Is Everything Getting Old?
Plus! Parity; Work from Home; Surgery; Buffett and Energy; Splinternet Update
In this issue:
Is Everything Getting Old?
Work from Home
Buffett and Energy
Is Everything Getting Old?
The Onion's Our Dumb Century is a classic satirical look at the twentieth century, of course, but it's also a nice tour through the American zeitgeist over that time. One of the headlines that hits a little harder than it used to is from 1985: "Dynamic New Soviet Leader Not on Brink of Death." In the early 80s, the USSR successively appointed Yuri Andropov (68 years old, died in office in a year and a half) and then Chernenko (who took power at the age of 72 and died after just over a year). But now the US Senate is the oldest it's ever been, the speaker of the House is 82, the party leaders in the Senate are 71 and 80, and the Presidency is held by someone who won at age 77, running against a 74-year-old.
It would be great to have a simple, pat explanation for this, maybe involving the increasing complexity of the political system, the importance of longstanding donor relationships rather than the energy necessary to campaign door-to-door and give rousing speeches, etc. But the phenomenon isn't just tied to politics: we're listening to older music, and watching older actors—in movies that are mostly revisions to existing franchises (Lord of the Rings has produced three big-budget multi-volume video franchises in my lifetime). From 2005 to 2019, the average age of CEOs at hire rose fourteen years. There must have been something really special about the class of '81 at every college.
So this might apply to fields where there's a public-facing franchise; whether it's music fans, moviegoers, voters, or shareholders, perhaps we're all fans of people we're very familiar with. But this trend is also happening in scientific research: in 1980, around 1% of NIH principal investigators who received grants were over 65, while 21% were 35 or under; now the numbers are 9% and just under 2%.
It's surprisingly hard to come up with a universal theory for what's happening here. But it's also important to understand, because it presents a few possibilities:
Something has broken in how we hire and promote people, and how we choose what kinds of entertainment to consume: we've gotten more fixated on what's well-known, which is narrowing our options and skewing them to whatever was trendy when they got popular.
Something was broken before, and has started working now; people retired before they should have, and companies and governments have been missing out on the talents of older workers.
There's been some fundamental change in the nature of aging, or of all of these roles, that either increases the relative importance of experience or reduces some of the costs of aging.
The last one is worth considering, because there is a health-based argument here. Consider this chart of life expectancy at age 45: even though most of the overall increase in life expectancy since 1900 has been from lower infant mortality, life expectancy at 45 went from around 70 in 1900 to 74 in 1950 and 81 today. And some of this is due to lifestyle changes: alcohol consumption per capita rose 76% from 1940 to 1980 and then declined, and US cigarette consumption per capita doubled over this time period and has since been in steady decline. Both habits will, for different reasons, make it less likely that people remain successful over long periods—alcohol as a social lubricant that can cause health and other problems might lead people to have broader social networks when they're younger (especially if alcohol consumption is common), but to run into problems when they're older.1 Nicotine, as a stimulant whose most popular method of administration is carcinogenic, is a way to burn the candle at both ends for a while, but not a good way to maximize your odds of still being at work in your 70s.
So the period since the early 80s might be a time when improvements in healthcare weren't being offset by worsening lifestyle choices, meaning fewer careers cut short.2 In that model, leadership turnover has been artificially high in the past, and it's finally getting to a more normal level—which is another way of saying that there's more competition for top jobs, and that the bar for quality (or luck) is higher.3 That's a fairly benign explanation, albeit an annoying one to anyone who isn't lucky enough to have started their career forty years ago.
But there are other possibilities. One is that aging in most fields is an instance of rising risk aversion. That, too, might be partly a result of a changing bias and partly a result of getting rid of a bias. The movie business produces lots of sequels, but Disney is not flailing around at random; they're trying to make what viewers will want to watch, and as it turns out that's a lot of Marvel and Star Wars. Big companies might have had a bias towards overweighting an executive's recent mistakes relative to their overall career, which would mean firing lots of people soon after they get enough responsibility to make a truly colossal mistake.
And there are definite advantages to older people. The predominant stereotype is that younger people are more flexible and older ones get set in their ways, but there's a very useful counterexample to that: in fields that skew young and grow fast, there's an interesting "sonic boom" effect, where a growing cohort in the industry has never been through difficult times. When those times hit, some people are unprepared, and the most adaptable people are often the older ones, who have seen a cycle or two before. You can try to fake this by reading history, but it's hard to make yourself inhabit the mindset of someone who went through difficulties you haven't personally faced. And what you'll often develop instead is a sort of reverse imposter syndrome: sure, that stockbroker's memoir of the 1930s makes it sound like a bad time, but he probably sensed that things would work out eventually, while you are going through a downturn that might never end. So having experienced people around—and in charge—can moderate some of the swings between mania and depression that someone would go through during their first difficult cycle.
Hiring for this is an instance of healthy risk aversion, since one classic risk is overreacting to near-term news in a catastrophic way. There were plenty of good excuses to keep 100% of your assets in cash in 2009 (the banks are still struggling, unemployment is still high!), 2010 (globally imbalanced recovery), 2011 (US debt downgrade!), 2012 (Europe!), 2013 (fiscal cliff!) etc. Some of the most thoughtful voices about the tech boom of the 2010s were people who remembered what 1999 was like and knew this wasn't it just yet. And politicians who were involved in serious decisions during the Cold War probably have a different attitude towards dealing with nuclear powers than the ones who didn't.
There are perks to getting older. You've had more time to learn, which means more ways to pattern-match; you've also had more opportunities to test high-level mental models and have them fail. You also tend to accumulate lots of weak ties to lots of people; if there's one thing I underestimated in my 20s and have started to appreciate in my 30s, it's just how many interesting friends-of-friends I'd end up with.
What a lot of these have in common is that they're traits that compound over long periods. Which raises one more possible explanation for the aging of everything: it's a real-world instance of the secular decline in real interest rates. An interest rate measures the value of some lump of productive assets, relative to its current output. When rates are high, current results are more salient; when they're low, the total accumulated value matters more. If you think of a political career as a process of forming valuable connections—to other politicians, to staffers, to donors, to voters and the people who influence them—it's an ongoing investment that accumulates value over time. And the natural converse of low rates producing high asset values for a given level of income is that low rates make accumulating productive assets with a given annual return that much more expensive. The political capital was locked up when it was cheap, and now it's expensive.
There is one last possibility: that part of what we're seeing is measurement error. If actors are getting older and the music we listen to is getting older, it may be because TikTok stars, Twitch streamers, and Roblox creators aren't being counted among entertainers, even if they have a similar-sized audience. One thing that drags down the average age of Fortune 500 executives is when tech startups with young founders go public, but many of those startups don't have the revenue to qualify for the Fortune 500, even if their market cap puts them in the S&P. Revenue is a lagging measure of impact, just as box office results are an output from fame. Political capital is scarce in systems where there isn't anyone famous or influential enough to be a gatekeeper, and where barriers to entry are low, so newer fields don't have the same compounding benefits from social networks and name recognition. There's too much turnover for the networks to stay valuable, and when everyone's getting better-known all the time, current fame is a weaker predictor.
And this can be further exaggerated when one segment of the economy is getting calcified because of risk aversion, which leads some young people to ignore it entirely. Working a low-paid service job while submitting articles to The New Yorker might at one point have been a viable way to break into writing, but now it raises the question: why wait for permission from gatekeepers instead of starting a Substack? If this two-track model happens, both tracks will exhibit weird demographics: the establishment will keep getting older, and the insurgents will stay young because they'll grow faster and take up a disproportionate share of the youngest people. And this low turnover in the establishment will also mean that social ties are disproportionately important—the people who make big decisions today will probably still be doing so a few years from now. Which further locks new entrants out. This model seems to work for entertainment and business, though it's hard to map it to politics; if there's a youth-friendly movement that goes around the traditional political hierarchy, it still has to contend with the fact that the establishment has jobs like the Presidency, majority leadership, and the Supreme Court. (On the other hand, what could be more fun than running an organization like, say, the FTC when your boss has a mandate to make a difference but lacks the expertise to closely supervise the process, and you have a specific theory for how it ought to be done?)
Demographics matter, at a macro scale and a micro one. And demographics as a topic almost never comes up unless someone has a big complaint. But sometimes problems at one scale become opportunities at another, or at least turn into an indication that the model you're using is missing something important.
A Word From Our Sponsors
Here's a dirty secret: part of equity research consists of being one of the world's best-paid data-entry professionals. It's a pain—and a rite of passage—to build a financial model by painstakingly transcribing information from 10-Qs, 10-Ks, presentations, and transcripts. Or, at least, it was: Daloopa uses machine learning and human validation to automatically parse financial statements and other disclosures, creating a continuously-updated, detailed, and accurate model.
If you've ever fired up Excel at 8pm and realized you'll be doing ctrl-c alt-tab alt-e-es-v until well past midnight, you owe it to yourself to check this out.
One of the most important long-term macro and political questions in the world is: are there complex products that can only be produced by mostly-capitalist, mostly-democratic systems, or, given time, is it possible for other countries to catch up? Two categories in particular that combine capital intensity, long lead times, and complex supply chains where a single mistake can render the investment worthless are widebody jets and advanced chips. In both cases, only a few places have been able to produce them profitably. (And, for chips, that list currently excludes the US.)
China has recently made significant advances in both: SMIC is shipping 7nm chips for crypto mining, putting them just two generations behind TSMC and Samsung. And this doesn't seem to be for bragging rights, either; as SemiAnalysis notes, the story broke when TechInsights did a teardown, not from a company announcement. Meanwhile, COMAC, a state-run aircraft manufacturer, is nearing certification for its C919 passenger plane. The plane still requires inputs from non-Chinese manufacturers, but it's much closer to being a standalone product. Over time, this makes it harder for the US to restrict what other countries do; one reason sanctions work is that some valuable goods can basically only be acquired by interacting with the dollar system, so a sanctions regime can almost completely cut off access to them. But if similar products are available from China, then neutral countries with big ambitions have two superpowers to think about, not just one.
Work from Home
Luxury rental buildings are increasingly incorporating work-from-home space into their offerings. Multifamily real estate tends to reflect tech changes on a lag, starting with the most expensive new units—you can estimate the age of a New York apartment building by looking at how much space they have for package deliveries, for example. These changes are incredibly durable, since the useful life of buildings is so long, and they're a way that temporary tech-driven swings can become more permanent behavioral changes.
In an open thread a week ago, I linked to a story about how better surgery robots have made it harder for surgeons to get training—there just isn't as much for an assistant to do, so there isn't much for them to learn. It's interesting to see technology solving a problem other technologies are creating: Theator, a startup that uses AI to analyze videos of surgery, just raised $24m. This doesn't perfectly solve the problem, of course; the training gap will still be there. But in some industries, part of how to deal with a weakening of apprenticeship models is to automate some of the work those apprentices would be doing.
Buffett and Energy
Yet Another Value Blog has a good in-depth look at what Warren Buffett might be seeing in energy: the stocks are incredibly cheap, even if you assume energy prices recede a bit (and that assumption embeds a bet on some combination of recession and green energy transition—but even then, it has to be aggressive about either). An even more useful piece is this follow-up, which provides an overview of how energy companies are thinking about prices: many of the big companies have reduced hedging compared to a few years ago, so they're more exposed to price decreases now that prices are high. This has historically been a bad sign ($, WSJ): energy companies are marginally worse than the rest of us at timing the cycle: they can respond both by changing their hedging strategies and by changing how much they invest in production, and they have to hedge more at low prices because their lenders insist on it. But they're also well-informed, and they could be reducing hedges because, having paid down so much debt with their recent cash flow, they can finally afford it.
It's useful to differentiate between hard power (being able to invade someone, bomb them, or embargo them) and soft power (being able to influence a country by affecting what music they listen to, what books they read, and ultimately and very indirectly how they vote). But these are intimately tied together; part of the point of soft power is to convince people that they'd really be better off not fighting, and that affects how well a country can recruit. It's interesting to see a blunt example of this: in areas of Ukraine occupied by Russia, local authorities are banning Google and, which will probably mean replacing it with Yandex.
Diff Jobs is our service matching readers to job opportunities in the Diff network. We work with a range of companies, mostly venture-funded, with an emphasis on software and fintech but with breadth beyond that.
If you're interested in pursuing a role, please reach out—if there's a potential match, we start with an introductory call to see if we have a good fit, and then a more in-depth discussion of what you've worked on. (Depending on the role, this can focus on work or side projects.) Diff Jobs is free for job applicants. Some of our current open roles:
A company building a universal API for accounting software is looking for mid- to senior-level engineers. (Austin)
A company using unique data sets to help investors create new insights is looking for data scientists. (US, remote)
A new quantitative hedge fund is looking for quant developers. (NYC/New Jersey)
A new fund applying systematic signals to early-stage investing is seeking quants (Bay Area).
Talented people with strong technical, product, and growth/marketing skills are encouraged to reach out—we'd love to chat even if there's not an immediate opening.
I've wondered if one of the reasons that fraternities and the Bullingdon Club seem to produce a disproportionate number of influential people is that they start with individuals who have already been selected for some combination of family influence and personal charisma, and then give them all some fond shared memories and/or blackmail material. It's a way to create synthetic social capital, especially if the club's practices mean that no member is entirely sure what they did and thus doesn't know the extent of how compromised they are.
Not that lifestyles have been universally improving. Obesity has gone up a lot, for example. However, this might have less of an impact at the tails of the distribution; a Washington Post study indicated that obesity leads to a nine-point lower voting preference. They're also less likely to get hired. So looking at a cohort of successful people is implicitly looking at a population that's less obese.
It's a bad mental habit to weight luck too heavily, but it's a good habit to know, on the margin, what makes events more luck-dependent. And one thing that does that is having lots of competitors who are angling for a small number of spots. The more of them there are, the smaller the skill gaps between candidates are, which means that random variance matters more. That's especially true in fields where the test can only approximate the skillset. Companies invest a lot of effort in figuring out which recent college graduates will make good programmers and traders, but they still get most of their information on skills after the hire is made. (If that weren't true, you'd see higher variance in starting offers followed by fairly lockstep raises. Instead, the variance shows up a few years after hire.) The more confident you are in your skills, the less you should aim for the obvious top job in whatever your chosen field is, because the closer you get to being the best, the more your outcome is determined by being the luckiest, too.