Longreads + Open Thread

Longreads

• Ben Smith and Liz Hoffman in Semafor on why Margin Call is such a great finance movie. There's a counterintuitive nepo-baby angle to it, which is that the reason JC Chandor could make a movie that so ruthlessly portrayed how investment banks behave when they suddenly switch from thinking like agents to thinking like principals is that his dad worked at Merrill Lynch for 40 years and was forced to retire after losing out in political infighting. (Turned out to be great timing: this happened in 2007.) The movie itself almost didn't happen: Chandor had actually given up on making movies and had a good job interview with an out-of-town company; the company asked him to stick around for a few days ahead of another interview, and he used them to write the script.
• In bene dictio, reflections on competitive math and college admissions. A very good piece: sometimes the logic of a system keeps pushing it incrementally further in some direction that leaves it totally disconnected from what it's supposed to be accomplishing (think of coal power plants sticking around because renewables and nuclear get stuck in endless environmental review processes). College admissions departments are trying to find hard-to-fake signals of some combination of smarts and grit, and while math competitions do index on brains, they're also not quite fair if they require someone to do something novel on the fly—so what they really have to test for is knowledge of lots of mathematical tricks and the ability to apply them. This is, to be clear, a useful skill! It's worth at least trying to get decent at this. But at some point, the next incremental trick doesn't help much outside of the domain of winning math competitions.
• Drayton D'Silva writes in defense of Buy-Now-Pay-Later loans for fast food. Thinking about the social utility of finance is tricky, because it's often easier to come up with high-level heuristics than to explain how any one behavior really makes sense. On the other hand, if those specific cases arise from coherent rules, they're a good way to illustrate how those rules really work, and that's what this piece does nicely.
• Andy Masley on why your personal use of AI tools almost certainly has a negligible environmental impact. Almost no one has good intuitions about how much energy or water they use, and where that usage comes from, so stats about the water use of a query, or how many households' worth of energy a model uses, are mostly decorative statistics. Once you do compare them to real-world quantities, you get a lot of fun examples of why curtailing your personal use of generative AI tools is an absurd way to reduce overall energy consumption: 100 ChatGPT queries uses as much energy as running your air conditioner for 18 minutes; if you're worried about the water consumption instead, just shower for 2.5 seconds less and you're even! The piece is repetitive after a while, but that's fine: first, it's a lot less repetitive than claims about how bad ChatGPT is for the environment, which don't seem to have been adjusted down for the radical reduction in cost per token since that meme started, and which also don't seem to have been adjusted back up to account for reasoning models. And second, the repetition comes from giving lots of real-world examples of the relative energy use of household appliances, which is actually useful information to know.
• Ghaderi et al.: Pricing of Corporate Bonds: Evidence From a Century-Long Cross-Section. If you've looked at long-run data on the performance of different asset classes, you might have noticed that for some of them, we mostly don't have it. Equities go way back (1926 is to market historians what December 31st, 1969 is to Unix users: as good a date as any for the Beginning of Time). Treasury bonds we can track. But corporate bonds are hard. Fortunately, they're now a lot easier, thanks to a ton of manual work that went into this study, leading to a pretty complete monthly time series of bond-level performance going back to the late 19th century. This dataset lets us put some real numbers on things that we'd previously just approximated: yes, the Great Depression had the highest credit spreads in our now-longer recorded history. And yes, the financial crisis came in second. It's also interesting to look at the count of corporate bonds outstanding as a proxy for the financialization of the economy: eyeballing the chart, the number of bonds didn't recover its depression-era peak until some time in the 1970s.
• This week in Capital Gains, a look at the favor-trading economy, and why some transactions happen entirely through free referrals even though fairly similar ones are often done through paid intermediaries.
• April's episode of FinTwit Book Club is out: we discussed Clashing Over Commerce, a history of US tariff policy.

Books

The Chosen: The Hidden History of Admission and Exclusion at Harvard, Yale, and Princeton: kicking this one off with a confession, I've been citing this Malcolm Gladwell review of the book off and on for two decades, and every time I did that I told myself that at some point I really ought to read the book. It was a good decision: this is a great history book, both as a narrowly-scoped look at the specific question of how Harvard, Yale, and Princeton have changed their admissions policies over a century, and as a more meta look at how institutions and elites evolve.

All of these schools are quite old, and it's tempting to draw a straight line between their founding in the 17th and 18th centuries and their status today. So it comes as a bit of a surprise to see that, around the turn of the century, they were mostly focused on football and partying. Harvard paid its head coach more than any professor, Yale had the largest football stadium in the country in 1914, and a fictional Yale football star, Frank Merriwell, was the subject of hundreds of books. (Later on, when they were getting more academically rigorous, Wilbur Bender worries that Harvard will go the way of U. Chicago—a once-proud football powerhouse now relegated to a focus on mere academics.) Princeton was not such a haven for jocks, but instead had a social scene that revolved around selective eating clubs. Each school had a tradition as being more of a religious institution—Woodrow Wilson was the first Princeton president not to also be a Presbyterian minister (in his defense, his father was quite a prominent one, and Yale required students to attend church service daily until 1926.

All of the schools recognized that they'd drifted from their educational mission and towards being an exclusive club for sons of rich men who might, when the mood struck them, periodically show up to a lecture or crack open a textbook. So, in the early twentieth century, they started opening up admissions a bit more, dropping Greek and Latin from their entrance exams and adding more questions that a bright public school student might reasonably expect to be able to answer. 

As it turns out, this led to what the elite schools perceived as a grave problem: the sons (and grandsons, and great-grandsons, and so on) of respectable WASPs were going to school with—Jews! Specifically, a large number of Eastern European Jews, who started migrating in larger numbers in the late 19th and early 20th centuries. These families were presumably sick and tired of a socioeconomic arrangement that automatically granted them lower social status and limited mobility, and pushed their kids towards academic achievement as the fastest way to advance in a more open society. The elites who ran that society were pretty alarmed by the prospect, and it was in response to this issue that top schools started using a more holistic approach instead of looking purely at academic standards (and, realistically, looking at the family tree to see if there were any alumni). They practiced varying degrees of honesty about this, both with the public and with each other. Yale, for example, imposed a quieter quota, while Harvard got in trouble with Boston-area political figures who worried, not about whether racial preferences were good or bad, but about whether they might end up applying to the Irish instead. 

And yet, the schools have muddled through. These quotas were gradually relaxed; postwar America considered antisemitism quite a bit less genteel than before. Just in time! They survived another period of upheaval and of negotiating changes in student demographics in the 60s and 70s, and, if history is any guide, they'll mostly get through the current changes with their roles intact. There's one sense in which the job of these schools is radically different than it was back when they were academically-unchallenging places for the scions of the rich to hang out before getting some similarly relaxing job or perhaps deciding to do something with their lives. But in another sense, their function is unchanged, as a mechanism for identifying future members of the elite.

Who are the elites today? If you ask someone in the media who the most powerful and unaccountable people in the country are, they'll probably point to Big Tech. Ask Big Tech, and there's a good chance they'll point right back in the same direction—the NYT can set a narrative that will constrain tech companies' behavior even if it isn't actually true. DC obviously matters, as does finance. But schools remain an important filter for elite status, and are thus a plausible candidate for the center-of-mass of America’s elite. There's still a very high correlation between having a Harvard/Yale/Princeton background and having an important job. 

The book is a story of elites deciding what criteria for membership will preserve both their immediate status as an elite (keeping the number of newcomers who don't follow refined upper-class rules to a minimum) and their long-term status as a deserving elite (by admitting that there are smart, effective people who did not go to the trouble of having their fathers attend Yale or send them to the right boarding school). Given enough time, an elite that can't renew itself with new talent is going to die, and it's a death best described as something between suicide and negligent refusal to address a chronic illness. But that means that if you're part of an upper crust that's worth being part of, one of the terms of that deal is that your kids will not necessarily inherit your status, though even if they don't they'll live in a society with reasonably competent people in charge.

Open Thread

• Drop in any links or comments of interest to Diff readers.
• Schools are a pretty common way to decide who's in the elite (two thirds of the UK's PMs have graduated from one of two universities). What else works? In the US, what are the other avenues for legible elite status?

Diff Jobs

Companies in the Diff network are actively looking for talent. See a sampling of current open roles below:

• YC-backed startup using AI to transform how companies quantify and optimize engineering productivity is hiring formidable full-stack and AI engineers. Experience with React + Typescript, Go, or Python on the ML side a plus. (SF)
• An OpenAI backed startup that’s applying advanced reasoning techniques to reinvent investment analysis from first principles and build the IDE for financial research is looking for a data engineer with experience building robust data infrastructure and performant ETL pipelines that support intense analytical workloads. (NYC)
• A multi-stage, fintech focused investment firm is looking for an investment associate to support thematic opportunity identification, diligence, and execution. Investing experience OR high-growth operating/investment banking/consulting experience and demonstrated interest in fintech required. (NYC, London)
• A hyper-growth startup (10x growth in 9 months) that’s turning customers’ sales and marketing data into revenue is looking for a head of deployments who is excited to work closely with customers to make the product work for them. Experience as a forward deployed engineer and leading enterprise deployments preferred. (SF, NYC)
• Well-funded, fast-moving team is looking for a full-stack engineer to help build the best AI powered video editor for marketers. Tackle advanced media pipelines, LLM applications, and more. TypeScript/React expertise required. (Austin, Remote)

Even if you don't see an exact match for your skills and interests right now, we're happy to talk early so we can let you know if a good opportunity comes up.

If you’re at a company that's looking for talent, we should talk! Diff Jobs works with companies across fintech, hard tech, consumer software, enterprise software, and other areas—any company where finding unusually effective people is a top priority.