is the second book I've read in which an American traveler visits pre-WW2 Japan and is followed around nonstop by not-especially-secret police (the other is this one). Presciently, he argues several times that conflict between the US and Japan is not inevitable as long as Japan is allowed to get rich.
The Laws of Trading: Extremely good book by a former Jane Street trader on how to apply trading knowledge to everyday situations. The book mostly focuses on defining principles outside of trading, but offers some excellent anecdotes on trading itself, and in particular on the mismatch between what strategies look like in a backtest and what they look like in real life. One useful instance of this: suppose a backtest shows that a strategy makes money 55% of the time, but in live trading it loses money—but a backtest over the same period still shows profits! What explains this? One answer is that of the 55% of trades that are worth doing, a trader will get fewer than 100% of them because other traders use the same signals better. But of the unfortunate 45%, an unlucky trader can get them all. So the meta argument is that it's trivial to model an agent who is doing what you're doing, but worse, and effectively impossible to model the situation where you're the subpar trader and someone else has a smarter version of your own strategy. This kind of adverse selection can show up in any competitive domain, where systematic approaches work well unless an adversary is prepared for them.
Open Thread
Drop in any links or comments of interest to Diff readers.
It's come up in a few links today that trading is a skill that can generalize to other roles. Which other skills fit. Programming, poker, and copywriting are good candidates; for some projects a background in philosophy can be surprisingly helpful. What else?
A mid-frequency systematic investment fund is looking for quantitative researchers and quantitative developers. (NYC)
A crypto incubator is looking for a Head of Security Research to support their portfolio companies. (US, remote)
An alt data consultancy that works primarily with hedge funds is looking for a data scientist. (NYC)
A company helping the mid-sized creators make a full-time living through social commerce is looking for a Senior Data Analyst with Looker experience to join them. (US, remote)
A company helping students and early career professionals navigate education options is looking for an experienced Chief of Staff. (US, remote)
A company which offers retail investors the opportunity to create their own systematic strategies is looking for a General Counsel. (US, remote)
All these companies listed, and 20+ more, are looking for people in all kinds of roles. Please reach out if you'd be interested in working for a company that's in a Diff-adjacent field.