| One occasional interest of mine is whether people should make investment decisions using astrology. The first-order analysis is: "Making investment decisions using astrology is dumb. There is no fundamental reason to think that the alignment of the planets influences asset prices, no mechanism other than medieval superstition." There is a relevant xkcd. But then the second-order analysis is: "Look, it doesn't matter that it's dumb. People do lots of dumb stuff, and financial markets reflect the stuff that people actually do, dumb or not. If, empirically, lots of people make trading decisions based on astrology, then the alignment of the planets will influence asset prices. The mechanism is not medieval magic, it's regular old investor psychology. Good investors and traders have to understand and take advantage of other people's irrationality. Your job as an investor or trader is to anticipate what other investors will do, and if those other investors are influenced by beliefs about the planets then you should start casting horoscopes." There is a relevant passage in Keynes. My casual sense is that in the US stock market you probably can't make any real money using astrology — "there are not so many astrology pods" at big multistrategy hedge funds, I have written — but you can't rule it out a priori, and it probably works in some markets. (Not investing advice.) We have discussed an academic paper finding that Chinese mutual funds show "a significant 6.82% reduction in risk-taking during managers' zodiac years, traditionally considered unlucky in Chinese culture"; maybe that's a signal you can trade on. There's a small weird literature. If you are a finance professor interested in working on astrology, I cannot promise that it will get you tenure, but there's a good chance that it will get you mentioned in Money Stuff. Meanwhile, what does move the US stock market is the whim of Elon Musk, and if Elon Musk starts making his own economic decisions based on the positions of the planets then guess what, you're in the astrology business now. The Financial Times reports: Elon Musk has proposed timing SpaceX's initial public offering to coincide with a rare planetary alignment and his birthday, as the world's richest man seeks an auspicious date for what would be the largest listing in history. The rocket maker is targeting mid-June for its IPO when Jupiter and Venus will appear very close together, known as a conjunction, for the first time in more than three years, said five people familiar with the matter. SpaceX is seeking to raise as much as $50bn at a valuation of roughly $1.5tn, the people added, which would make it the largest IPO in history and far exceed the $29bn raised by Saudi Aramco in 2019. They cautioned that all figures remain preliminary and may change. ... On June 8 and 9 Jupiter and Venus will be "within a little more than 1 degree of each other in the sky, about the width of a thumb held at arm's length", according to non-profit organisation The Planetary Society. A few days later, Mercury will also align diagonally with the two planets. I guess if your business is shooting rockets into space you have somewhat more reason to care about planetary alignment than the rest of us do, though I can't see how Mercury's position would be relevant to the IPO or even really to the rocket launches. Also I don't exactly know how you would trade this particular piece of information — SpaceX is not currently public — but the point is that, if Elon Musk sometimes cares about astrology, then the market sometimes has to care about astrology too. Also his birthday.  | | | Lots of companies have multiple divisions that make and sell different products, and it is interesting to think about how they fit together. [1] There are some companies that take an extremely free-market, arm's-length approach to their different divisions: Each division has to stand on its own, the divisions are free to compete against each other for business, and if one division wants something from another division, it has to negotiate and pay a market price for it. Every business head is compensated purely for the success of her own division. Other companies take a more collective approach: The divisions are supposed to help each other out; each employee is supposed to work for the overall success of the company, not the success of her division at the expense of others. Every business head is paid a bonus that reflects her own division's results, but also her contribution to the broader company. [2] At many companies, one division will be favored more than the others by senior management. The favored division might be the especially profitable one, or the one that is growing the fastest, or the one that gets the most media attention, or the one managed by the chief executive officer's son. If the favored division goes to the CEO and asks for more capital to build a factory, the CEO will be happy to give it the money; the favored division has, as it were, a lower internal cost of capital than the less-favored divisions. And if the favored division wants something from one of the less-favored divisions — if the favored division asks another division's salespeople to cross-sell its products, or to lend it a few employees for an engineering sprint — the managers of the less-favored divisions will do their best to help. After all, they are good team players; they are working for the overall success of the company, and the overall success of the company depends more on the favored division than on the rest of them. Again, sometimes the favored division is the CEO's son's division, but at a well-run company the favored division really will be the one that is most important to the overall success of the company. One way to think about this, particularly at a public company, is using market valuation signals. If the stock market tells you "we will pay $20 for $1 of annual profit in the widget division, but only $10 for $1 of profits in the sprocket division" — definitely something that the stock market is capable of telling you — then that is, roughly speaking, a signal that the widget division is twice as important to the company as the sprocket division. Widgets are the future, and increasing the revenue and profit margins of the widget division will create more shareholder value than you could get out of sprockets, so management should focus on widgets. And the head of the sprocket division, if she wants to be a team player and help the company succeed, should do her best to help out the widget division. As a matter of incentives, the head of sprockets might get a bonus that reflects not only sprocket sales but also her contributions to the widget division. But even if she doesn't, she will probably get paid largely in stock, and her stock will be worth more if the widgets division does better, because that's the division that the stock market really values. So the head of sprockets will want to help out with widgets, because she is a good team player but also because that is in her personal financial interest. This creates, at least potentially, an accounting problem. Let's say the sprocket division has some resource, and the widget division wants it. Widgets will go to sprockets and say "hey I know you were planning to use that steel to make sprockets, but we are out of steel to make widgets, can you give us yours?" In a pure arm's-length free-market company, the head of sprockets might say "no," or she might say "sure but it'll cost you" and charge widgets the market price of the steel plus a markup for her trouble. But in a more congenial team-oriented company, she'll be more congenial and team-oriented. "Sure, I can help out," she'll say; "take some of my steel." But what will she charge? On the company's books and records, there has to be some sort of entry. She has transferred some steel to the widget division; the widget division will have to transfer some money back to her. That price will be negotiable, and in the negotiation, both sides might want to get a good result for the widget division. If the widget division pays less than the market price for the steel, its profits will increase, and the sprocket division's profits will correspondingly decrease. But the head of widgets gets paid for increasing the widget profits, and the head of sprockets also kind of gets paid for increasing the widget profits, and the CEO gets paid for increasing the widget profits, and everybody wants to increase the widget profits, so everyone will be happy if the sprocket division sells the steel to the widget division at a discount. Certainly the shareholders will be happy: They value $1 of widget profits at $20 and $1 of sprocket profits at $10, so increasing widget profits by $1 at the expense of $1 of sprocket profits increases the overall value of the company by $10. Good deal all around! But that's cheating. That's not really $1 of extra widget profits; that's not $1 of profit created within the widget division by being good at making widgets. That's just a subsidy from sprockets; it's just moving money that "really" belongs to sprockets to widgets, because shareholders will pay more for widget profits than for sprocket profits. It's securities fraud. A crude simplified model of Archer-Daniels-Midland Co. in 2021 and 2022 might be [3] : - Its Ag Services and Oilseeds ("AS&O") division was in the business of acquiring soybeans and processing them into something called "white flake."
- Its Nutrition division was in the business of acquiring white flake and processing it into "food products it sold to external customers."
Just from those words, you can probably fill in the rest of the story. "White flake," as a product, just sounds less appealing than "food." [4] Turning soybeans into white flake seems like a commodity business. Turning commodities into food seems like a better business that the market might value more. And, since (some of) the outputs of AS&O were (some of) the inputs for Nutrition, ADM had a lot of flexibility to allocate profits however it wanted: If AS&O spent $10 for some soybeans, and then spent $5 turning them into white flake, it might sell the white flake on the open white-flake market for $16. And Nutrition might buy white flake on the open market at $16. But if AS&O sold the white flake to Nutrition for $7, then that would reduce AS&O profits by $9 and increase Nutrition's profits by $9, and that would be better for everyone. But that's cheating, and yesterday the US Securities and Exchange Commission sued. (It settled with ADM for $40 million, and also settled with two other executives; it is suing ADM's former chief financial officer, Vikram Luthar, whose lawyer "said he would fight the allegations.") We have talked about these allegations before, when they were first reported in 2024, though I had to guess at some of the details. But now we have the SEC's complaint, and all the details are pretty much what you would have guessed. Nutrition got a higher multiple than AS&O or Carbohydrate Solutions, ADM's other main division: ADM pursued this investment strategy [of investing in Nutrition] because equity shares of food processing companies were trading at higher multiples than mature, commodity-based companies like ADM. In other words, ADM recognized that investors viewed food processing companies as having higher growth potential (and less volatile earnings) than ADM's traditional, commodities-based business. … Although Nutrition was ADM's smallest business segment by gross annual revenue, at all times relevant to the Complaint, Luthar and other ADM executives: (a) routinely described Nutrition – both internally and to the investing public – as a key driver of ADM's growth strategy; and (b) knew that Nutrition's operating profit growth was a key metric evaluated by investors and prospective investors in ADM. And so ADM's executives were paid based in part on Nutrition, even if they worked in other divisions: The Nutrition segment was so important to ADM and its investors that ADM developed executive compensation plans that were tied to Nutrition's performance – even for executives who worked for ADM's other business segments (i.e., AS&O and CarbSol). To incentivize management to bolster the Nutrition segment, Nutrition's operating profit growth target was added as a performance metric in ADM's 2020 and 2021 cash bonus and equity long-term incentive plans for eligible ADM employees, including Luthar and senior executives of CarbSol and AS&O. In practical effect, as of 2020, ADM's compensation plans incentivized AS&O and CarbSol executives to assist Nutrition (even though they were not working for that segment). In addition, executives and other employees at AS&O and CarbSol widely understood that to be considered a "good corporate citizen" by senior management, they should at times be willing to assist Nutrition – even when that meant adversely affecting their own business segment. And so, allegedly, they would shift profits from other divisions to Nutrition to meet profit targets. Stuff like this: Nutrition faced an unanticipated obstacle in late 2020 that threatened its ability to meet its publicly disclosed 15% to 20% operating profit growth target for FY 2021. As part of its routine operations, Nutrition bought a processed soybean product called "white flake" from AS&O, which Nutrition then used to manufacture food products it sold to external customers. Nutrition agreed that it would pay AS&O for white flake based, in part, on the market price for soybean meal (white flake's main component). Nutrition and AS&O memorialized each purchase in a purchase order reflecting the then-current market price for soybean meal. But, in the fourth quarter of 2020, the market price of soybean meal rose significantly, causing a corresponding increase in the price of white flake sold by AS&O to Nutrition. This substantial and unexpected increase in the price of white flake would have negatively affected Nutrition's operating profit margin. ... Luthar executed a plan to retroactively re-price the white flake transactions between Nutrition and AS&O to reflect the pre-spike August 2020 market price for soybean meal, which would: (a) improve Nutrition's operating profit; and (b) allow Nutrition to achieve its initial FY 2021 operating profit growth forecast. As this plan took shape in late 2020, Luthar knew that AS&O executives were incentivized to be compliant partners rather than arm's length negotiators. Nutrition's President reminded Luthar that AS&O's President was financially motivated to help Nutrition even though a transfer of operating profit would adversely affect AS&O. Nutrition's President told Luthar that AS&O's President "is willing to work with us on white flake too. We spoke privately last night and he is severely motivated by [ADM's Performance Incentive Plan] and [Performance Share Units] to help us." Right, yes, good teamwork, arguably bad accounting. Investment bankers mainly get two kinds of pep talks from upper management: - "You need to go out, pitch more clients and win more deals."
- "You need to go out less, pitch fewer clients and do fewer deals."
We talk from time to time around here about the second variety, which, as a former investment banker who never did many deals, I find more congenial. Sometimes bankers get the "actually do less" talk for compliance reasons (if you pitch a lot of clients, you will end up working with some bad ones), but sometimes they get it for pure efficiency reasons. Investment banking is a business that involves doing a lot of free work for prospective clients in the hope that they will eventually pay you a fee to do a deal, and sometimes senior management will notice that their bankers are doing a lot of free work for slim chances of small fees on marginal deals and tell them to knock it off. But the first pep talk — "go do more deals" — is surely more common. Bloomberg's Crystal Tse and Swetha Gopinath report: JPMorgan Chase & Co.'s leadership has told the firm's investment bankers that they need to work harder to close the gap with rivals including Goldman Sachs Group Inc. on mergers and acquisitions, according to people with knowledge of the matter. John Simmons and Filippo Gori, co-heads of global banking, delivered the message to staff during an internal meeting this month, the people said. The co-heads said the group had underperformed in M&A in 2025 and that improvement was needed to start winning back market share, according to the people. Sure. I suppose if you work in sales at, say, an enterprise tech company, and management wants you to sell more software, they sit you down and talk about all the new features of the product and all the new customers it could help. But investment banking is just kind of investment banking, and if you want to do more of it you just have to do more of it. You could imagine that the people who are most worried about the risks of artificial intelligence would not be the most productive builders of AI. You could have a model like "actually AI is good, and the people who are worried about it are wrong, and being wrong does not make you a good AI researcher." Or you could have a simpler model like "well, people who are really scared of AI probably won't work 20 hours a day to build AI as fast as possible, and we want to hire the people who will." But in fact there seems to be a surprisingly strong positive correlation between noisily worrying about AI and being good at building AI. Probably the three most famous AI worriers in the world are Sam Altman, Dario Amodei and Elon Musk, who are also the chief executive officers of three of the biggest AI labs; they take time out from their busy schedules of warning about the risks of AI to raise money to build AI faster. And they seem to hire a lot of their best researchers from, you know, worrying-about-AI forums on the internet. You could have different models here too. "Worrying about AI demonstrates the curiosity and epistemic humility and care that make a good AI researcher," maybe. Or "performatively worrying about AI is actually a perverse form of optimism about the power and imminence of AI, and we want those sorts of optimists." I don't know. It's just a strange little empirical fact about modern workplace culture that I find delightful, though I suppose I'll regret saying this when the robots enslave us. Anyway if you run an AI lab and are trying to recruit the best researchers, you might promise them obvious perks like "the smartest colleagues" and "the most access to chips" and "$50 million," but if you are creative you might promise the less obvious perks like "the most opportunities to raise red flags." They love that. The Information has a story with the wonderful title "OpenAI One-Ups Anthropic with New Whistleblower Policy," suggesting that this is the sort of thing that AI labs compete on now: OpenAI's new policy is the strongest whistleblowing policy that any of the leading AI companies have published. It goes further than Anthropic's, which was published in December. Anthropic's version is narrowly focused on employees reporting violations of its own safety and security policy and lacks a timeline when it pledges to respond to complaints. A spokesperson for Anthropic said in a statement that the company has a non-retaliation policy similar to OpenAI's that covers additional topics, such as fraud and conflicts of interest, including an anonymous reporting line. Anthropic has also said employees do not have to sign non-disparagement agreements that would make it difficult for them to publicly raise safety concerns. "Actually we have even more ways for you to report complaints" is not the most obvious recruiting advantage. Do you want to recruit the people who are most excited to report their complaints? In AI, counterintuitively, you do. AI hiring is weird, though, because do you even want the most productive researchers? The contrarian logic of AI is not just "the people who are most afraid of AI are the best people to build it," but also "the people who are building productive useful AI quickly are not the most prestigious AI people." They are sullied by commerce. The most prestigious AI research is completely disconnected from boring stuff like applications or revenue or shipping products; the most prestigious AI research is about fundamental questions of what intelligence is and how to create it. This is sort of how everything works in academia — the most prestigious mathematicians or physicists or literary theorists are the ones who are most removed from the grubby realities of everyday commerce — but modern AI is like fancy academia with billions of dollars of venture capital. "My basic model of modern artificial intelligence labs is that they are like emotionally intense graduate school programs but with unlimited money," I wrote the other day. So here's a Wall Street Journal story about "neolabs," small AI startups with excellent researchers who keep themselves completely untainted by commerce so they can raise a lot of venture capital: Ben Spector had an unusual pitch for investors last fall. A Ph.D. student at Stanford University with a highly prized artificial-intelligence background, Spector had no near-term plans to make money and no traditional pitch deck. He didn't even have an idea for a hit AI product. What he did have was a lab, called Flapping Airplanes, a novel idea for training AI models and a zeal to hire talented young researchers eager to tackle AI's biggest problems. Venture-capital firms jumped at the chance to back him. ... Flapping Airplanes—a reference to the biological cues future AI should take from nature—is part of a new wave of startups some have dubbed "neolabs," which give priority to long-term research and developing new AI models over immediate profits. It's not an unusual pitch! We talk about it all the time! It was Mira Murati's pitch for Thinking Machines, which was essentially "we have a lot of top AI researchers, give us money and don't ask about products." It was also Yann LeCun's pitch for his new startup, about which I wrote: It is really an amazing time to be an AI researcher. You can come into work to pursue pure science, uncontaminated by any need for near-term practical applications or making money. And when your boss comes to you and says "hey guy not to be annoying but I am paying you $100 million, is there any way we could turn any of this stuff into money," you can quit in a huff and find venture capitalists who will give you billions of dollars to be even less commercial. And they're not even wrong! Betting on long-term theoretical AI research has turned out to be hugely lucrative for a lot of investors; there's no reason not to keep doing it. It's a golden age of academic researchers getting bazillions of dollars to pursue pure research.
Yeah I mean keep getting those bags, guys. Popular Dutch Pension Trades Are Backfiring After Months of Hype. Trump's Embrace of Weaker Dollar Fuels Bets on New Downtrend. SoftBank in Talks to Invest Up to $30 Billion More in OpenAI. Private Wealth Fuels a $240 Billion Market for Secondhand Investments. Private Equity Braces for Major Changes in 401(k) Investing Rules. How private equity's pioneer in tapping retail money lost its edge. Wells Fargo Cuts Proxy Adviser Ties in Latest Blow to Industry. Star Kirkland & Ellis partner's future in doubt as firm sides with PE in legal tactics dispute. Deutsche Bank Raided in Money-Laundering Probe of Employees. Silver squeeze leaves solar panel makers feeling the heat. Amazon to Cut 16,000 Corporate Positions to Trim Bureaucracy. Amazon to Shut Down All Amazon Go and Amazon Fresh Stores. Amazon inadvertently announces cloud unit layoffs in email to employees. BlueCrest Goes to Supreme Court in Senior Trader Tax Dispute. France presses Capgemini over ICE contract amid international backlash. Fish pile up at European ports as new digital system falters. Weill Gives $120 Million to Vet School That Treated His Dog. If you'd like to get Money Stuff in handy email form, right in your inbox, please subscribe at this link. Or you can subscribe to Money Stuff and other great Bloomberg newsletters here. Thanks! | 
No comments:
Post a Comment