One job of modern finance is to give you exactly the bets you want. If you want to buy Tesla Inc. stock, you can do that. If you want to make a more levered bet on Tesla stock, there are options at a wide variety of strike prices and expiration dates that are listed on exchanges; you can click a button on your brokerage app to buy a call option giving you the right to buy 100 shares of Tesla at, say, $465 per share any time between now and, say, Nov. 21.

But what if you have more arcane tastes? What if you want to buy an instrument like “you get $1,000 if, on Nov. 18, the price of Tesla stock is above $469 and the S&P 500 index is below 6,600”? You can buy Tesla calls and S&P 500 puts and get some payoff related to that bet, but not exactly that. Those two bets would pay out independently; you want, for your own reasons, an all-or-nothing bet. How can you get it? 

If you are a hedge fund, you can call up a bank, describe the bet you want, and ask the bank to write you an over-the-counter derivative giving you exactly that bet. The bank will price the bet using the market prices and volatilities of Tesla and the S&P and the correlation between them. [1]  (Intuitively, if Tesla and the S&P are very correlated, a bet like “Tesla goes up and the S&P goes down” should be cheap, because it is unlikely to happen; if they are uncorrelated or negatively correlated the bet is more valuable.) It will put those inputs into its models, get a fair value, add on a profit margin and quote you a price. 

If you are a wealthy individual investor, you might call up your broker or wealth adviser and ask her to sell you a structured note giving you the bet you want. A structured note is not quite the same thing as an over-the-counter derivative, [2]  and most banks will be willing to put more effort into structuring a $100 million hedge-fund trade than a $1 million structured-note trade, but the principles are similar. 

But to offer structured notes at scale through brokers, the banks need some automation. Big banks have technology platforms that let brokers type in the terms of a structured note and spit out a price. The math of pricing these bets is reasonably well-understood, the inputs (prices, volatilities, correlations) are reasonably observable in the market, and, you know, for a bespoke retail product, the bank can probably build in a big enough profit margin to cover any mistakes. [3]

We are not quite at the stage yet where you can go to your broker’s phone app, input exactly the bet that you want to make, get a price, push a button and execute it, but we are not that far from that stage either. Your broker’s phone app is probably sending your stock and options orders to big sophisticated electronic market makers, and those electronic market makers probably have the mathematical models and market data required to price pretty much any combination you want, and it’s just a matter of getting the right contract structure (FLEX options? exchange-traded funds? structured notes?) to execute the bet.

Do you want a bet that pays off $1,000 if Tesla is up and the S&P is down, and $0 if either of those conditions fail? I don’t know why you would. No, sorry, probably somebody does want that bet, and I kind of do know why they would. They would because they like gambling. That’s a way to get a little action. You have a somewhat out-of-consensus prediction of the future and you want to put some money on it, in part because it’s a fairly high-leverage way to win money if you’re right, and in part because it is just entertaining.

But realistically not a lot of people are doing this with Tesla because they are doing this with sports. Sports, not stocks, are the traditional locus of gambling. You have some narrative in your mind of how this weekend’s Patriots/Bills game is going to go, you convert that narrative into a series of linked propositions (“the Bills will win by at least 8, the total score will be at least 53 points and Josh Allen will throw for a touchdown”), and you make a joint bet on all of them. If they all happen as you predict, you get paid a lot of money; if any one of them fails, you lose your bet. 

This is called a “same-game parlay,” and sportsbooks are very much in the business of offering it, because (1) it is a product that gamblers want, (2) the sportsbooks have the mathematical models and market data required to price pretty much any combination you want and (3) you know, for a bespoke retail product sold to gamblers, they can build in a big enough profit margin to cover any mistakes. Like, a really really big profit margin. 

And so we have talked about the correlation desks at sportsbooks, who are in charge of pricing these bets. (There is some correlation between the Bills winning by 8 and Josh Allen throwing for a touchdown, etc., and you have to estimate that correlation to price these bets properly.) Again it is pretty automated: They make the models and pull in the market data, you choose the combination you want from a menu of possible outcomes, and their models give you a price. [4]  

But can you do it from your brokerage app? Well, why not. One thing we discuss a lot around here is the push that prediction markets, particularly Kalshi, are making into sports betting. Briefly:

1. Kalshi is a prediction market that lets you bet on various events, including particularly the results of football games.
2. Robinhood Markets Inc. lets you trade Kalshi sports contracts from inside its brokerage app, because football bets are “an emerging asset class” for financial investors and not, you know, gambling.
3. Kalshi is a commodities exchange regulated by the US Commodity Futures Commission, which gives it some huge regulatory and tax advantages over traditional sportsbooks.
4. Kalshi really wants to be big in sports gambling, because that’s a much more stable and lucrative business than predicting local elections.

But to get big in sports gambling, you need to offer same-game parlays, because (1) the profit margins are large and (2) that’s what the most devoted (addicted?) gamblers want.

There is something conceptually a bit tricky about offering parlays on prediction markets. The conceit of a “sports prediction market” is that, unlike traditional sportsbooks, it is just a neutral marketplace where investors/gamblers can trade with each other. Kalshi is not on the other side of your bets: If you bet that the Bills will win, you are betting against some other person who thinks the Patriots will win. 

But this is a hard conceit to maintain when your bet is a custom multileg parlay. If you want to bet on some precise combination of eight outcomes occurring, with a low probability of winning but a huge payout if you do win, what are the chances that some other retail gambler wants the other side of that bet? You’re doing this for fun; you want a lottery ticket. It’s no fun to sell lottery tickets, though doing it in size is lucrative. Buying lottery tickets is a popular hobby; selling lottery tickets is a profession.

This is a completely solvable problem. When you sell stock on Robinhood, it’s not like you expect another retail Robinhood customer to buy it. Most of the time, a professional electronic market maker buys your stock, and later perhaps sells it to another customer. Same with options. And more or less same with sports bets: When you bet on the Bills on Kalshi, some other retail bettor might be on the other side, but there are also professional electronic market makers who will trade against you. 

And they have correlation traders and pricing models. And so Kalshi can offer you a menu of parlay options, and you can put together the parlay that you want, and Kalshi can beam it out to market makers, and the market makers can price it and give you exactly the bet you want. And you can push a button to do it.

And that is happening. Yahoo Sports reported yesterday:

[Kalshi] rolled out a “build your own” parlay product on Monday afternoon, just ahead of the dual Monday Night Football matchups between the Dolphins and Jets and the Bengals and Broncos. The product launch comes four weeks after Kalshi first self-certified to offer multi-leg bets. The prediction market launched some pre-built parlays before the NFL season opener, but rollout of those bets had been limited since then.

A Kalshi user can select different bets — such as a game winner, a player to score a touchdown, or total points — to make up legs of the parlay. It is then sent out to certain other users on the Kalshi exchange as a “request for quotations,” asking them to give a price to the opposite side of the parlay. In practice, the deals are priced by institutional market makers — large-scale traders who typically receive rebates or other incentives for offering odds on certain bets.

Sportico adds:

Available legs to choose from included points spread, moneyline, total points scored and anytime touchdown scorer. Once the legs were selected, parlays were filled almost immediately. ...

Unlike other trading markets on its platform, during the time the customizable parlays were available, Kalshi’s public app had no apparent feature enabling “limit orders” in which retail customers could propose a betting line for others to take. Instead, odds were presented by Kalshi as take it or leave it, and it is unclear who set the prices and whether the process conflicted with its peer-to-peer messaging.

Here is Kalshi’s self-certification of the parlay contracts, the form that it files with the CFTC to allow it to list those contracts (and to preempt state gaming regulation). Sadly the appendix with “an analysis of its risk mitigation and price basing utility” — the justification for treating these contracts as commodities futures rather than gambling — was filed confidentially. But I am sure that somebody somewhere is hedging a real economic risk and contributing to price discovery by, you know, betting on both the Bills and the over.

The Wall Street Journal notes that “Shares of DraftKings and Fanduel parent Flutter Entertainment sank around 11%” on the news, as Kalshi gets one step closer to becoming a full-featured federally regulated sportsbook, though “Kalshi’s parlay product is still bare-bones compared to what traditional online sportsbooks offer” and “prediction markets are having a limited impact on DraftKings’ and FanDuel’s handle and engagement so far this football season.” Still. We are one step closer to democratizing finance with the emerging asset class of same-game football parlays.

The thing investors want is uncorrelated returns. The stock market mostly goes up, but sometimes it goes down, and if you are a big institutional investor who owns a lot of stocks, you will find this annoying. You need money every year, and you do not want to report losses to your clients. You own stocks because they mostly go up, but you worry about the times when they go down.

And so if you can find another thing that (1) mostly goes up, (2) sometimes goes down but (3) goes down at different times from the stock market, that is good and you should pay up for it. A classic modern case is catastrophe bonds, where you invest money, get paid a high rate of interest, but sometimes lose a lot of money if there is a big hurricane. This is, for many institutional investors, a perfect risk profile: You understand that getting paid a high return requires taking risk, but what you want is to take uncorrelated risk. There’s no reason to think that hurricanes are correlated with stock market crashes. If there’s a big hurricane and stocks are up, or if stocks are down but there are no hurricanes, the catastrophe bonds have improved your risk-adjusted returns.

This works even for individual stocks. When the stock market goes down, some stocks go up, or at least don’t go down as much. Consumer staples and utilities are traditional examples of defensive stocks: Generally the stock market goes down when the economy slows, but when the economy slows people still need food. In financial theory, investors should be willing to pay more for these sorts of stocks; that is, those companies should have a lower cost of capital (a lower expected return) than other, regular stocks that move with the stock market. [5]

There are, however, measurement problems. What you want is an investment that will hold up well in the next market crash. What you can get is some historical measure of realized correlation, most classically beta (a measure of the correlation between a stock’s return and the market return). You can say “the daily correlation between this investment and the stock market is low, so it probably will hold up well in the next crash,” which is reasonable but not certain. Perhaps the price of the investment moves independently from the stock market on most normal days, but when there’s a crash it also crashes. [6]  

A classic case is private equity. Private equity is a business of owning the equity of companies, generally with a lot of debt. You might naively think that owning companies would be highly correlated to the stock market (which also involves owning companies): When the economy slows, that is bad for corporate profits, which should be bad for the value of public and private companies alike. (It should be worse for private-equity companies because they have a lot of debt.) But one cynical but important model of private equity is that it is the stock market, but on a delay. [7]  When the economy starts to slow, the stock market drops, because investors anticipate lower cash flows in the future. But private equity companies don’t trade, so their prices don’t go down immediately. The private equity owners, who mark their own investments, say “ehh the cash flows have held up fine, no change in value.” They report good performance to their institutional investors, who are thrilled to get good performance in a rough year for the stock market. Much later, the cash flows go down, the companies are marked down, and the private equity funds report bad performance to their institutional investors, who are still thrilled, because now the stock market is up. The private equity fund has given the investors what they wanted, uncorrelated returns. But this is not because of some intrinsic economic difference between the companies in the private equity fund and the companies in the stock market; it is because the stock market reacts faster to economic news than the private equity valuations do. [8]

This could work for individual stocks, too, a little bit. For small, illiquid stocks, anyway. The intuition is: You have a small company that makes widgets. Widgets are cyclical, and when the economy slows you will sell fewer widgets and make less money. But it’s a small company; it’s not in the big stock-market indexes, not covered extensively by sell-side analysts, not owned by a lot of big sophisticated institutions. When the economy slows, nobody thinks to dump your stock. You just bop along selling widgets. Eventually you sell fewer widgets, and you announce bad earnings, and your stock drops. But your stock drops later than the rest of the market, because no one was paying attention. 

This makes your company look good, by the important standard of uncorrelated return. Your stock mostly goes up and sometimes goes down, but it goes down at different times from the rest of the market. This makes you an attractive investment, so investors are willing to pay up for your stock, and your cost of capital is lower than it would otherwise be.

But this is an unstable equilibrium. If more people pay attention to your stock, then they will do things like sell it when the rest of the market goes down. Your low correlation to the stock market is not an intrinsic fact about your business; it is a contingent fact about how people trade your stock. That is, mostly, they don’t. If they traded your stock more, it would move more in line with the stock market, and it would be less attractive.

Here (via Byrne Hobart) is a fun Bank for International Settlements working paper on “The speed premium: high-frequency trading and the cost of capital,” by Matteo Aquilina, Gbenga Ibikunle, Khaladdin Rzayev and Xuesi Wang. Intuitively:

1. High-speed algorithmic traders can pay attention to all the stocks, so they trade all the stocks. This makes smaller stocks trade more often.
2. They can only pay so much attention to all the stocks. Their trading of smaller stocks will be based largely on heuristics like “if the economy is slowing then widget sales will slow,” or more plausibly “if all the other stocks went down then this stock should go down.” 
3. Therefore, high-frequency trading makes many smaller stocks more correlated to the market; they trade more on systemic news and less on idiosyncratic news or simple inattention.
4. This makes those stocks less attractive and raises their cost of capital.

From the abstract:

Using co-location and latency improvement upgrades at NASDAQ as natural experiments, we find that, on average, high frequency trading (HFT) leads to higher cost of capital. However, the impact is not uniform. HFT raises the cost of capital for low-beta stocks by amplifying their systematic risk, as HFT’s correlated trading strategies make these stocks more responsive to market-wide information. For the most liquid stocks, HFT reduces the cost of capital by lowering the liquidity premium required by investors.

Investors want stocks that are liquid, but they really want stocks that do not move with the market. In a world of slow idiosyncratic trading, there are a lot of stocks like that. In a world of fast systematic trading, there aren’t.

The basic theory is that if you own a professional sports team you get both a financial return and consumption benefits. You get the cash flows from the team (television rights, etc.), but also you get to, like, hang out with the team and feel important and cool. After your team wins the big game, you get to go to the locker room and high-five everyone. You can call up the quarterback and ask him to play catch with your kid and he probably will. 

This implies that the price of sports teams should be high, and their expected (financial) return should be low: People will pay a premium to high-five the team, so they will bid up the price of teams well beyond what is justified by their cash flows. (This is such a problem that sports leagues have rules essentially requiring owners to try to be profitable, and we talk sometimes about owners cheating to get around these rules to be less profitable but have more fun.) Eventually the price of teams will get so high that they will all need to be owned by diversified institutional investment pools rather than rich individuals.

Which … defeats the purpose? Like if you are part of a diversified institutional investment pool that owns a sports team, do you get to high-five the team? Maybe? Here’s this:

Sports and media business veteran Matt Rizzetta is opening private-equity firm Underdog Global Partners to focus on investing in sports, real estate and media intellectual property.

Part of his pitch to investors includes “priceless experiences and access” to what the firm describes as “a global ecosystem of relationships, networks and influencers.”

Among Underdog’s assets at inception are global sports and media properties as well as associated real estate such as stadiums and arenas. Early backers of the nascent firm include high-net-worth individuals and family offices.

“Priceless experience and access” but there does seem to be a market for it.

Yesterday we talked about Charlie Javice’s prison sentence for defrauding JPMorgan Chase & Co. when she sold it her startup Frank. I was skeptical of her lawyer’s argument that, because JPMorgan looked pretty dumb in this deal, she should get a lighter sentence. I wrote: “I feel like normally it’s the opposite? ‘My victims trusted me so completely that they didn’t ask any questions’ does not normally make your fraud look better.” And: “I would have thought that defrauding dumb victims would go worse for you.” That was just speculation on my part, and certainly not legal advice, but there is a literature. Here is “Beyond Misconduct: What Explains DOJ Involvement in SEC Enforcement,” by Nathan Herrmann, Sara Toynbee and Matthew Kubic of the University of Texas:

While the Securities and Exchange Commission (SEC) is the primary regulator of the U.S. securities markets, it must rely on criminal authorities such as the Department of Justice (DOJ) to threaten criminal sanctions. If criminal sanctions deter misconduct, then cooperation between the SEC and DOJ is crucial for effective enforcement of securities laws. We show that factors other than the type and severity of misconduct explain DOJ involvement in SEC enforcement. Cases are more likely to involve the DOJ when they have more prosecutorial appeal (e.g., vulnerable victims and dislikable defendants) and the associated DOJ office has more white-collar crime experience. DOJ workload, but not SEC workload, also reduces the likelihood of SEC-DOJ cooperation. These results suggest that the ability to threaten the most severe penalties in SEC enforcement depends on factors beyond the SEC’s control.

That is: Some sorts of securities fraud are only of interest to the SEC (which cannot put you in prison), while others are also of interest to the DOJ (which very much can), and one thing that makes prison more likely is “vulnerable victims.” Again I do not normally think of JPMorgan as a particularly vulnerable victim, but that was the claim here, and Javice got seven years in prison.

Meta to Start Using Chatbot Conversations to Target Advertising. Jain Global Buys Firm to Fast-Track Physical Natural Gas Trading Push. 
Currency Arbitrage Costs Milei $2 Billion a Month in Argentina. UBS Says Bank’s Business Model, Swiss Future at Stake in Regulatory Overhaul. Japan’s Firebrand Investor Is Back, Along With the Country’s Stock Market. Berkshire Hathaway Near $10 Billion Deal for Occidental’s Petrochemical Unit. Texas Stock Exchange Could Start Listing Shares Next Year. Education Tech Provider Anthology to Hand Over Keys to Oaktree, Nexus in Bankruptcy. Student-Loan Debt Is Strangling Gen X. Taiwan Rejects US Demand for Half of Chips to Be Made in America. Supreme Court Refuses to Let Trump Immediately Oust Fed’s Cook. US vows to maintain tariffs regardless of Supreme Court ruling. 

If you'd like to get Money Stuff in handy email form, right in your inbox, please