Money Stuff: There’s a GameStop Treasury Company

Well. The other day, I wrote that the “crypto treasury strategy” trade is very good, but saturated. The US stock market will pay $2 for $1 worth of Bitcoin, so lots of companies have acquired Bitcoin to make their stock go up, but the returns are diminishing. Companies have moved into other cryptocurrencies, Ethereum and Tron and Solana and BNB and Trumpcoin, but they are running out of open space. I wrote:

If you want to do this trade now, you might try something else. “Maybe the US stock market will pay $2 for $1 worth of _____,” you think, where what goes in the blank is (1) big and attention-getting, (2) relatively liquid (so you can buy it incrementally as money comes in and mark it to market as the price goes up), (3) somewhat crypto-adjacent in spirit but (4) not a big obvious crypto token.

So we talked about a company that pivoted to holding gold: not crypto, but vaguely spiritually aligned. I speculated about other possibilities:

I feel like someone should try a “MicroStrategy treasury strategy.” Like, you take over a small public company and you announce that you are raising $200 million to buy MicroStrategy Inc. stock. Maybe that would be worth $400 million?

I’m an idiot. There is one obvious answer, and I am ashamed that I missed it. Quantum BioPharma found it:

Quantum BioPharma Ltd. (NASDAQ: QNTM) (CSE: QNTM) (FRA: 0K91) (Upstream: QNTM) (“Quantum BioPharma” or the “Company”), a biopharmaceutical company dedicated to innovative therapies for neurodegenerative disorders, today announced the purchase of 2,000 shares of GameStop Corp. (NYSE: GME) to hold on the Company's balance sheet as a strategic investment. This move aligns with Quantum BioPharma's ongoing commitment to combating market corruption and enhancing shareholder value through prudent financial strategies and advocacy against manipulative trading practices.

Oh sure, sure, sure sure sure. Two thousand shares of GameStop is like $50,000, which is not going to move the needle even on a $72 million public company. The stock was down this morning, but good effort!

The simple way to think about exchange-traded funds is that “an ETF is a mutual fund that doesn’t pay taxes.” If you own shares of a stock mutual fund, that fund will sell stock from time to time (to meet redemptions or change its holdings), and if it sells appreciated stock it will incur capital gains taxes. As a shareholder of the fund, you will have to pay your portion of those taxes: Even if you don’t sell your shares, the fund’s trading creates tax liability for you.

ETFs are, for the most part, not like that. ETFs have found ways to never sell stock [1] : If an ETF needs to get rid of stock, it will not sell that stock for cash. It will do an “in-kind redemption”: An “authorized participant” (a trading firm, like a bank or Jane Street) will give the ETF issuers shares of the ETF, and will get back some basket of the underlying stock; the AP will then sell that stock. Under US tax law, this transaction — the ETF’s in-kind exchange of its own shares for its underlying holdings — is not a tax realization event, so the ETF’s holders don’t have to pay taxes. (Technically the taxes are deferred, and the holders eventually pay them when they sell their shares.) There is nothing obvious or inevitable about this — you could imagine taxing these in-kind trades — but that happens to be the rule, and the US ETF industry has been built up around it. 

The simple case here is when an S&P 500 index ETF needs to meet redemptions. Investors want to redeem, so they sell their shares of the ETF to an authorized participant. The AP then hands the shares to the ETF issuer, which cancels them and hands back a similar-sized basket of the 500 underlying stocks to the AP, which sells them. The AP makes a tiny bit of money on the trade (it sells the basket of 500 stocks for slightly more than it paid the investors for their ETF shares), the ETF shrinks its portfolio to match its shrinking investor base, and the investors get cashed out for their shares. Everything works as you’d expect.

The more complicated case is when an ETF wants to do a trade: It’s not getting rid of stock to meet redemptions, but it wants to sell its holdings of Stock X and buy Stock Y. [2] ETF issuers realized that they can do that via in-kind creations and redemptions; the trade is:

1. The ETF provider tells an authorized participant what it wants.
2. The AP buys some Stock Y and delivers it to the ETF as a “special creation basket,” getting back a similar amount of ETF shares. 
3. They wait a bit.
4. Then the AP delivers the ETF shares back to the ETF and gets in exchange a “special redemption basket” consisting of Stock X.
5. The AP sells the Stock X.

The ETF has traded its Stock X for Stock Y, but without ever handling cash. Only in-kind redemptions, no tax realization event, no taxes for its investors. This trade is called a “heartbeat,” and we discussed it in more detail back in 2019.

Back then, though, the technology was really in its infancy. An ETF was a mutual fund that doesn’t pay taxes, but with an emphasis on the mutual fund part: The point of most ETFs was (1) to do normal mutual-fund stuff and (2) not incur taxes on trades or redemptions. 

But once you realize “we have a technology to not pay taxes,” all sorts of other possibilities open up to you. If you have any sort of tax problem, you might think “can an ETF fix this?” So we talked a couple of weeks ago about an ETF that avoids taxes on dividends. Ordinarily, if you own a mutual fund or even an ETF, and the underlying stocks pay dividends, you get your share of the dividends and you owe taxes. [3] But an ETF issuer realized that it could heartbeat that problem away: “Sell” stocks right before their ex-dividend date (via in-kind redemptions), “buy” them back right after (via in-kind creations), and avoid the taxable dividend. [4]

What I said about this ETF was: You can’t do this. If you sold stock before the ex date, you’d incur taxes. This is a trade that is not available to normal humans, but that is available to ETFs because of their heartbeat technology.

The most general-purpose use of the technology might be: 

• You are a person.
• You own a ton of one particular appreciated stock: Perhaps you were an early employee at a tech company, and now it has gone public and you have millions of dollars of your employer’s stock with a very low tax basis.
• As part of your retirement planning and risk management, you would like to sell that stock and buy a diversified portfolio (the S&P 500, Treasury bills, whatever).
• If you sell the appreciated stock yourself, you will have a big tax bill.
• Can you contribute the stock to an ETF, and then have the ETF heartbeat it away and leave you with a diversified portfolio?

Like the ideal use here would be for issuers to create custom ETFs for each person’s stock, and then heartbeat the stock away. [5] Bloomberg’s Justina Lee reports:

The Twin Oak Active Opportunities ETF … debuted in February without much marketing. Yet it launched with nearly $450 million in assets and almost immediately received a $99 million inflow — amazing numbers for a virtually unknown ETF.

The day after that big inflow, there was an almost identical outflow. A similar pattern repeated in the next two trading days, and then again in the two days after that. And as all that capital rushed in and out, a change took place within TSPX that offers a vital clue as to its true nature.

Having launched with just five positions — around $245 million across three fixed-income ETFs, a $99 million stake in Snowflake Inc. and $92 million in Datadog Inc. — TSPX swiftly replaced its two equity holdings with a roughly equivalent slice of a simple S&P 500-tracking fund. ...

TSPX is one of a breed of exchange-traded funds created via what’s known as a 351 conversion. It’s a tactic to help rich investors minimize capital gains tax liabilities — one of a plethora of strategies that form Wall Street’s flourishing tax-optimization complex.

In a 351, a fund is seeded with appreciated assets before launch so that after listing it can use an infamous ETF industry loophole to rebalance without realizing a taxable gain, usually harnessing artificial flows to do so. While they’ll often retain their original strategy afterwards, some — like TSPX — use the process to significantly alter the portfolio without incurring taxes.

Yeah I mean it’s a good service! If you own $100 million of zero-basis stock and sell it, you will pay $20 million of capital gains taxes. Some wealth manager and ETF issuer will happily go through the work and complexity of setting up an ETF for you, taking your stock and heartbeating it into an S&P 500 fund; that all costs money, but not $20 million, so it’s a good trade for them and you. [6]  (Less so for the Internal Revenue Service.) If you own, like, $500,000 of low-basis stock, probably nobody is going to offer you this trade yet — the tax savings aren’t worth enough to pay for the complexity — but one day we’ll get there.

An important business for a bank is making collateralized loans to institutions and wealthy individuals. There is a range of collateral the banks will accept, and different sorts of collateral will have different processes. If you come to a bank and say “I have $100 million of Treasury bonds and I want to borrow some money against them,” you will probably be sent to the repo desk, which will quickly do a very standardized transaction, with a low interest rate and a high loan-to-value ratio. If you say “I have a $100 million portfolio of stocks and derivatives,” you will be sent to the prime brokerage division, which will be slightly slower, more bespoke and more expensive. If you say “I have a $100 million collection of blue-chip modern art,” they’ll find you some art lender in the private wealth management division, who will spend more time evaluating the portfolio, and who will probably give you less money and a higher interest rate. If you say “I own a private business that I think is worth $100 million,” you’ll be sent to a commercial loan officer who will do a lot of due diligence and lend you less than $50 million at a pretty high rate.

If you say “I have a collection of sneakers,” or data centers or Beanie Babies or port facilities or yachts or speculative lawsuits or rare wine or music royalties or cool rocks or really any other thing that people value and trade, and the collection is valuable enough, some banker will sit down with you and come to a valuation and lend you some money against it. That’s just what banks do. The interest rate and the loan-to-value ratio will depend on the bank’s level of comfort with the category, but for most things there is a price. Not everything. If you come to a bank and say “I have a collection of cocaine with a street value of $100 million,” the bank will absolutely not lend you money against that collateral. But most things.

And then if your stuff loses value, (1) you might not pay back the loan, (2) the bank will have to sell your collateral, (3) it might not get back the full amount of the loan and so (4) it might lose money on the trade. It happens! At some theoretical level, this means that the risk of the banking system is tied to the market price of whatever banks are lending against. But it’s a matter of degree, of concentration and correlation. Banks lend a lot of money against Treasuries and stocks, so if the prices of Treasury bonds or stocks move dramatically, that can cause problems for the banking system. If the bottom fell out of the yacht market, a few banks might have embarrassing losses, and a few yacht bankers would lose their jobs, but for the most part everything would be fine. It’s not like yacht loans make up a big percentage of the aggregate balance sheet of the banking system. You might convince a bank to lend you some money against your collection of medieval armor, but nobody is going to go around worrying about the systemic risk to the banking system if the medieval armor market collapses, because there are just not very many of those loans.

Anyway:

JPMorgan Chase is exploring lending against clients’ cryptocurrency holdings, in the latest sign that the biggest US banks are endorsing the move of digital assets into the mainstream.

The policy would mark a big shift for JPMorgan’s chief executive Jamie Dimon, who eight years ago branded bitcoin a “fraud” that would “eventually blow up” and was only useful for drug dealers and murderers.

JPMorgan could start lending directly against crypto assets such as bitcoin and ethereum next year, according to people familiar with the matter, who cautioned that the plans were subject to change. JPMorgan declined to comment. 

The move would underscore the extent to which big banks, and the regulated financial industry more broadly, are opening up to closer interaction with cryptocurrencies.

At some level, meh, yes, if you have a $100 million collection of anything, some JPMorgan banker will want to give you a loan against it, even if Jamie Dimon is not particularly enamored of your particular thing. (He’s not buying it; he’s making a loan.) At another level, in five years, will some biggish bank (not JPMorgan!) have a highly concentrated book of loans against crypto? Maybe? Is that a potentially worrying systemic linkage between crypto markets and the banking system? Kind of?

Elsewhere, Coinbase Global Inc. announced yesterday that it’s doing perpetual futures in the US:

For years, U.S. crypto traders have looked on as their international counterparts utilized one of the most popular tools in the digital asset marketplace: perpetual futures. These innovative derivatives, offering higher leverage while eliminating the need to navigate monthly expiration dates, dominate 90% of global crypto trading volumes. Due to a complex regulatory landscape, they remained just out of reach for traders in the United States. Until now.

We are thrilled to announce that U.S. customers, starting July 21, 2025, can trade CFTC-regulated perpetual futures via Coinbase Financial Markets (CFM), bringing one of the world's most traded derivatives to a secure and trusted crypto platform. …

U.S. traders can now access enhanced capital efficiency and amplify their market positions within a secure, regulated framework. For our newly launched crypto perpetual futures, you can trade with up to 10x intraday leverage.

That doesn’t exactly say “now Coinbase will make 90% LTV loans against Bitcoin,” but it doesn’t exactly not say that either. [7]

Floating-rate loans in the US used to be based on Libor, the London Interbank Offered Rate; now they are based on SOFR, the Secured Overnight Financing Rate. Libor was, roughly, the interest rate that big banks paid on unsecured short-term loans; SOFR is, roughly, the interest rate that big banks pay on short-term secured loans collateralized by Treasury bonds. From first principles, SOFR should be lower than Libor: Lending money with no collateral is riskier than lending money secured by Treasuries, so you should charge more. And in fact, when they coexisted (SOFR started in 2018; Libor ended last September), SOFR was normally lower than Libor. Canonically the typical spread between 3-month Libor and 3-month SOFR was 0.26161%.

Floating-rate loans are normally priced at some fixed spread to the floating reference rate: In 2015, a company might have had a loan with an interest rate of “L+300,” meaning that each quarter it would pay interest of that quarter’s Libor plus 3.00%; in 2025, a company might have a loan with an interest rate of “S+325,” meaning that it pays SOFR plus 3.25%. The spread for any particular loan will depend on the company’s creditworthiness, the state of the market, the structure of the loan, the negotiating dynamics, etc. But as a very general matter, you would expect spreads over SOFR be higher, ceteris paribus, than spreads over Libor. The transition from Libor to SOFR didn’t change anything structural about the costs of loans: If the right cost for a company’s loan this month is 7%, then the right cost is 7%, which you can express as “SOFR plus 272” (SOFR is 4.28% today) or “Libor plus 246” (if, counterfactually, Libor existed and was 4.54%). You need to add a higher spread to a lower number to get the same answer.

The transition from Libor to SOFR was long and drawn-out, and one aspect of it was that a lot of floating-rate loans with interest rates like “Libor plus 300” had to be renegotiated to say “SOFR plus ______.” Plus what? Well, it’s a matter of negotiation. The lenders say “it should be SOFR plus 350, because SOFR is lower than Libor and we want more money”; the borrowers say “it should be SOFR plus 300, because it’s just a change of risk-free rate and we want to pay less.” And then you negotiate. There was not exactly a correct answer, though there kind of was: The 0.26161% number that I cited above was based on average actual trading spreads, published by Bloomberg and endorsed by the International Swaps and Derivatives Association and the administrator of Libor. You could do worse than saying “SOFR plus 326,” but if a lender said “SOFR plus 326,” a borrower could quite rationally have replied “what about 305?” 

It is quite a mess to do all of these negotiations one by one, with each borrower negotiating with each lender for each loan. And many loans are syndicated, meaning that each borrower had to negotiate with many lender, some of whom were collateralized loan obligation managers with their own investors. You could imagine some more organized and coordinated process, where some central administrator announced “Libor now means SOFR plus 26 basis points in every contract,” and then that automatically happened. Government agencies and benchmark administrators did have some power to do that, and there were various partial efforts in that direction.

On the other hand, if like a dozen CLO managers got together and said “we are all going to agree that Libor now means SOFR plus 26 basis points, and negotiate that into all of our loan agreements,” that would be (1) efficient but also (2) an antitrust violation, oops:

The US Justice Department is conducting a criminal antitrust investigation into whether some investors in collateralized loan obligations colluded to bolster their positions as markets transitioned away from the scandal-plagued London interbank offer rate in early 2023, according to people familiar with the matter. ...

In the last few months of 2022 and early 2023 — shortly before the final phaseout of Libor — a flurry of companies in the leveraged loan market rushed to switch the benchmarks on their debt. Often, they tried to exclude an adjustment that was meant to compensate investors for the fact that the Secured Overnight Financing Rate — the debt’s new benchmark — consistently printed below Libor.

CLO managers, who repackage leveraged loans into bonds of varying risk and size, saw how some companies were about to reap benefits during that transition if that additional spread wasn’t added, as it lowered the interest the companies paid. The most junior holders of the bonds they issued, also known as equity, stood to lose millions as they get paid last after every other investor in the bond has received their payments. ...

Antitrust law bars competitors from colluding for economic gain. Because each CLO investor is a separate entity, it could potentially be illegal for them to agree with each other on the financial terms for an investment.

I often find myself reading about alleged antitrust violations and thinking “no, of course they should have colluded, that was just more efficient for everyone.” If everyone agreed on how to replace Libor, that really would be more efficient than renegotiating each loan individually! But not allowed.

I wrote yesterday about what a meme-stock company should do with its meme status. If your stock is temporarily trading at a high price for no good reason, I wrote, you should (1) sell stock into that demand and (2) use the money you raise to make your company permanently valuable. I suggested that the best way to do this, in 2025, is to sell stock to buy Bitcoin, because (1) that’s very easy and (2) investors love it. There are other approaches, though; perhaps the most obvious is to sell stock to build a good enduring business. This one seems much harder to me, but if you run a meme-stock company you might be tempted. “The market thinks my company is good,” you think, “so probably I have unusual insight into capital allocation, so I’m going to build the best movie-theater-and-gold-mining company the world has ever seen.” 

There is a literature. Here’s a fun 2024 paper on “Retail Trading Frenzies and Real Investment,” by Clifton Green and Russell Jame, finding that (1) meme stock companies do sell stock to invest in building their businesses but (2) it’s hard and they’re not especially good at it:

Retail frenzies strongly predict equity issuance and increased investment, with the relation strengthening in the post-zero-commission era. The increased investment is associated with incrementally lower future returns and more negative earnings surprises, and it is concentrated among financially constrained firms led by lower-ability managers. Overall, our findings suggest that retail frenzies adversely impact the real economy by relaxing financial constraints, leading to overinvestment by less-skilled managers.

One lesson here is that a meme-stock craze is not a particularly efficient allocator of capital, oh well. Another possible lesson is that crypto solves this: If, instead of selling stock to make bad business investments, meme-stock companies sell stock to buy Bitcoin, perhaps that does not “adversely impact the real economy”? I suppose there is a third possible lesson along the lines of “perhaps when meme-stock managers sell stock to buy Bitcoin, that’s not a good idea,” but never mind that.

The Private Equity Boom Is Leaving Midsize Players Behind. SoftBank and OpenAI’s $500 Billion AI Project Struggles to Get Off Ground. McKinsey Changes How It Elects Its Leaders to Avoid Succession Dramas. Startup Claims Its Green Steel Will Be Cheaper Than Regular Steel. Mike Lynch estate and business partner must pay £740mn to HP Enterprise, court rules. BlackRock Restricts Use of Company Devices for China Travel. UniCredit Wins More Time for Banco BPM Bid as Regulator Acts. Billionaire Fired by His Family Trust From Poland’s Cyfrowy. Fusion energy start-up claims to have cracked alchemy. “Heavyweight is a free, online, and open-source tool that lets you give any complaint you have extremely law-firm-looking formatting and letterhead.”

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!