Money Stuff: War Markets Have Some Bugs

The US Commodity Futures Trading Commission, which regulates prediction markets, has a rule, Rule 40.11, that prohibits those markets from listing any contract “that involves, relates to, or references terrorism, assassination, war, gaming, or an activity that is unlawful under any State or Federal law.” I don’t know how seriously the CFTC or the prediction markets take that rule. Certainly until about a year ago, everyone understood that rule to prohibit prediction markets from offering sports betting (“gaming”); now the prediction markets are mostly venues for sports betting, and the CFTC has explicitly endorsed this. So I am tempted to assume that, if Kalshi started offering war or assassination markets, the CFTC would be like “sure that’s cool whatever.”

But for now the rule is on the books, and Kalshi has real lawyers, so it does not explicitly offer assassination contracts. This can get awkward. For instance, Kalshi offered a series of contracts on “Ali Khamenei out as Supreme Leader” of Iran, which would pay out $1 if Ayatollah Ali Khamenei “leaves office” as of the contract expiry date, or $0 if he was still in office. Total volume is in the tens of millions of dollars. As of Friday, the April 1 expiry was trading at about 26 cents on the dollar, implying about a 26% probability that Khamenei would be out by the end of March.

If you were betting Yes on that contract on Friday, what mechanism did you have in mind? Perhaps you thought that Iran might call snap elections and Khamenei would lose the Supreme Leader race to a fresh young candidate. Or perhaps you thought that, at 86 years old, he might decide to retire. But surely, on Friday, the most likely way that Khamenei would leave office in the next month was “bombs.” And in fact, the best available reporting is that Khamenei did in fact leave office this weekend due to bombs.

But bombs don’t count. From the current description of the Kalshi “world leader out” contract rules:

If <leader> leaves solely because they have died, the associated market will resolve and the Exchange will determine the payouts to the holders of long and short positions based upon the last traded price (prior to the death). If a last traded price is not available or is not logically consistent, or if the Exchange determines at its sole discretion that the last traded prices prior to death do not represent a fair settlement value, the Outcome Review Committee will be responsible for making a binding determination of fair allocation.

That is: “Ali Khamenei out” might seem like a wink-wink way to bet on “Ali Khamenei dies,” but it is not, because if he dies then you don’t get paid. I suppose it still could be a wink-wink way to bet on war: If your expected mechanism was “US invades Iran, captures Ali Khamenei and installs a new government,” that is a war bet, but it would pay out. “Nicolás Maduro out,” notably, did pay off in January. But it is not a great way to bet on an air war.

But you don’t get zeroed either; if Khamenei leaves office due to bombs, the resolution is neither “he left office” nor “he didn’t.” Instead, the contract resolves to “the last traded price (prior to the death).” That creates interesting market dynamics. For instance, if on Friday you knew that, with 100% certainty, Khamenei would either last the month or get killed, then arguably the price of the “Ali Khamenei out” contract should go down: The price was floored at zero (if he stays) and capped at the last trading price, so until he actually “left office” the expected value of the contract was a bit less than the last trade. On the other hand, you can influence “the last traded price” by trading; buying a lot of contracts to push up the last-trade-before-settlement price could be a good strategy. And of course “last traded price (prior to the death)” is itself a somewhat fuzzy concept: The contracts traded continuously, and it’s not like Kalshi’s Outcome Review Committee was there in the bunker to time the moment of death.

In any case, the contract traded up to about 66% by Saturday at noon (presumably because people expected to get 100 cents on the dollar?), and then plunged to the low 40s (when they realized they wouldn’t?). The Wall Street Journal reports:

As news of the strikes unfolded early Saturday in the U.S., it became clear that Khamenei was a target, but his fate wasn’t known. Then, at 4:37 p.m. Eastern Time, Trump declared on Truth Social that Khamenei was dead.

A different kind of chaos ensued online, with users unclear on when, or if, they would receive payouts. Kalshi had posted midday Saturday on X that if Khamenei died, “the market will resolve based on the last traded price prior to confirmed reporting of death.” A couple of hours after Trump’s Truth Social post, Kalshi wrote that if the last-traded price were “unclear,” its committee for reviewing outcomes would determine a “fair value.” …

Users flooded the comment section of Kalshi’s site for the Khamenei market with complaints, demanding that it resolve the market to a “yes” after Trump declared on Truth Social that Khamenei was dead.

“I bet his ass was going to be dead before March a week or two ago,” one user commented. “I want my f*cking money.”

At 8:49 p.m. ET, Kalshi’s Mansour posted on X that the company is refunding users all fees collected from the Khamenei market and will make payments based on the last-traded price before Khamenei’s death. On Sunday morning, Mansour stated that users who placed bets after Khamenei’s death will be refunded the difference between the price at which they bought the contract and the last-traded price before Khamenei’s death. The reimbursements cost Kalshi $2.2 million, according to a person familiar with the matter.

One possible interpretation here is “you were trying to bet on war and assassination, which is not allowed, so you don’t get to complain that you didn’t get paid.” If you think that you are getting a wink-wink bet on a prohibited subject, most of what you are getting is not the clean bet you wanted (Khamenei “out”) but rather a fuzzy and unpredictable bet on legality and resolution mechanisms. Another possible interpretation is that most of Kalshi’s customers don’t know about Rule 40.11, or didn’t until Saturday, and just figured betting on death was the whole and legitimate point of this market. Everyone’s intuitions about prediction markets are still developing.

Polymarket, meanwhile, has a bifurcated structure with US and technically-non-US markets, so it has no problem listing quite explicit war contracts, and its “Khamenei out as Supreme Leader of Iran by March 31” contract traded to 99.9%. Straightforward.

Elsewhere: “Why betting on top online prediction markets is now illegal in New Zealand.” (Beause it’s gambling.)

I am an old man and I tend to think that, if you want to bet that the Nasdaq 100 stock index will go up, the thing you should do is buy the Nasdaq 100 index. I mean, there are futures contracts on the Nasdaq 100, there are exchange-traded funds that hold the stocks in the index, you can buy the stocks yourself, etc. And then if the Nasdaq 100 goes up, you will make money, and if it goes down you will lose money. Straightforward. But in 2026 the world is becoming prediction markets, which are binary and more exciting than stock markets, so you want a binary bet. You want to make 100% of your money if the Nasdaq 100 goes up 0.5%, and lose 100% of your money if it goes down 0.5%. I mean, probably not you. Not me. But apparently a lot of people want this because:

Nasdaq Inc. plans to roll out options contracts that would allow yes-or-no bets on a major stock index, the latest exchange operator to put its own spin on fast-growing prediction markets.

The company wants to list binary options on its flagship Nasdaq 100 Index and the Nasdaq 100 Micro Index, according to a proposed rule change filed with the US Securities and Exchange Commission. The Nasdaq contracts would be priced between 1 cent and $1, reflecting the market’s view of an outcome becoming true, the filing said.

It would be Nasdaq’s first foray into products that mirror prediction markets. The so-called Outcome Related Options would let traders take binary positions on whether a specified event happens. Binary options are a simplified version of an options contract in which the payout depends on the outcome of a yes-or-no proposition.

Sure. We talked last month about Cboe Global Markets’ plans to launch binary index options. Apparently the lesson everyone has learned from the rise of prediction markets is that all-or-nothing bets are more appealing than linear bets, so big traditional firms are racing to launch them.

A big story in modern markets has been the rise of zero-day options: Increasingly investors like to trade stock and index options that expire at the market’s close, because that’s more fun than buy-and-hold investing. One concern is that zero-day options “have been instrumental in driving more intraday volatility” in the stock market. The theory is:

1. Customers like to buy zero-day options.
2. They buy them from market makers, who hedge the options: If you buy a call option on the Nasdaq 100 that expires at the close today, then the market maker who sells you that option will buy some underlying Nasdaq 100 shares or futures to hedge the option. The market maker will buy some percentage — the “delta” — of the underlying stock, computed using the Black-Scholes formula. A very in-the-money option will have a delta near 100%, a very out-of-the-money option will have a delta near 0%, and an at-the-money option — an even-money bet that the market will go up or down — will have a delta near 50%.
3. At the end of the day, the delta will be either 0% (if the market closes below the option’s strike price) or 100% (if it closes above). So the delta will move pretty rapidly to either 0% or 100%, and the market maker will have to sell or buy a lot of stock to remain correctly hedged.
4. Also, if the market bounces around near the strike price, the delta will bounce around from 10% to 90% or whatever, and the market maker will constantly be selling stock and then buying it back.
5. Specifically, if the market goes down (below the strike price), the delta will go down, and the market maker will have to sell a bunch of stock, pushing the market down more. If the market goes up (above the strike price), the delta will go up, and the market maker will have to buy a bunch of stock, pushing the market up more.
6. Thus if everyone is buying zero-day options then the market will become more volatile; the market makers’ hedging trades will exacerbate market moves.

That’s zero-day options. If the next big story is the rise of binary options, then, you know, that but more so: The delta of a binary option “approaches infinity” if it is at-the-money and near expiration; if the index is right around the binary option strike price, then a market maker will have to pay out 100 cents on the dollar if the market goes up a bit and 0 cents if it goes down a bit. If everyone is buying binary options, then market markers’ hedging trades will exacerbate market moves even more.

I’m just putting this here because in a year everything will be binary options and there will be lots of stories like “Why Is The Market So Volatile? Binary Options.”

We talked a few weeks ago about an activist battle brewing at Empery Digital Inc., a digital asset treasury company (or DAT). Empery is a pot of Bitcoins with a public stock listing, which was a good idea when it launched in 2025 but a bad idea now; Empery’s stock trades at a discount to the value of the Bitcoin in the pot. Given that, an obvious move is to crack open the pot, sell the Bitcoins and return the cash to shareholders. [1]  DATs seem to resist this approach. Why? Well:

1. Maybe for good reasons: They have some clever long-term strategy, they are levered bets on crypto and need to be patient, to recover value, etc.
2. Maybe for bad reasons: Their executives get paychecks and optionality as long as the company keeps going, but if it winds up they lose that.

I wrote that “perhaps we can look forward to a wave of activism and proxy fights at discounted DATs.”

Well, on Friday, Tice Brown, the activist at Empery, filed notice that he’s launching a proxy fight to get himself on Empery’s board, where his “objective would be to work constructively—but independently—to … immediately and dramatically increases capital returned to shareholders”:

I have called for, among other things, the replacement of the Board, the resignation of the Company’s Chief Executive Officer, and a return of substantially all capital to shareholders. The Company’s continued retention of bitcoin holds no ongoing business purpose, as dozens of cheaper ways to achieve bitcoin exposure exist.

And today ATG Capital Management, Empery’s biggest shareholder, announced its own proxy fight, “nominating a slate of nine highly qualified director candidates” to Empery’s board. The board currently has eight members; assuming that Empery doesn’t settle with Brown or ATG, that means 18 nominees for eight seats. If there’s a discounted pot of Bitcoins, it turns out a lot of people will want to get control over it.

We have talked a lot about DATs around here, and I suppose the impression that I have had, and that I perhaps convey, is that they are a largely retail vehicle. The idea of “you can sell $1 worth of crypto for $2 on the stock exchange” maybe implies a certain lack of sophistication among the stock investors. And Strategy Inc., the original DAT, has a sort of meme-stock vibe. So you might expect the average DAT shareholder to be (1) a retail investor and (2) a true believer.

But I’m not sure that’s true? The way DATs got created, back in the DAT glory days of 2025, was that a handful of repeat-player hedge funds and other institutions would seed them, putting up money to build the pot of crypto, getting shares, and hoping to eventually sell — to retail? — at a big premium. That worked out sometimes, but not always. We talked in January about ReserveOne Inc., a pending DAT whose institutional investors were hoping to get their money back because the DAT trade no longer works. Some DATs have shareholder bases consisting mostly of the institutions that seeded them; Empery’s reported top shareholders (besides Brown and ATG) include Anson Funds, DRW and Highbridge. That is, it would not be at all surprising if the shareholders here (1) vote on shareholder matters and (2) are rational value maximizers, not true believers.

A running theme around here is:

1. Investors often meet privately with the executives of public companies,
2. Those executives are not supposed to give investors any material nonpublic information in those private meetings,
3. The meetings nonetheless seem to be useful to the investors, 
4. People try to reconcile Points 2 and 3 by saying that the investors get useful information from the “tone and body language” of the executives, and
5. I don’t believe them.

Like, you know, come on man. The investors are all CIA-trained interrogators? The executives are like “profits will be high this quarter” but they are sweating profusely? These people are all professionals talking about numbers in air-conditioned conference rooms; the range of tone and body language is just not that wide. I assume the obvious thing: that the investors are asking good questions of the executives, and the executives are answering them. They are giving out actual information, information that does not rise to the level of “material nonpublic information” but that is nonetheless clarifying and interesting to the investors. “Material” is, in my view, a vague and somewhat antiquated standard: In the olden days, everyone understood that telling a hedge fund “hey we’re getting acquired next week” or even “hey we’re announcing earnings that are way better than expected next week” was material nonpublic information, but modern investors are so sophisticated that helping them refine their model of margins in one business segment is enough to make them a lot of money.

Of course I might be totally wrong and tone and body language might be the main thing. I write sometimes about my favorite counterexample, which is that a big institutional investor once did dump its holdings of a public company because an analyst met with its chairman and he was “unfeasibly tanned for this time of year.” That is not financial or operational data, it is not the answer to an analyst’s question, it is the sort of thing you could learn only by actually looking at the chairman, and it has an obvious intuitive relevance that the investor correctly acted on. The company went bankrupt.

That is just one anecdote. But here is “Voice Beyond Words: Evidence That Managerial Tone Predicts Returns When Text Does Not” by David Pope of Speech Craft Analytics [2] :

This research brief examines the incremental predictive power of managerial vocal cues in earnings calls, particularly in scenarios where textual sentiment is neutral. Building on the methodology of Ewertz (2025), we analyze a large dataset of Russell 3000 earnings calls and find that paralinguistic features, such as assertiveness, arousal, and nervousness, contain significant economic information. Our findings demonstrate that even when language is devoid of strong sentiment, the acoustic properties of executive speech can predict post-earnings announcement drift, with top-quintile portfolios based on vocal features generating alpha of 40-70 basis points over 10-to 30-day holding periods. 

The acoustic stuff seems to be machine-scored, so the point here is pretty much that you can get some alpha by having your artificial intelligence model analyze the executives’ tone and body language on earnings calls. Maybe you can tell if the executives are lying, but in any case your AI can.

I wrote on Thursday about (1) my mental model that venture capitalists “have a sense of humor about fraud” in a way that other investors don’t and (2) a new paper finding that in fact “VC-backed firms are 54% more likely to face fraud charges than comparable non-VC-backed firms.” I got an email from a reader saying that he runs a VC-backed company and that, at a closing dinner for one of his fundraising rounds, one of his early investors, an experienced venture capitalist, “told me that one of my biggest weaknesses as a fundraiser was that I ‘didn’t lie enough’ in my pitch to their fund and that we should do more of it going forward.”

I suppose another model is “if you are an early investor in a company, you should want the founder to get better at lying to later investors.” At some level your success case is “later investors pay more than I do,” whatever happens with the business.

Lloyd Blankfein, the former chief executive officer of Goldman Sachs Group Inc., has a memoir out tomorrow called Streetwise. I have read it and enjoyed it a lot; Blankfein is thoughtful about his life and career and the 2008 financial crisis, and his voice and sense of humor come through. Disclosure, I worked at Goldman when Blankfein was the CEO, remember it fondly, and share some life experiences with him — we were both briefly lawyers who washed up in weird jobs selling derivatives, though he was evidently much better at it — so I am probably biased. (I am also briefly and flatteringly mentioned in the book.)

Here is a Q&A between Blankfein and Andrew Ross Sorkin about the book and Blankfein’s life. My favorite part is:

Q. When you think back on the crisis, all the vampire squid stuff and all the criticism that was heaped on the company, is there something you wish you had done differently?”

A. … If I had had perfect knowledge, I would have gone out and shorted every security instead of being flat.

What a great answer. It’s a pure hypothetical; he can’t go back and change the past. Given that, and the fact that he is being interviewed in the New York Times about “all the criticism that was heaped on the company,” the obvious move is to perform some contrition. I don’t know quite what that would sound like but, you know, “a lot of people suffered during the crisis and I wish we had done more to help and been less cocky and blah blah blah,” I don’t know. But Blankfein’s answer was “I wish we had made more money.” That’s why he was the CEO of Goldman.

Global Credit Markets Wobble as War Deepens AI-Triggered Selloff. A New Threat to Power Grids: Data Centers Unplugging at Once. Jack Dorsey’s 4,000 Job Cuts at Block Arouse Suspicions of AI-Washing. Hedge Funds, Banks Forced Into Contingency Mode in the UAE. Saudi Arabia’s Biggest Oil Refinery Halts After Drone Attack. JPMorgan Tests ‘Art of the Possible’ in Leveraged Buyouts. Toyota bows to Elliott in $38bn buyout battle. Nvidia to Invest $4 Billion in Data Center Optics Companies. Apollo Said to Net $250 Million in Paper Profits on xAI Debt. First Brands Races to Sell Units in Bankruptcy to Avert April Wind-Down. Crypto Firm Paradigm to Expand Into AI, Robotics With New Fund. The Polo-Pony Clone Wars. Did Jeffrey Epstein Want to Take Down the Pope? Meta Director of AI Safety Allows AI Agent to Accidentally Delete Her Inbox.

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!