Money Stuff: Buy the Truth, Sell the News

Programming note: Money Stuff will be off next week for spring break. We’ll be back April 21. In recent years, my vacations have often been occasions for Elon Musk to do weird stuff, and that is certainly a live possibility for next week. Also, though, I once took a vacation during which Lehman Brothers filed for bankruptcy. Probably nothing to worry about though.

One informal rule that I set for myself in writing this newsletter is that I try not to pay too much attention to politicians’ tax and regulatory proposals, because often they come to nothing and it’s silly to take them too seriously. I spent the last week breaking this rule for President Donald Trump’s tariffs, and look where that got me: 

President Donald Trump announced a 90-day pause on higher tariffs that hit dozens of trade partners after midnight, while raising duties on China to 125%.

The president’s about-face came roughly 13 hours after high duties on 56 nations and the European Union took effect, fueling market turmoil and stoking recession fears. Trump faced massive pressure from business leaders and investors to reverse course. ...

Stocks staged their best rally since 2008 as euphoria gripped markets after Trump’s abrupt announcement. The S&P 500 Index soared 9.5%, rebounding from bear-market territory. The tech-heavy Nasdaq 100 surged 12%. Goldman Sachs Group Inc. economists rescinded their forecast for a US recession.

Sure! There was a formula! There was a whole worked-out theory of trade! This was a centerpiece of his policy! He was kidding! In my defense, financial markets also took the tariffs seriously, and were down; now the tariffs are off and the markets soared yesterday.

Well, the tariffs aren’t off. There are still 10% tariffs on everything, and 125% tariffs on China, with “little prospect for a near-term detente.” Adam Tooze writes: “China is so huge that with 125 percent tariffs on that one country, the overall tariff level is actually higher than before Trump announced his ‘concessions.’” It is still a dramatic shift in US economic policy, but now it comes as a bit of a relief.

Also though this is just a pause, right? Technically, in 90 days we’re supposed to be right back here, with high pseudo-reciprocal tariffs on every country and the formulas and the non-tariff barriers and the reindustrialization of America and the wholesale restructuring of global trade. I will definitely have forgotten all of this by then, though, and so will you, and so will financial markets, and probably so will Donald Trump. “Remember the 29% tariff on penguins,” people will idly muse, like it was a dream.

Anyway if you were on Truth Social yesterday morning you got shockingly good financial advice:

President Donald Trump on Wednesday urged calm as global markets continue to be upended by his wide-reaching trade war.

"BE COOL! Everything is going to work out well," Trump on Truth Social minutes after the New York York Stock Exchange opened. "The USA will be bigger and better than ever before!"

He added shortly after, "THIS IS A GREAT TIME TO BUY!!! DJT”

The “be cool” post hit at 9:33 a.m. yesterday, the “great time to buy” one at 9:37 a.m. The official tariff pause was also announced first on Truth Social, at 1:18 p.m. If you look at a chart of the S&P 500 index yesterday, there was a little bump at 9:37, and then a huge vertical line at 1:18.

So if you bought stocks yesterday morning on Trump’s advice, you made a lot of money. What does it mean? What does anything mean? Like, the possibilities are:

1. Trump believed, yesterday morning, that his tariffs were good for America and that this was a great time to buy, because of the tariffs. So he Truthed that. And then a few hours later he changed his mind and paused the tariffs.
2. Trump knew, yesterday morning, that he was going to pause the tariffs, and he accurately predicted that that would cause the stock market to rip, so he gave his followers on Truth Social a little preview.

Each of those options is pretty funny, but the first one is (1) less funny and (2) surely correct. The “great time to buy” stuff meant that the tariffs were good, that you should buy stocks because of the high tariffs on every country; the four-hours-later tariff pause was a completely separate and unforeseen event. You can tell in part from first principles — why would Donald Trump decide something and then wait patiently for four hours to post about it? — but also from Bloomberg’s reporting about how and when he changed his mind yesterday:

The president started his day monitoring the reaction on Fox Business, where a parade of executives and traders on the normally friendly network sounded alarm. …

Trump turned to social media to encourage followers it was “a great time to buy.” The missives underscored to Wall Street that the president was paying attention, helping fuel a stock rebound.

“BE COOL! Everything is going to work out well. The USA will be bigger and better than ever before!” Trump posted to his Truth Social site.

Still, as the morning stretched on, Trump decided to gather some of his top economic advisers at the White House. Notice went out to US Trade Representative Jamieson Greer, who was testifying before a House panel, that Trump was rethinking the tariff levels.

Treasury Secretary Scott Bessent, who had been scheduled to address the congressional Republican Study Committee to discuss tax legislation, hastily canceled his appearance to return to the White House. Commerce Secretary Howard Lutnick was also summoned. The financial markets were weighing heavily on his mind. …

The trio sat to draft a social media post to tell the world that he was, at least temporarily, rolling back the tariffs, writing the text without even consulting the president’s legal team.

“Just wrote it up,” Trump said. “It was written from the heart, and I think it was well written, too.”

“Be cool” and “great time to buy” were not previews of the tariff pause; they were the last effort to change the market’s mind about the tariffs being good. They didn’t work, so he paused the tariffs.

Still! Many, many readers emailed me with variations on the question “is this insider trading” or “is this securities fraud” or “is this market manipulation,” and … no? [1] Even if the “great time to buy” post was based on an intention to pause the tariffs, it’s not fraud if it’s true and publicly disclosed! I mean, if he spent the morning calling a bunch of his buddies to say “this is a great time to buy, wink wink,” and they bought stocks and then he announced the tariff pause, sure, insider trading. Obviously in a “unitary executive” no one would bring a case about it, but still.

But if you say it publicly it’s not insider trading! (Not legal advice!) Everyone who follows him on Truth Social, or who reads news articles about his Truth Social activity, could trade on this tip or whatever it was. I was vaguely aware, yesterday morning, that Donald Trump had been Truthing that it was time to buy stocks, though I did not buy stocks. I just, you know. Howard Lutnick said on television last month that Tesla Inc. stock would “never be this cheap again,” and then it was. The public investment advice of the Trump administration has not generally been infallible, nor does it seem motivated by dispassionate objective analysis. I just assumed that part of the administration’s job was to make public statements pumping stocks to make Trump’s economic policy look good. It did not occur to me that Trump was transmitting useful inside information about the economic plans of the president of the US. It’s not like he knows what those plans are! Though he is the president.

This is all pretty stupid, but it’s probably less stupid than the alternative of leaving the tariffs in place, so I’ll take what I can get.

The stylized story of bond trading over the last few decades is that, before 2008, when a customer wanted to buy or sell bonds, it would call a big bank. “I want to sell some bonds,” the customer would say, and the bank would say “sure I’ll buy them.” The bank would use its own money to buy the bonds, and would hold them on its balance sheet. Eventually it would find another customer who wanted to buy them, and it would earn a spread on the trade.

This was a risky and capital-intensive business, and 2008 happened, and banks retreated a bit from it. Regulators raised capital requirements, so it was more expensive for banks to hold bonds; the Volcker Rule in the US also discouraged banks from trading with their balance sheets. Meanwhile technology got a bit better, and you could more often trade bonds without making a series of phone calls.

And so the business model of bank traders shifted somewhat. It used to be that, when a customer called a trader and said “I want to sell 100 XYZ bonds,” the trader would say “I’ll pay you $97 for them” and plan to hold them on her balance sheet. Increasingly, though, when a customer calls a trader and says “I want to sell 100 XYZ bonds,” the trader will say “let me put you on hold a minute” and will call around to other customers to see if they want to buy them. If she finds a buyer, she can do the trade, matching up the buyer and seller and taking a cut for her efforts. That way she doesn’t have to hold the bonds on her balance sheet, tying up capital and taking market risk.

This is a broad stereotype; of course banks did riskless matching trades like that before 2008, and they still do plenty of risky principal trades on their balance sheets now. But stereotypically there has been something of a shift to a more agency, less principal model, as it becomes harder for banks to take big trading risks.

This arguably created a new business niche. The niche is: Be the other customer that the bank calls. If some customer calls a bank trader and says “I want to sell 100 XYZ bonds,” and the trader says “let me put you on hold a minute,” you want her to call you during that minute. She offers you the bonds, you take them. And then an hour or a day or a week later, she calls you again and says “hey now I have a buyer of XYZ bonds, you want to sell?” And you sell. In expectation, the price that you sell at is higher than the price you bought at, in part because you are good at predicting bond prices but mostly because you are providing a service. The service you are providing is liquidity. It is balance sheet.

Historically, this was a service that the bank provided, a service that came bundled with the bank’s other services like “answering the phone when customers call” and “taking the customers out for drinks.” But as banks retreat from providing liquidity with their balance sheets, there is an opportunity for someone else who can provide that service. Arguably that someone else could provide the full bundle — liquidity, phones, drinks — but that is not strictly necessary. You could let the bank keep answering the phone, and just provide liquidity, through the bank, on the back end.

Bloomberg’s Katherine Doherty had a story yesterday about perhaps the most obvious firm to fill this niche, Citadel Securities:

More than 30 banks are engaged in talks — some more advanced than others — with billionaire Ken Griffin’s market-maker, hashing out an arrangement that would let them submit orders to the firm without revealing their clients’ identities. Citadel Securities is marketing the concept as a way for small- and mid-tier banks to provide better pricing on fixed-income trades. But it also would give it more insight and clout in markets. …

Citadel Securities is zeroing in on fixed-income operations, where the strongest banks have invested heavily in systems to consolidate market share. Its executives canvassed dozens of banks and smaller broker-dealers to discuss their pain-points and concerns.

Banks are highly protective of their client rosters. To defuse concerns that Citadel Securities might later swoop in on customers directly, the proposed system wouldn’t reveal their names. Instead, Citadel Securities is enlisting a third-party firm, TransFICC, that will install its technology at banks and brokerages while sparing them the cost to connect.

TransFICC’s systems would broadcast requests for quotes across banks, which could accept or decline prices they see. If banks choose the price from Citadel Securities, it would fill the order without knowing which client is on the other side. …

“The solution here is for us to be able to sit behind a bank, to be able to provide them with liquidity and competitive pricing that they can send back to their end-user,” said Amit Bhuchar, head of liquidity solutions for fixed-income, currencies and commodities at Citadel Securities. “What the partner chooses to do with their client is completely at their discretion.”

We talked about this model last year, when Doherty first reported that Citadel Securities was trying to get into the “white-label trading” business. It is roughly what Citadel Securities and other market makers already do in the retail equity market: If you put in an order to buy stock through your Robinhood app, Robinhood will send the order to Citadel Securities or a similar competitor to fill it. Robinhood has the app and the customer relationship; Citadel has the balance sheet and the risk appetite. There’s no reason that those things have to be bundled together. Historically, in institutional bond trading, they were, but that can change.

The basic deal is that if you are a high-level quantitative researcher at a hedge fund or proprietary trading firm, and you get an offer to go work at a rival firm, your current firm will usually (not always) require you to take months or even years of “gardening leave,” during which your old firm pays you not to work at your new firm. In some ways this is nice for you — lots of paid time off — but in other ways it is not: You probably get paid less, considering bonuses etc., than you would at the new firm, you might get bored, and your skills atrophy when you are away from the cutting edge of quantitative finance. By the time you start at your new firm, you are less useful than you would have been when you left your old firm, which is of course the point of the arrangement (for your old firm). 

I have always thought of this as a particularly financial-industry thing. You do not see it so much in tech, because tech mostly exists in California where noncompete agreements are mostly unenforceable. But not exclusively in California. Business Insider reports:

The battle for AI talent is so hot that Google would rather give some employees a paid one-year vacation than let them work for a competitor.

Some Google DeepMind staff in the UK are subject to noncompete agreements that prevent them from working for a competitor for up to 12 months after they finish work at Google, according to four former employees with direct knowledge of the matter who asked to remain anonymous because they were not permitted to share these details with the press.

Meanwhile, here is a story about how “Liverpool [Football Club] are set to appoint Laurie Shaw as their new chief scientist when his gardening leave from the City Football Group ends.” It would be interesting if professional athletes had gardening leave, [2] but as far as I can tell the main gardening leave in sports is for data scientists.

One curious discovery of modern life is that a lot of apparently disparate businesses are actually machine learning businesses. We talked a few months ago about how DeepSeek, an influential artificial intelligence startup, grew out of a quantitative hedge fund. It turns out that those things are the same thing. You might naively have thought that “teach a computer to write lengthy texts in grammatical English that read like they were written by humans” and “pick stocks that will go up” were wildly different activities, but in fact they are approximately the same activity. I wrote:

It turns out that there is a sort of general skill like “program a computer to take a huge pile of analogous data and predict the most likely next _______,” where the blank can be filled in with “word in the sentence” or “pixel in the image” or “stock that will go up.” And people and companies with this general skill can move between various applications, using similar techniques to pick words or images or stock picks.

Similarly you might have thought that running a professional soccer team is very different from running a quantitative hedge fund or programming a computer to write essays, and in some crude mechanical sense it is, but in a broader sense the soccer team is also essentially a machine learning project. That Liverpool story has more on Shaw’s background:

Shaw had been at CFG since 2021 but had previously worked for a £30billion hedge fund while also acting as a policy advisor to the British government. He holds a PhD in Astrophysics from Cambridge University, and an Msci from Imperial College, London.

Oh yes of course astrophysics; astrophysics is also the same skill set as stock picking, language models and soccer analytics. 

Anyway I have previously somewhat-but-not-really jokingly said that this is good, that “the general background condition of ‘a lot of people want to learn statistical techniques to get rich’ creates a deep pool of talent for all of the applications” of deep learning, that we have a pipeline of astrophysicists because getting an astrophysics PhD is a good path to running a hedge fund or a soccer team. 

But of course all the gardening leave feels like a loss to society? All these highly trained astrophysicist hedge fund researchers and AI researchers and soccer analysts, sitting around twiddling their thumbs or gardening between hedge-fund or AI or soccer jobs. Really there should be an exchange program, a gap year for quants. Your gardening leave from a [hedge fund][AI lab][soccer team] should prohibit you from working for another [hedge fund][AI lab][soccer team] for a year or whatever, but it should encourage you to work for a [soccer team or AI lab][hedge fund or soccer team][hedge fund or AI lab] or, in each case, university astrophysics department, [3] during that year. Have some sort of centralized exchange or website where the hedge-fund astrophysicists can match up with AI labs or sports teams for their gap years, where the AI-lab astrophysicists can match up with hedge funds, etc. (I feel like the sports teams would get a lot of takers for one-year jobs?) Just seems good for everyone for all the astrophysics PhDs to work across all the applications.

We talked last month about a Texas lottery arbitrage. In certain situations, it is a positive expected value trade to buy every combination of numbers for a Lotto Texas drawing, and in 2023 someone did, “collecting a one-time payment of $57.8 million, by acquiring virtually all of the 25.8 million possible number combinations.” (The headline jackpot was $95 million; the $57.8 million is the one-time lump-sum version.)

One reason I wrote about it is: That’s cool. Another reason I wrote about it is that people seem to find it miscellaneously unsavory. You are not supposed to be able to buy 25.8 million different tickets for the lottery drawing. It’s not exactly that it’s illegal to buy every ticket. It’s more that you’re supposed to buy your tickets from regular stores that sell things other than lottery tickets. The syndicate that bought all the tickets seems to have set up special offices with lottery terminals to automate printing every ticket, which violates at least the spirit of the rules. Even they think so: They asked Texas lottery officials to help them do it, but they testified in the state legislature that “we fully expected that they would laugh at us and say, ‘Well, no, of course you can’t do this.’” But the officials let them do it, so they did it, and they won the lottery. And now there are investigations.

Let’s assume that, in fact, this was not allowed. It was bad, a violation of some rule or law or norm of behavior. Who was the victim of the bad thing? One possible answer is “the state of Texas, which gave this syndicate $57.8 million,” but that can’t be right; the $57.8 million was going to get paid out for lottery prizes anyway. Another answer might be “the state of Texas, because this undermined confidence in the lottery and thus reduced future lottery revenue.” That answer feels about right — if lottery prizes are just for big professional syndicates, why would you play? — and I assume it’s why the legislature has investigated it.

But there is another, more direct answer. Lotto Texas has a jackpot that builds each time no one wins. The person who won the jackpot after this syndicate missed out on the money that the syndicate won; if they had never played the lottery, the $57.8 million that they won would have stayed in the jackpot and grown the next week’s prize. [4]  Their arguably illicit jackpot mechanically reduced the next winner’s jackpot; the next winner is the direct victim. It seems a bit silly to say “I was victimized by winning a lottery jackpot that was lower than it should have been,” but I guess it is technically correct.

Anyway that guy sued:

A Texas lottery player filed a lawsuit in Travis County on Tuesday alleging his May 2023 Lotto Texas jackpot prize was severely diminished by $95 million because a group of lottery retailers and a London-based sports gambling company conspired to rig a lottery drawing less than a month before his win.

The plaintiff’s name is Jerry B. Reed of Hood County, Texas. On May 17, 2023, the lawsuit says Reed won the Lotto Texas jackpot which, at the time, was worth $7.5 million. Less than a month earlier, an entity known as Rook TX purchased a winning Lotto Texas jackpot ticket worth $95 million. That April 2023 win has drawn criticism from Texas lottery players and state lawmakers who have called the event a money laundering scheme. 

Reed is seeking “recovery of funds fraudulently and illegally obtained” by the defendants in the case.

The lawsuit alleges four lottery retailers, Lottery.com, Lottery Now, Inc., ALTX Management, LLC, and Qawi and Quddus, Inc., worked with a sports gambling company, Colossus Bets, to orchestrate a bulk purchasing event where they bought more than $25 million worth of lottery tickets that covered nearly all possible number combinations to almost guarantee a jackpot win. …

Reed’s lawsuit claims the legal avenues for purchasing a lottery ticket are far too slow to print that many tickets in that amount of time. The suit claims the retailers “used custom-designed software, loaded onto smartphones, to generate a system of counterfeit QR codes that tricked the state-approved Texas Lottery terminals into recognizing the codes as if they had been generated by the Texas Lottery Commission’s authorized mobile app.”

As a former finance guy, I am tempted to say “no, figuring out the operational details to buy every lottery ticket is a shining example of American ingenuity, and they should be rewarded for it.” But as a former lawyer, I am even more tempted to say “no, figuring out that you can sue the previous lottery jackpot winner for reducing your lottery jackpot is the real shining example of American ingenuity, and this guy should really be rewarded for it.” I might have to go to Texas to cover the trial.

Trump’s Trade Math Ignores a Major Export: American Services. The $40 Billion Startup Mystery Shaking Up Silicon Valley. Justice Department Scales Back Crypto Enforcement. UBS’s Colm Kelleher slams ‘extreme’ plan for 50% more capital. Prada Strikes $1.4 Billion Deal for Versace. Buffer ETFs are working. Hedge Fund Veteran Set to Lead Trump’s Sanctions Strategy. ICE Chief Says Deportation System Should Run Like Amazon Prime. Musk’s Doge fired self-drive car safety experts at agency that regulates Tesla. Flashbacks to Covid Send Americans Rushing to Stock Up. “You have lawyers in big firms getting $20mn a year. God bless them. I was overpaid, too.”

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!