Money Stuff: GameStop Doesn’t Have Enough Stock

Well! Yes! This is the point! Of everything! Like:

1. You do good business stuff;
2. Your stock goes up;
3. Now your stock is worth a lot of money; 
4. You use your valuable stock to buy things — executive talent, other companies — that help you do more good business stuff.

A virtuous and self-reinforcing cycle. Step 2 is not, perhaps, the most important part of all of this, but it’s not the least important either. At some point in early 2021 the mad scientists at GameStop Corp. realized, somewhat to their own chagrin, that their stock was going up quite a lot for reasons of its own. At first, they were terrified by this dark magic and declined to use it. Later, they tentatively started to use their stock-price powers to, like, hire Amazon.com executives. I am sure that I have somewhere joked that they would eventually use their stock to buy Amazon itself. They’re not quite there yet but man we are getting closer:

GameStop Corp. is proposing to buy eBay Inc. for about $56 billion in cash and stock, a bold attempt by Ryan Cohen to take over a storied e-commerce name several times larger.

The gaming retail chain offered $125 per share in cash and stock for the online marketplace, or about a 20% premium to its Friday close. GameStop, which built a roughly 5% stake in eBay, said it’s secured an initial, non-binding “highly confident letter” from TD Bank to provide about $20 billion of debt financing. In a memo to investors Sunday, Cohen’s company pledged to find some $2 billion of annual savings within 12 months of closing. …

The takeover bid follows the surprising ascent of GameStop, a chain of video game outlets that shrank its brick-and-mortar footprint after gamers increasingly bought software in digital stores. In 2021, it became the center of a retail-investor frenzy. …

Cohen is now proposing to take over a company roughly four times larger than the retail chain he operates. GameStop had a market value of $12 billion as of Friday. EBay was much bigger at around $46 billion, though the game retailer has about $9 billion in cash. The takeover offer is split evenly between cash and GameStop common stock.

Here is GameStop’s proposal, which envisions the combined company as, like, a hybrid online/in-person trading venue for collectibles? The presentation says that GameStop’s stores would give eBay “a national network for authentication, intake, fulfillment, and live commerce.” I do not understand business at all.

Here’s a question. What does it mean that, in this proposal, GameStop would “buy” eBay? The Wall Street Journal notes:

GameStop has around $9 billion in cash on its balance sheet to put toward a deal. It wasn’t immediately clear how it would come up with the rest of the money needed for a $56 billion acquisition. It is possible Cohen could tap outside investors, such as Middle Eastern sovereign-wealth funds, to back the deal, people familiar with the matter said.

And here’s a funny Business Insider article about how Cohen and Andrew Ross Sorkin talked past each other on “Squawk Box” this morning:

Sorkin asked Cohen to walk viewers through how the deal would work.

"Yeah, we'll see what happens," Cohen replied, drawing laughter in the studio.

"I hear you. I understand that," Sorkin replied. "I'm just trying to understand where the rest of the money would come from."

"It's half cash, half stock," Cohen said, before Sorkin pushed back.

"I'm just saying that that math doesn't get you to the price that you're offering," Sorkin said.

The video truly is wild, but I think the math is fine? The crude math is (1) $8 billion of balance-sheet cash plus (2) $20 billion of “highly confident” debt financing from TD plus (3) $28 billion of new GameStop stock gets you to $56 billion. Half cash, half stock. It all adds up. But I’d make a few points.

First, GameStop, sort of obviously, could not go out and borrow $20 billion itself. (Bloomberg tells me its estimated fiscal 2027 earnings before interest, taxes, depreciation and amortization are about $469 million, and it already has $4 billion of debt.) TD is “highly confident” that GameStop could borrow $20 billion against eBay: If it takes over eBay, the combined company would have more cash flow and be able to issue more debt. But this is not GameStop debt; this is a leveraged buyout where the new debt capacity would come from eBay.

Second, in some arithmetic sense, GameStop could issue $28 billion of new stock. The stock closed at $26.53 on Friday; if you divide $28 billion by $26.53 you get about 1.05 billion. So if GameStop issued 1.05 billion more shares, that would cover the stock portion of the purchase price. But:

1. If GameStop issued 1.05 billion more shares, would they be worth $26.53 each? Meh. The stock is down a bit today. GameStop has about 448 million shares outstanding, so this deal would more than triple its share count. “Any credible offer would require substantial dilution,” note Bloomberg Intelligence analysts Poonam Goyal and Sydney Goodman. In a sense this is like the debt: GameStop couldn’t issue 1.05 billion more shares against its own business without driving down the stock price a lot, but if it owned eBay, it would be much bigger and could support a larger stock market valuation. But the new stock that GameStop would issue would not really be “GameStop stock”: It would be stock in a combined GameStop/eBay in which eBay represents the large majority of the business.
2. Relatedly, that means that, in any deal, legacy eBay shareholders would own twice as much of the combined company as legacy GameStop shareholders. GameStop would “buy” eBay in the sense that (1) that’s what the press release says and (2) Ryan Cohen would be the CEO of the combined company. But eBay shareholders would own most of the combined company, which would be mostly eBay.
3. Under New York Stock Exchange rules, this would require a GameStop shareholder vote, for intuitive reasons. Ordinarily if a big company acquires a smaller one for cash, or for a small amount of stock, the acquirer’s shareholders don’t get to vote, but the target’s shareholders — who are giving up their company — do. But, here, GameStop is more or less getting acquired, so its shareholders get to vote. (As do eBay’s.)
4. Also GameStop can’t actually issue 1.05 billion more shares! It is often reasonable to assume that companies can freely print as much of their own stock as they want: Stock is just the fractions in which ownership of the company is divided, and the company can make up as much as it wants. But that is not exactly legally true. A company’s certificate of incorporation will normally authorize a large but finite number of shares, and the company can’t go around issuing more than that. In fact, GameStop has only 1 billion authorized common shares, and about 448 million are already outstanding, meaning that GameStop doesn’t have anywhere near enough authorized shares to pay $28 billion in stock for eBay. Oops!

That last point is not an insurmountable problem: GameStop’s shareholders could vote to authorize new shares. They’d have to vote to approve the deal anyway, so getting more shares authorized is not impossible. Still it’s odd that it’s not even mentioned in the proposal. GameStop is offering to buy eBay for cash it doesn’t have, and also for stock it doesn’t have. One gets the sense that GameStop did not entirely think this through. 

Incidentally. The Journal notes:

Cohen also has a potentially massive payday at stake if he can pull it all off. GameStop adjusted Cohen’s compensation package at the beginning of the year to give him extra incentive to boost the company’s market value and profitability. He stands to make as much as $35 billion in stock if certain criteria are met, including if its market value hits $100 billion. 

This sort of thing is not uncommon these days; it is often called a “moonshot” pay package and, like so many things, is largely Elon Musk’s doing. Instead of paying your chief executive officer a reasonably large amount of money for hitting reasonably ambitious goals, you pay her (1) not very much money plus (2) an extraordinarily large amount of money if she hits extraordinarily unreasonable goals. In 2018, Musk famously got a pay package that would be worth about $56 billion if he took Tesla Inc.’s market value from $59 billon to $650 billion; he did, and it was. He now has a similarly ambitious 2025 Tesla pay package that will be worth something like $1 trillion if he takes the market cap to $8.5 trillion; he also apparently has a pay package at SpaceX that will trigger on $6.6 trillion in market cap and a lot of data centers in space.

Lots of CEOs who would like to think of themselves as Musk-esque have persuaded their boards to give them similar pay packages, Cohen among them. In January, GameStop granted him a package of options contingent on hitting targets for market capitalization and earnings before interest, taxes, depreciation and amortization. If he hits all of the targets — $10 billion of “cumulative performance EBITDA” and $100 billion of market cap — within 10 years, he’ll get 171.5 million options worth, at those levels, about $35 billion. [1]

Now. Musk has in recent months added to SpaceX’s market capitalization by acquiring xAI (and X, formerly Twitter). There is occasional fun speculation that he might at some point merge Tesla and SpaceX. Cohen, as GameStop CEO, is trying to acquire eBay. Intuitively you might think: “Well, yes. These guys have enormous, entirely at-risk pay packages that depend mostly on size. They are incentivized to get a large market cap and large earnings rather than, say, a high stock price per share, or high earnings per share, or a high profit margin. If you have a $12 billion company earning $500 million a year, and you need to make it a $100 billion company earning $2 billion a year, [2] merging it with a $46 billion company earning $3 billion a year is a helpful shortcut.”

I think that intuition is approximately correct. The boards of directors who agree to these pay packages understand this, and they understand that it is in a sense a problem: Turning a $10 billion company into a $100 billion one organically (1) is probably harder than just merging with a $50 billion company and (2) certainly creates more shareholder value; just adding $50 billion of already-existing company to GameStop doesn’t really create any new value. The goal of the compensation package is to encourage value creation, not just indiscriminate acquisition. And so Cohen’s pay package says:

The Performance Hurdles will be adjusted by the Committee equitably and proportionately as determined by the Committee in a manner designed to preserve the economic opportunity provided under the Award, (a) higher to account for acquisition activity for which stock is provided as consideration; and (b) lower to account for a split-up, spin-off, dividend or other distribution (whether in the form of cash, shares, other securities, or other property) or divestiture activity, in each case, that could be considered material to the achievement of the Performance Hurdles, as applicable.

Though it doesn’t spell out how the hurdles will be adjusted, other than “equitably and proportionately.” Musk’s 2025 Tesla pay package is a bit more explicit:

In the event that Tesla acquires a business with a purchase price of more than $20 billion, any then-unearned Market Capitalization Milestone will be increased by the purchase price of such acquisition. Similarly, if the target of an acquisition transaction has Adjusted EBITDA (based on cumulative four consecutive quarters prior to the transaction) of more than $2 billion, the then-unachieved Adjusted EBITDA Milestones will be increased by such target’s Adjusted EBITDA. … These features of the 2025 CEO Performance Award are designed to prevent achievement of milestones based on acquisition activity that could be considered material to the achievement of those milestones.

But notice that that doesn’t quite solve the problem, because it is arithmetically easier to grow (say) a $60 billion company to $150 billion than it is to grow a $10 billion company to $100 billion. Assume Cohen is very talented and it is easy for him to grow his company’s market capitalization by 20% per year. If he starts at $10 billion, after 10 years he’ll be at about $62 billion: terrific, but not enough to meet all the milestones. But if, right out of the gate, he buys $50 billion of market cap, and then compounds that at only 10% per year, at the end of 10 years he’ll have a market cap of about $156 billion: enough to meet the milestone, even if it is adjusted from $100 billion to $150 billion to account for the acquisition. [3]

Or put another way: Let’s say in rough numbers that Tesla is worth $1.5 trillion and Musk gets maximally paid for getting it to $8.5 trillion, and that SpaceX is worth $2 trillion and Musk gets maximally paid for getting it to $6.6 trillion. Let’s assume that SpaceX’s pay plan adjusts the same way as Tesla’s. Right now, Musk needs to get to a total market capitalization of $15.1 trillion for his two companies ($8.5 trillion for Tesla, $6.6 trillion for SpaceX) to get his full pay packages. If they merged, loosely speaking, his Tesla pay package milestone would adjust up by the value of SpaceX, from $8.5 trillion to $10.5 trillion, and his SpaceX one would adjust up by the value of Tesla, from $6.6 trillion to $8.1 trillion. But notice they’re the same company now. Getting the combined company to $10.5 trillion would hit all the targets, $4.6 trillion earlier than if they were separate companies.

I don’t actually know how GameStop’s board would adjust Cohen’s pay package if GameStop buys eBay. (Or what would happen if Tesla and SpaceX merge, for that matter.) You could imagine doing it multiplicatively — “if you go from $10 billion of market cap to $60 billion by acquisition, the target is adjusted from $100 billion to $600 billion” — but not really. That seems unfair too; it’s harder to add $500 billion of market cap to a big company than it is to add $50 billion to a smaller one. My guess is that most moonshot pay packages would adjust for big acquisitions the way Musk’s does, by adding the acquisition to the target market capitalization in a way that seems roughly fair but that gives CEOs a powerful incentive to pursue bigness by acquisition.

Of course here I am assuming that these moonshot pay packages are intended to encourage organic value creation, and that assumption might be wrong. Maybe the boards, and the shareholders, actually just want bigness. Perhaps it would be fun for Tesla and SpaceX shareholders if their companies merged. Surely it will be fun for GameStop shareholders if GameStop takes over eBay; that is just the sort of random swashbuckling that GameStop shareholders should enjoy. GameStop shareholders should be pleased with this little adventure; of course Cohen’s pay package should encourage it. 

In the governance literature, this sort of thing — CEOs pursuing size rather than return on investment — is called “empire building,” and is considered a bad thing. But, come on. If you’re a shareholder of the Elon Musk complex, or of Ryan Cohen’s GameStop, don’t you sort of want your CEO to build an empire?

We talked a couple of weeks ago about sportsbooks. There is a popular misconception that sportsbooks set their betting lines to balance their action, so that they make a little bit of money no matter who wins a game. But in fact that’s largely not true: Sportsbooks set their betting lines to maximize profit, and if public bettors are systematically wrong — if they always overestimate the chances of the home team or the underdog or whatever — then sportsbooks will take the other side of that bet. They will take directional risk, if it’s good risk.

Modern sports betting increasingly happens not at sportsbooks but at prediction markets. In theory, in a prediction market, if you want to bet on the Mets you are betting against some other customer who is betting on the Nationals. In practice, there’s a decent and increasing chance that the person on the other side of your bet is a professional or semi-professional market maker, someone (or some algorithm) who makes two-sided markets, buying “Mets win” contracts at $0.40, selling them at $0.41 and collecting the spread.

I pointed out that financial market makers are a lot like sportsbooks, in that (1) the simple model is that they balance buying and selling, collect the spread and do not take any directional risk, but (2) in fact sophisticated well-capitalized market makers do often have fundamental value models and lean into one side or the other. If retail investors all want to get long the same dumb stock, a market maker might end up short that stock and happy with its position.

Also, we have talked from time to time around here about insider trading on prediction markets. Intuitively, if you are betting on the capture of Nicolás Maduro using inside information, you are betting against someone, someone who does not have inside information and therefore loses money. Loosely speaking the person you are betting against is probably a market maker. 

Thus market making in prediction markets consists of:

1. Sometimes buying at $0.40, selling at $0.41 and clipping a $0.01 spread for a few minutes of risk;
2. Sometimes making a considered fundamental bet that public bettors overstate the probability of an event happening, selling at $0.41, holding until resolution, and making a $0.41 profit when the thing doesn’t happen [4] ; and
3. Sometimes making that considered fundamental bet, selling at $0.41 to someone trading on inside information, and losing $0.59 when the thing happens and you have to pay out $1.

Here’s a fascinating paper on “Adverse Selection in Prediction Markets: Evidence from Kalshi,” by Robert Bartlett and Maureen O’Hara:

Using 41.6 million trades, we measure adverse selection in prediction markets. Adapting Kyle's λ and the Glosten-Harris decomposition, we show single-name markets exhibit greater informed price impact than broad-based markets. Despite this, effective spreads are only modestly wider, and market makers earn twice as much per contract. A frequency-magnitude decomposition resolves this puzzle: traders systematically overbet YES in markets that predominantly settle NO, generating a behavioral surplus that cross-subsidizes adverse selection. Adapting the VPIN toxicity metric, we find that one-sided order flow predicts maker losses in single-name markets but not broad-based markets. Our research suggests a new microstructure equilibrium concept for bilateral settlement markets.

They divide Kalshi contracts into “either single-name (referencing a specific company or individual) or broad-based (referencing macroeconomic aggregates, market indices, or asset prices),” and find that contracts referencing a specific company or individual have a lot of “informed trading,” which doesn’t necessarily mean “insider trading” but who knows: “Markets referencing outcomes known to a handful of insiders exhibit more informed trading, and that information is incorporated permanently into prices.”

That is bad for market makers in those contracts: If you sell the “Nicolás Maduro out” contract, it’s probably to a Special Forces soldier planning his capture, and you lose money. Why, then, would market makers trade those contracts, given that they’re probably trading against insiders? Intuitively the answer is that they’re trading against some insiders and also a lot of randos who are systematically wrong, which subsidizes the market makers (and thus the insiders):

Events trading at an implied 46% YES probability actually settle YES only 21% of the time, reflecting a miscalibration of 25 percentage points. Takers on single-name markets exhibit not simply a classical favorite-longshot bias but a pervasive YES optimism, most pronounced where outcomes are most contested. 

So market makers systematically fade public bettors, betting No where the public naively bets Yes, and make money. 

Incidentally, when we talked about market making on prediction markets a few weeks ago, I quoted a tweet from a guy who set up “a Polymarket bot that automatically buys ‘No’ for every non-sports market and holds to resolution.” “Right,” I said, “that’s (one-sided) model-based positive-expected-value passive liquidity provision.” It basically works!

One question is: Why make two-sided markets at all, betting “No” and “Yes” on all of these markets where the public predictably overvalues the “Yes” side? Bartlett answered by email:

The temptation is to think: if YES is overpriced, why not just sell YES all day? We ran the counterfactual. An offer-only maker (never bids YES) earns roughly twice as much in gross profit on single-name markets as the full-book maker, so the instinct is right that the sell side is where the money is. But the offer-only maker’s standard deviation is 3X higher, and her 20 worst markets are 100% events that actually happened. By refusing to bid YES, she concentrates her exposure on exactly the markets where informed YES takers show up knowing the event will occur. The full-book maker gives up half the gross for a 70% reduction in variance, because her YES bids accumulate cheap long positions from wrong-side NO takers that cushion her when events do happen.

Intuitively, the offer side is the revenue engine, and the bid side is the hedge. The maker who only runs the engine blows up on every TikTok-gets-banned or CEO-gets-fired market where someone on the other side knew the answer. The maker who runs both earns less per market but survives the tail across the portfolio.

Mostly the public is wrong, but sometimes it’s insider trading.

“Maybe the most important paper in economics,” I once wrote, “is the one about how people sometimes give themselves painful electric shocks just because that is an option that’s available to them.” They should teach that paper on the first day of Evil Business School. It might give you ideas for business models. One business model is: I give you something of value, and in exchange you give me money. This is probably the most common business model, and the most socially beneficial, but the problem with it from my perspective is that it is a lot of work for me to give you something of value.

Another business model is: I promise to give you something of value, and in exchange you give me money, and then I don’t give you anything of value. I talk about this model a lot around here — loosely speaking it is called “fraud” — but it has some big downsides for me too. For one thing, it’s a fair amount of work for me to trick you into giving me your money. Plus I might go to jail, etc.

A third business model is: I promise you something worthless, and in exchange you give me money. I suppose I talk about this model a lot around here too. It is called “crypto,” ahahahahahahahaha, no, kidding, sort of. I mean. If I can create some tokens that cost me nothing, and that promise you nothing, and then I can sell them to you for real dollars without technically lying to you, that’s a good business for me. What is it for you? Well. It’s an option that is available to you. If you’re sitting around, bored, thinking “you know what I could go for right now is a painful electric shock,” maybe you’ll buy my worthless tokens?

At Bloomberg Businessweek, Peter Robison and Vernon Silver report:

Tens of millions of people play [social casino apps] daily. High 5 Casino, Jackpot Party, Slotomania and similar apps are billed as just-for-fun diversions that can be played for free.

And they can be: You download the game at no cost and start with a small stash of coins. But you’ll almost certainly run out—because, as at every casino, the house always wins. Then you’ll be prompted through a stream of pop-ups to pay real money for more coins, to avoid waiting (maybe an hour, maybe all day) for the game to dispense more free ones. Even when you pay, and win, you can’t cash out. It’s the defining element of a social casino; the prize is the make-believe coins, and perhaps some dopamine.

“Perhaps Some Dopamine” would be a good name for a startup. Just an underrated explanation for all sorts of economic phenomena. Robison and Silver write about how people become addicted to these games:

Considering you can’t win real money, casino-style games generate an astonishing amount of actual revenue, more than $11 billion in 2025, according to the market intelligence firm Sensor Tower Inc. And actual pain: Some players have spent more than $1 million, contemplated suicide, mortgaged houses or divorced their spouses over the games.

Big spenders, a jury was told, were described within High 5 as “whales” or “monetizers.” When a staffer reported that one user—“good ol’ Patty… the whale with crazy monetization habits”—had bought several coin packages in a single day, another suggested sending “a box of wine or three” to her home to keep her tapping away at her favorite game, Ciao, Roma! “Thumbs up,” responded Patrick Benson, High 5’s vice president for integrated marketing. “I hope she sits down with a box of wine tonight and plays C. Roma all night.” High 5 declined to comment for this story.

Like if I just offered you the proposition “if you give me $1 million in real money, I will give you one trillion imaginary coins that can’t be redeemed for anything,” you would probably say no. But if I did it over time and packaged it in a colorful slot machine on your phone you might say “sure yeah whatever” and then we’re in business.

Jane Street’s Record Year Translates Into a $2.68 Million Payout Per Employee. OpenAI Finalizes $10 Billion Joint Venture With PE Firms to Deploy AI. Anthropic Unveils $1.5 Billion Joint Venture With Wall Street Firms. U.S. Senators Now Prohibited From Trading on Prediction Markets. Why Almost Everyone Loses—Except a Few Sharks—on Prediction Markets. Trump-Linked World Liberty Sues Justin Sun for Defamation. Trump Family Crypto Project Quietly Sold as Holders Got Stuck. Trump Media’s Latest Pivot Is a Leadership Shake-Up. Private Credit Could Spark Psychological Contagion, Barr Warns. Bill Savitt profile. Hotspan extension. The Mystery of ‘Human Meat’ for Sale on Temu. 

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!