Money Stuff: Strategy Sells Perpetual Strife

Strategy, the company formerly known as MicroStrategy Inc., [1] is in the business of selling securities to raise money to buy Bitcoin. The securities that it has sold to buy Bitcoin include:

1. Its common stock, which trades at a premium to the value of its underlying Bitcoin, making this a very nice trade for Strategy and a somewhat perplexing one for the people buying the stock. [2]
2. Convertible bonds, bonds that can be converted into common stock, which benefit from the extreme volatility of Strategy’s stock: Convertible arbitrageurs make money as Strategy’s stock bounces up and down, which it does a lot. There is also a simpler investment case for the convertibles: If you buy a Strategy convertible and Bitcoin goes up a lot, you get exposure to Bitcoin’s rise, but if Bitcoin goes down a lot you (maybe?) still get your money back.
3. Convertible preferred stock, which is sort of like a convertible bond: You give Strategy money, it pays you fixed quarterly interest (technically a dividend), and if the stock goes up you can convert your preferred stock into common stock.

What’s missing from that list? An obvious one is straight bonds — debt instruments that can’t convert into common stock — but “lend money against our pot of Bitcoins” is less appealing as a pitch than the stock or convertibles are. (I once wrote: “My gut sense is that the one-sentence intuitive credit analysis of MicroStrategy — ‘it has a huge pot of Bitcoins, has borrowed significantly less than the market value of that pot in the convertible market, and has like a $90 billion equity cushion’ — is just fine for convertible investors in a way that it might not be for, like, ratings agencies or traditional credit investors.”)

Another possibility is straight preferred stock, an instrument that (1) is junior to bonds (but senior to common stock), (2) never has to be paid back, (3) pays a relatively high cash dividend and (4) is not convertible into common stock. Strategy has been teasing preferred stock for a while, and when I first saw that I assumed it meant non-convertible preferred. I was puzzled, writing:

I have to say that perpetual preferred stock of a software company is not exactly a traditional product; there are not a ton of, like, mutual funds devoted to holding that sort of paper. But I guess a mezzanine claim on a giant pot of Bitcoin — junior to the convertibles, but senior to a giant pile of common stock — is a decent fixed-income investment?

But in fact Strategy went and did an odd convertible preferred, one with a very high conversion price, which it called a “perpetual strike preferred stock.” I thought that was kind of weird too: A convertible preferred with an 8% coupon and a 900% conversion premium is an unusual instrument — it doesn’t give you the equity exposure you normally want in a convertible — and in fact it struggled to find buyers. The name is also odd — I had never heard of “perpetual strike preferred stock” before, and I used to underwrite convertible preferred stock for a living — but you could parse it. “Perpetual preferred stock” is a common term for preferred stock (since it never has to be paid back), and is sometimes used to distinguish regular (non-convertible) preferred from convertible preferred; “strike” is a common term for the exercise price of a convertible instrument. So a “perpetual strike preferred stock” means something like “a convertible preferred stock with a strike price that is so high you might never convert.”

An alternate parsing might be “if you buy this preferred stock we will constantly punch you,” but at the time that did not occur to me.

Anyway Strategy announced last week that it might sell as much as $21 billion more of that perpetual strike preferred stock which, sure, okay, why not. But now here’s this:

Strategy™ (Nasdaq: MSTR; STRK) today announced that, subject to market and other conditions, it intends to offer, in a public offering registered under the Securities Act of 1933, as amended (the “Securities Act”), 5,000,000 shares of Strategy’s Series A Perpetual Strife Preferred Stock (the “perpetual strife preferred stock”).

Strategy intends to use the net proceeds from the offering for general corporate purposes, including the acquisition of bitcoin and for working capital.

The perpetual strife preferred stock will accumulate cumulative dividends at a fixed rate of 10.00% per annum on the stated amount. 

This does appear to be a non-convertible perpetual preferred stock. [3]  I have no special technical knowledge about how to parse the name; the first four times I read it I thought it was a typo. It is not. So I assume it’s the obvious parsing: “If you buy this preferred stock you will constantly be engaged in bitter struggle and conflict.” Seems right. I could not love this company more.

We talked recently about a Texas lottery arbitrage. The way Lotto Texas (and many other important US lotteries, including Powerball and MegaMillions) works is that you pick some numbers, there’s a drawing of random numbers, if you match every number in the drawing you win the jackpot, and if no one wins the jackpot it rolls over to the next drawing. So your chances of winning are fixed — for Lotto Texas, they are about 1 in 25.8 million; for Powerball they are about 1 in 292 million — but the amount that you can win varies. As the jackpot grows, the expected value of buying a lottery ticket might become positive: If the Lotto Texas jackpot is $100 million, the expected value of a 1-in-25.8-million chance at $100 million is greater than the $1 cost of a ticket. [4]

That’s not exactly right, because if two or more people match every number, they will split the jackpot. This is not a huge worry if you are buying a $1 lottery ticket — live a little! — but it is if you are doing the obvious lottery arbitrage, which is buying every ticket. If the Lotto Texas jackpot is $100 million and you spend $25.8 million buying every possible number, then you will definitely win the jackpot, but if you split it four ways you’ll lose money. And other people might notice this too! If five people each spend $25.8 million buying every possible number, they will all definitely lose money. [5]

The actual problem here turns out to be: Is there some reason to think that (1) you can buy every ticket and (2) no one else can? There are two components to this analysis. One is operational: Can you actually, like, go around to convenience stores to get every ticket printed out in time for the drawing? This is meant to be hard — Lotto Texas’s rules discourage it — but there are examples; someone figured it out with Lotto Texas, and there is a famous earlier example in Massachusetts. 

The other is financing: If you want to win the $2 billion Powerball, you can buy every ticket for $584.4 million (tickets are $2), but you need $584.4 million. Where are you going to get that kind of money? It would be pretty amazing if Jeff Bezos won the Powerball just for fun, but realistically the people with the operational skills and, like, desire to do this trade do not have $584.4 million lying around. [6]  You’d need to raise a fund.

Here’s a fun post from anonymous blogger “No Dumb Ideas” proposing a hypothetical lottery exchange-traded fund, why not:

Introducing: $LUCKY, the ETF that buys lottery tickets. ...

$LUCKY takes a pool of money (the assets under management, or AUM) and puts it all into short term treasury bonds. When the Powerball hits a threshold — call it $584m, the cost of buying every combination — $LUCKY spends 10% of its assets on lottery tickets. If it loses, it waits for the next jackpot. If it wins, $LUCKY holders get paid. Finally, you can get lottery exposure in your IRA.

The basic problem is that spending 10% of assets on lottery tickets is risky — you have a high likelihood of losing 10% of the fund — unless you buy a ton of lottery tickets, in which case you have a high chance of winning the lottery. “Using our 10% rule, we get a 100% chance of winning the Powerball at $5.8b in AUM,” which I guess returns 10%. [7] “So scale really matters.” How do you get to that scale? How do you raise billions of dollars quickly? “$LUCKY may be the most memeable ETF of all time,” argues the post, and, maybe? 

The obvious risks are:

1. Ehhhhh the US Securities and Exchange Commission is not going to like this, though who knows these days.
2. The lottery authorities might not like this, and you do have to address the operational problem of actually buying every number.
3. You might split the jackpot: If this works for you, it could work for someone else. “With two players, it would be a race to see who can white knuckle their way through split prizes long enough for the other fund to give up. Unfortunately, there’s only room for one $LUCKY in this market.”

I don’t think this will happen; this is just a joke or a thought experiment. But … I kind of think we are about two years away from a sports gambling ETF? Sorry sorry sorry sorry sorry sorry sorry, a “sports event contract” ETF. There are filings for memecoin ETFs right now, and last month Robinhood Markets briefly launched (and then withdrew) a Super Bowl event contract. “Robinhood’s mission is to democratize finance for all,” said Robinhood at the time. “With an emerging asset class like event contracts, we recognize an opportunity to better serve our customers as their interests converge across the markets, news, sports, and entertainment.”

I made fun of that language — sports betting is not an asset class! — but clearly Robinhood is the future and I am the past and financial markets are for betting now. Why shouldn’t all the technology of finance be applied to emerging asset classes like gambling? Put lottery tickets in ETFs, sure.

Investment bankers work on deals (mergers, stock offerings, etc.) and pitches (meeting with potential clients to get them to do deals). Senior investment bankers have a list of companies they cover, and they make their own work: They call up companies and say “I’d love to come in to talk to you,” if the company says yes then they put together a pitch, and if the pitch works then they have a deal. All pretty organic.

Junior investment bankers do not bring in their own work; they do the work created by senior investment bankers. Sometimes this also works pretty organically: The senior banker has a few junior bankers on her team, and when she schedules a pitch or lands a deal, her juniors work on it for her.

But for junior analysts in big groups at big banks, this is not the norm. Instead, a group will have a bunch of senior bankers and a bunch of junior bankers, and when a senior banker creates some work, she will pull some junior bankers from a centralized pool to do the work for her. Each junior banker will work on several pitches and deals at a time, for several different senior bankers. And there is a person, generally called a “staffer,” who coordinates this: The senior banker will call the staffer and say “I need an analyst to build me a merger model,” and the staffer will look at his list of analysts and say “I see Alice has some free time” and hand the work off to Alice.

The staffer is traditionally a midlevel investment banker, and it has always seemed to me that it’s a terrible job. It doesn’t seem particularly fun to keep track of junior analysts’ abilities and availability, and it takes away from the time that you would otherwise spend actually doing investment banking. Senior bankers will always be mad at you, because they will call you to say “I need three analysts to format a pitchbook” and you will have to say “everyone is super busy, could you live with two?” Junior bankers will always be mad at you, because you will constantly be calling them at midnight to say “can you make a new pitchbook by 8 a.m.?” 

The consolations for the staffer are that (1) it’s usually a short-term job and (2) it’s probably good training: Being a staffer is often a midlevel banker’s first exposure to management, and it’s a way to prepare the staffer for bigger roles in the future. Ideally the staffer, in addition to being the staffer, is also a good banker, so the senior bankers respect him (and accept his pushback when they demand too much) and the junior bankers respect him (and don’t push back too much when he calls them at midnight). And then he gets promoted out of the role and goes on to bigger and better things and in 20 years, when he wins the power struggle to become CEO of the bank, thinks “everything I needed to know about managing egos I learned in my year as a staffer.”

I don’t know, that’s roughly the theory. There are ways for it to break down. The staffer, being ambitious, might be too keen to please the senior bankers (who will decide his promotions) and work the junior bankers too hard. Or the staffer, being fairly junior himself, might be ignored by the senior bankers: A managing director might just walk into the bullpen, point at an analyst and say “you, LBO model, now” rather than bother with the staffing system.

If you are worried about overworked junior bankers, I suppose you might try to tweak this system. The Wall Street Journal reports:

Bank of America says it is stepping up its oversight of young bankers’ workloads, the firm’s latest change since the death of an employee working 100-hour weeks sparked an outcry across Wall Street.

The bank previously relied on midlevel employees on one-year rotations to assign work to its thousands of junior investment bankers. One of their responsibilities was supposed to be ensuring young bankers aren’t overworked. But the midlevel employees weren’t always provided incentive or equipped to enforce limits, according to people who work at the bank.

Bank of America has now made that role a permanent position for senior bankers, with the goal of strengthening supervision of young employees who may be facing heavy workloads. 

One risk is that “one-year staffer” is a prestigious position for an ambitious midlevel banker looking to become a manager, while “permanent staffer” might be kind of a dead-end position for a less-ambitious senior banker. If you’re a senior banker, the goal is to manage other senior bankers, not to manage first-year analysts.

Every stock trade involves an extension of credit. [8] If I go to my brokerage app and click “buy” and you go to your app and click “sell” at the same time, you might sell me some stock, but you won’t get the money — and I won’t get the stock — immediately. I will owe you the money, and you will owe me the stock, and we will settle up — actually delivering the money and the stock — on the next business day. This gives us time to get things all lined up: I can make sure I have the money, maybe moving it out of a bank account or converting it from euros to dollars, and you can make sure you have the stock, maybe recalling it from stock borrowers. 

What if I change my mind? What if I just decide not to show up with the money? (What if I lost all my money?) Well, in the US stock market, there is a pretty robust backup system called “clearing.” As soon as we agree on a trade, it gets submitted to a clearinghouse, the Depository Trust & Clearing Corp. [9] DTCC becomes our counterparty: I owe the money to DTCC, and DTCC owes it to you; you owe the stock to DTCC, and DTCC owes it to me. For the stock this works very well: DTCC holds all the stocks anyway, so transferring it from you to me just means updating DTCC’s own database. For the money there is more risk — if I don’t pay DTCC, someone still has to pay you — but the risk is pretty carefully managed. If I don’t pay DTCC, my broker (or its clearing broker) will have to pay, and DTCC takes collateral from brokers to make sure that they’ll be able to pay. (We once talked about how this got Robinhood Markets in trouble during the January 2021 meme-stock frenzy, when DTCC demanded a lot of collateral from Robinhood that it didn’t happen to have.)

There is a small weird technical problem: What if you and I agree on a trade, say at midnight, and we try to submit it to DTCC, but DTCC isn’t open? Well! DTCC can’t clear the trade — it can’t become our counterparty — until it opens and acknowledges the trade. What if something goes wrong at 2 a.m.? What if I message you and say “never mind, I’m fleeing the country with the money”? Does DTCC still owe you the money for your stock? It’s just a gap. Between midnight and when DTCC opens, there is a creepy residual credit exposure. Seems bad.

If we are trading in a dark pool run by a bank, I guess the bank takes the credit risk, but on a stock exchange it’s weirder. From a February New York Stock Exchange blog post:

We strongly believe exchanges should not match buyers and sellers when centralized clearing is not available because exchange clearing is fundamentally different than alternative trading systems (ATS) or other broker-to-broker clearing. When the clearing house (Depository Trust and Clearing Corporation or DTCC) is open, trades are submitted immediately, and counterparty risk is well defined. However, when DTCC is not open, the clearing firm absorbs counterparty risk. Clearing firms are specifically managed and regulated to serve this function; exchanges are not clearing firms and current market rules do not contemplate exchanges assuming counterparty risk by “holding” trades until the clearing house opens.

This is meaningful, even for small windows of time, as world events could create extreme volatility and volume. In such scenarios when the clearing house is closed, market participants trading on an exchange could have substantial exposure to an unknown counterparty, whereas on an ATS during such an event their exposure is to the ATS’s clearing firm.

The obvious solution is: DTCC is in the first instance mostly a computer, so keep the computer open later. Bloomberg’s Katherine Doherty reports:

The Depository Trust & Clearing Corp. will begin clearing equities trades 24 hours a day, five days a week by the second quarter of next year, the company said in a statement Tuesday. The move, which needs regulatory approval, will take place in phases over the coming months to support both off-exchange and on-exchange stock venues.

“As interest in near round-the-clock trading of US equities grows, we are meeting this demand by extending our clearing hours to support our clients and further strengthen the safety and soundness of the markets,” Brian Steele, a managing director and president of clearing and securities services at DTCC, said in the statement.

I suppose if you agree on a trade on a stock exchange at midnight, the settlement risk is particularly acute. (What if your counterparty says “I was drunk, I don’t want these shares”?) So you do want to make sure that it gets cleared immediately.

A standard worry about artificial intelligence is that there will be an inflection point where AI models get better at making AI models than humans are. Right now, highly paid researchers at big AI firms build large language models that are pretty good at computer programming, but eventually the LLMs will be better at building LLMs than the human researchers are, and they will just build themselves into self-improving superintelligent AI models that enslave us all. Something to look forward to!

In the meantime, though, what happens when AI models get better at selling AI models to enterprise customers than human salespeople are? The Financial Times reports:

Artificial intelligence start-up Anthropic is preparing to release a range of new features focused on business users, pitching its AI models at enterprise over seeking mass market adoption. ...

Internally, the company has been testing a prototype which it plans to “commercialise” and build into Claude in the coming months.

The feature examines workers’ calendars and prepares a report ahead of their meetings, analysing internal and external information about the client, such as who else they have met internally and previous notes about the company.

“Ideally, you’re a salesperson who walks into every one of those meetings as prepared as possible, but they don’t have time to do that report for every single company,” [chief product officer Mike] Krieger said.

“That’s a way in which the models do what they’re really good at: go off and do asynchronous work, think about the problem, and then create something consumable. Hopefully, that will be an accelerant for our sales folks.”

What if Claude is really really good at selling itself to business users? What if it is, like, unsettlingly good? What if every pitch is like “you should replace all of your executives with Claude and just let Claude take over your business,” but written so hypnotically that every customer says yes? It seems unlikely. Still it feels very Silicon Valley to worry that AI will replace humans when it becomes better than humans at software design; what if really you should worry about AI getting better than humans at business-to-business sales? 

Morgan Stanley Went Big on DEI, and No One Is Happy About It. Google Agrees to Buy Cloud Security Firm Wiz for $32 Billion. PepsiCo to Acquire Prebiotic Soda Brand Poppi for $1.95 Billion. JPMorgan Stock Traders Reel In Windfall as Trump Whipsaws Market. Hedge-Fund Momentum Bets Crater All at Once in Volatile Markets. A Secret Mortgage Blacklist Is Leaving Homeowners Stuck With Unsellable Condos. Saudi Fund Broadens Cash Quest to Meet MBS’ Spending Goals. Spain’s Biggest Bank Is Making a Big Push in the U.S. Spanish defence champion accuses General Dynamics of ‘destroying’ arms maker in battle for control. Short sellers make $16bn profit from Tesla’s share price plunge. UBS Braces for Most Complex Phase of Credit Suisse Integration. “European credit markets are trading at a premium to the US for the first time in more than three years.” Thames Water Says It’s Been Assessing Six Equity Proposals. Forever 21 Files for Bankruptcy for Second Time in Six Years. San Francisco Giants Sells 10% Stake to Sixth Street Partners. How Bribes Helped a Crime Ring Steal Thousands of iPhones From Porches. Equities in Dallas.

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!