Longreads

• Homevoters and the haut precariat: Noah Smith has an interesting analysis dividing class into the people who will probably eventually be able to buy a home, and people who fear they won't. (Class wars seem to follow a Hotelling's Law pattern, where the struggle is between two groups that are roughly equally influential, but whose exact power is a bit blurry.

• An interview with Mark Zuckerberg on the metaverse: Mobile is not the last platform pivot Facebook will have to make, but as Zuckerberg notes in the interview, this time the company is big enough to decide how the next platform should work.

• Applied Divinity Studies proposes the doped Olympics. If drug testing is good but imperfect, the Olympics are partly an athletic competition and partly an expensive medical one anyway. (Perhaps the best semi-controlled experiment here is that since reunification, Germany has yet to match the number of Olympic gold medals earned by East Germany in 1980 and 1988.)

• Why don't rich people do more patronage? One reason is reputational risk: a good target for patronage is someone who is much less famous than the person funding them, but that means the funder faces downside depending on who they associate with. This is a broad explanation for why rich people generally don't do the things people think they'd do if they were that rich: the monetary downside is manageable, but the social downside is actually bigger.

• American Affairs has a great piece on Operation Warp Speed as an example of industrial policy. It ticks many of the right boxes for successful state-driven investment—a tight deadline, an existential threat, and the removal of most red tape. (Even within OWS, there was some red tape-cutting; Pfizer turned down funding for research, but accepted a preorder, because they didn't want to get slowed down.)

Books

• The City and Man: In the introduction to this book, Leo Strauss says that replacing political philosophy with the history of political philosophy means replacing "a doctrine which claims to be true by a survey of more or less brilliant errors," then proceeds to do approximately that—the book is a very careful reading of Aristotle, Plato, and Thucydides, with much attention paid to what the authors say, what their characters say, and what that says about what they really thought. Strauss is worth reading because his framework—interrogating the text, understanding the context, expecting double meanings and selection effects in what's left unsaid—is even more applicable in a world where there's an abundance of media and it's heavily filtered by algorithms or editorial decisions before it gets to you.

• Bits, Bytes, and Barrels: This book is partly a very long-form sales letter encouraging energy companies to build more software (the co-authors are consultants who work with energy companies), but this book is a good overview of how modern software is changing the oil and gas industry. The most important point is that many of the assets involved have long lives, and get built around cyclical peaks, so some parts of the industry are naturally a few decades behind; if a refinery runs roughly 24/7 for half a century, and shutdowns have to get scheduled a year in advance and planned down to the minute, that limits how many sensors, connected devices, and Smart Components can be added. For assets with a faster replacement cycle, there's a different problem: when prices are high and companies are flush, it's easier to have a narrow confidence interval for the expected return of conventional assets than for new kinds of investment. But the natural consequence of this is that as existing infrastructure reaches the end of its life, it will get replaced by much more efficient new versions.

• The First Junk Bond: the Platonic ideal of a business book—this is a book about a fairly middling oil company, Texas International, with the distinction of having issued the first original-issue junk bond ever. The company went bankrupt, but not before engaging in all sorts of interesting financial engineering, including: lots of debt swaps, acquisitions, divestitures, paying suppliers in stock, forming a joint venture with its investment bankers in which Texas International used the proceeds of junk bonds it issued to buy other junk bonds to which it could apply a tax-loss carryforward, and even issuing a bond whose maturity value was stated in barrels of oil rather than dollars. (No, they did not promise to physically deliver the oil, just to pay you whatever WTI closed at on a particular day.) What's great about this book is that it has complete financial statements for the entire period in question. It's been a bugaboo of mine for a long time that business books will give you some of the numbers, in text form, but they rarely give you everything; for growth companies, that's especially annoying—a million-dollar deal for Microsoft was very different in 1977 than in 1987, but if you want to know how different it was you have to Google around to find out what their annual revenue was in each historical period. (Distant Force is another business book that includes financial statements, although not at nearly the same level of detail as The First Junk Bond.) Because the company was so energetic in the pursuit of deals, the book has cameos from more famous companies—they bought assets from King Resources, a company tied to Investors Overseas Service, and later did an exploration joint venture with Enron. It's hard to measure this, but it seems anecdotally like the companies that generate lots of scandals are more likely to do business with one another. This could be a side effect of the complexity/books-cooking correlation: it's easier to fudge the results for a complicated business. But it's an interesting pattern regardless.

Open Thread

• Drop in any links or books of interest to Diff readers.

• I'm interested in cyclical phenomena, in markets, industries, politics, and culture. And that raises a meta question: what are some industries (or other things) that used to be cyclical and stopped, or that used to follow a secular trend and switched to a cycle instead?