Money Stuff: Can the Lenders Stick Together?

Here is my schematic description of modern “liability management exercises”:

1. A company borrows $1,000 in good times from a bunch of lenders.
2. It runs into bad times, can’t pay back the $1,000, and needs more money to stay afloat.
3. It goes to some of its lenders — ones who hold 51% of its debt — and says “Look, we don’t have $1,000. But here’s a deal. If you lend us $100 today, on top of the $510 you already gave us, we will pay you back $800 in three years, a nice $190 bonus. But in exchange, you have to vote to amend the credit agreement so that we can pay the other guys back $0.”
4. The 51% lenders are like “sure, fine, that’s a good deal for us, too bad for those other guys.” And they vote to amend the credit agreement to zero the other guys.
5. Net, the company has borrowed $1,100 and only has to pay back $800 (a $300 windfall). The 51% favored lenders have loaned $610 and get back $800 (a $190 windfall). The 49% disfavored lenders have loaned $490 and get back zero, a $490 loss that pays for everyone else’s gain.
6. The 49% disfavored lenders complain a lot and sue, with mixed success.

This is exaggerated and schematic; in particular, you cannot actually amend the credit agreement to say that some lenders get back zero. (What you do is amend the credit agreement to strip collateral and other protections from the disfavored lenders, or restructure the company to move assets out of the collateral package, or give the favored lenders new super-priority liens, etc.) But it captures the rough intuition of a lot of recent trades.

Given this dynamic, it is very important for lenders to be in the 51% (who get paid with a bonus) rather than in the 49% (who get hosed). There are various ways to do this. One way is to be a big important scary lender, so companies (and their private equity sponsors) won’t want to hose you. Another way is to come to the company first: If, at the first sign of trouble, you go to the company and say “heeeyyy what if you paid us a little bonus in exchange for some new money and voting to zero the other guys,” you’re more likely to end up in the group of favored majority lenders. (Occasionally two groups try this and get into a bidding war to be the favored group.) Thus liability management exercises are sometimes described as “creditor-on-creditor violence,” suggesting that the lenders are jockeying to help themselves at the expense of other lenders.

A third approach is for all the lenders to say “no, this is terrible, the whole point of being credit investors is that we all get paid back, we are not going to stand for these liability management exercises anymore.” One way to implement that approach is in credit agreements (and bond indentures): Lenders can refuse to make loans whose documents have loopholes that allow for these sorts of shenanigans. This is a bit hard, though, because (1) the credit agreements are long and complicated and creative lawyers keep finding new loopholes and (2) generically, the credit agreements were written in good times, when lenders feared missing out on deals and didn’t worry too much about distress, and are exploited in bad times, when it’s too late to change them.

And so another way to implement the approach is for the lenders to say “no”: If a company comes to them and says “hey we’ll give 51% of our lenders a nice bonus for hosing the other 49%, do you want to be in the 51%,” each lender can say “no thank you.” Even if it is individually advantageous, in the moment, to be in the 51%, in the long run it is bad for lenders to have to deal with this stuff, and they’d prefer to cut it out. If everyone says no, then these deals won’t happen, and lenders generally will be better off. [1]

Obviously there is a prisoners’ dilemma here: If you say “no thank you,” and everyone else says “sure sign me up,” then you will be in the 49%, which is very bad for you. Just individually refusing to participate doesn’t work. You need everyone — or at least most lenders — to refuse to participate. 

And so in some deals lenders sign up to “cooperation agreements,” where the lenders (often majority holders) agree among themselves not to take a separate deal with the company. Sometimes the cooperation agreement is a way for some set of lenders to get to 51% and negotiate a collective deal with the company, but other times the cooperation agreement is a way for the lenders to prevent any sort of asymmetric deal and force the company to treat all of the lenders equally.

At some intuitive gut level, that sounds nice to me. There is something off-putting about this whole process, a broad norm of “distressed companies should have to treat all of their creditors the same” [2]  seems fine, and if all the creditors want to get together and agree on that then who am I to complain.

On the other hand, if a bunch of competing firms get together and agree not to cut separate deals with the company, is that … an antitrust problem? Is that a conspiracy in restraint of trade? I dunno, maybe. Bloomberg’s Reshmi Basu and Chris Dolmetsch reported last week:

Altice USA Inc. filed an antitrust lawsuit against some of its largest creditors Tuesday, an opening salvo ahead of potential discussions to restructure its $26 billion debt pile.

Optimum Communications Inc. — Altice’s name following a rebranding earlier this month — brought the case in federal court in New York against lenders including Apollo Capital Management LP, Ares Management LLC and BlackRock Financial Management Inc., alleging that they worked together to freeze the company out of the US credit market.

The suit argues that a so-called cooperation agreement between the defendants binds “nearly every creditor holding Optimum’s debt” and bars them from dealing with the company unless two-thirds of the firms approve. …

“The cooperative is a classic illegal cartel,” the company said in the complaint. “Competing debt investors have agreed to lock Optimum out of the credit market unless Optimum offers terms the entire cooperative deems acceptable.

“It is also classic price fixing: the cooperative is collectively dictating terms to Optimum by forcing it to transact only at loan and bond prices the whole group will accept,” the complaint reads. “These restraints decimate competition and obstruct the capital markets from working as intended.”

Here is the complaint, which gives a description of liability-management exercises that is maybe a touch more sympathetic than mine, but only a touch: 

Defendants and their advisors have sought to justify their conspiracy by claiming they need it to curtail “creditor-on-creditor violence.” In this telling, leveraged debtors like Optimum often play creditors off each other by enticing some creditors to strike unilateral deals that help themselves while potentially undermining competitors. Because this process induces rival creditors to act against each other’s interests, Defendants like to call it “violence.” But that is just a pejorative term for competition. The antitrust laws’ founding principle – that competitors should compete – does not evaporate just because the competitors here collectively dislike it. In a free market, competing creditors should work against each other to unilaterally help their own position. That process, of course, may ultimately help some creditors at the expense of others. But that is hardly a reason to allow this conspiracy. When these creditors decry “creditor-on-creditor violence,” they are really just articulating the traditional logic of a cartel: the belief that collusion between rivals yields better outcomes for those in the cartel.

Right, I mean, from a distressed borrower’s perspective, it’s good for the lenders to compete to offer the company a good deal to avoid being left out entirely. The company has a lot more leverage, if each lender is afraid that all the other lenders will betray it. Is it required by antitrust law? Maybe?

I should add that not all cooperation agreements work like this. Here, allegedly, “the Cooperative has captured a staggering 99% of [Optimum’s] outstanding loans and bonds,” making it impossible for Optimum to cut a deal with some favored creditors. In other cases, though, a cooperative will control, you know, 51% of the debt, and will then cut its own deal with the company at the expense of the other 49%. Just a standard liability management exercise, but with the 51% holders coordinating beforehand. Is that an antitrust problem? In October, Bloomberg’s Giulia Morpurgo reported that “a group of lenders to Selecta is suing the vending machine company and some of its bondholders — Strategic Value Partners, Invesco, Man Group and Diameter — in the first legal challenge to the kind of creditor pacts gaining traction in distressed debt negotiations,” also alleging antitrust violations. “By using the Cooperation Agreement to enhance their own market position at the expense of other market participants, the Favored Holders violated Section 1 of the Sherman Act,” says the complaint in that case. If lenders get together to stop creditor-on-creditor violence, or to commit it, either way there is some antitrust risk.

Here are some things about the stock market that are approximately true:

• It’s hard to know which stocks will go up and by how much. People who can reliably beat the market — who can reliably buy stocks that will go up a lot and avoid stocks that will go down — are rare and in high demand, most people should own index funds, etc.
• There are, however, some stocks that are obvious garbage, and it’s easy to tell that they will go down. If you spend five minutes looking at one of them, you will think “this stock is garbage, I should sell it short to profit when it goes down.”
• But if you try to do that, you will run into a problem: To sell a stock short, you need to borrow it from someone who owns it, and you have to pay the owner a fee to borrow the stock. And this stock — the one that you immediately knew was garbage — has a very high borrow fee. Because, it turns out, everyone who looks at this stock immediately knew that it’s garbage and decided to short it, and all the demand for short selling bid up the borrow cost until it was very high — 10% or 20% or 100% or 1,000% of the price of the stock per year. [3]

How big is this problem? You could imagine three possible answers:

1. It’s expensive to short the stock, but it is such garbage that it is nonetheless profitable: Even paying 100% per year to short the stock, you’ll still make a profit, because the stock will go to zero in like a month. The borrow cost is lower than the expected return of shorting. If this is true, you should short the stock.
2. It’s much too expensive to short the stock: The borrow cost is so high that, even if the stock goes to zero in the time you expect, you will lose money. The borrow cost is higher than the expected return of shorting. If this is true, you should buy the stock, lend it out in the stock borrow market, get paid the stock borrow fee, and make a profit even as the stock goes to zero.
3. The borrow cost and expected return of shorting are in precise equilibrium: In expectation, you will make about as much money on your short as you will spend on borrowing the stock. The combination of (1) the stock market and (2) the stock borrow market is efficient: The stock is knowably garbage, but once you factor in the borrow cost you should not expect to beat the market either by shorting it or by going long.

Obviously the pleasing answer would be No. 3: The stock market is generally efficient, and in the pockets where it is not efficient — where it is easy to identify a stock that is overvalued — it is nonetheless meta-efficient, so you can’t profitably trade on that insight.

And that’s approximately right. Here’s a fun paper on “Inefficiencies in the Securities Lending Market” by Kent Daniel, Alexander Klos and Simon Rottke. (Here is a blog post from John Cochrane about it.) On the one hand, the paper is a story of inefficiency: Stock borrow costs have gone up, leading to bigger stock mispricings. From the abstract:

In contrast with a general trend of declining trading frictions, over the last several decades the cost of borrowing securities for short-selling has increased dramatically. Using a portfolio approach, we show that as the borrow costs have increased so has the mispricing associated with portfolios of high-borrow-cost names. This decline in market efficiency has resulted from a lack of competition in the intermediation chain that links share lenders with borrowers, and a growing and rational unwillingness among institutional investors to hold and lend high-borrow-cost names. Since 2020, we estimate that the inefficiencies associated with these frictions have exceeded $300 Million/day.

On the other hand, the stock-plus-borrow pricing seems eerily accurate, in some sense. They construct a portfolio of stocks with the highest borrow costs and find that those stocks are on average, garbage, and that their borrow costs almost perfectly reflect what garbage they are:

Our high-borrow-cost portfolio ... earns an annualized CAPM alpha of -81.4% (t = -5.87). Thus, an investor in this portfolio who did not lend out the securities would have earned a risk-adjusted return of -81.4%.

Given the large negative alpha earned by this portfolio, one might think a short position in this portfolio would have earned a high positive alpha of 81.4%. However, the ex-post cost of borrowing the shares over this period was approximately 84.8%, so net of the borrow cost the return to shorting this portfolio was an annualized -3.4% (t-statistic: -0.25), based on the daily indicative fees reported by Markit.

A net-of-borrow-cost alpha of zero is consistent with a market in which short-sellers are rational, informed, and competitive. The net return for our high-borrow-cost portfolio is statistically indistinguishable from zero. 

But intermediation in the stock borrow market is expensive enough that, while the short sellers are making approximately $0, the long side of this trade — the investors who own the stocks and lend them out — is losing money:

By comparing this borrow cost with the revenue received by the fund, we estimate that the intermediation costs are about 40% of the fees. Thus, the lender receives only 60% of every dollar that a short-seller pays to borrow its shares. The intermediaries’ share does shrink with the borrow cost but is still roughly 67% for even higher borrow cost names. Thus, once one accounts for the spread between the cost paid by the share borrowers (as reported by Markit), and the fee received by the lenders, the alpha from a strategy of holding and lending out a portfolio of high-borrow-cost names are strongly negative and highly statistically significant.

Anyway the paper also has a sample list (Table 1, page 2) of 30 stocks with the highest borrow costs in June 2025 (more than 400% annualized), and, uh, let me put it this way, several of them will be familiar names to Money Stuff readers. Sometimes it’s not that hard to tell.

Some people like solving puzzles and are good at it, and the world is their oyster. Many of them would like to pursue a career that (1) consists mainly of solving fun puzzles and (2) pays them lots of money. In 2025, the field where those career opportunities are densest and most lucrative is arguably finance, particularly the quantitative/prop-trading/market-structure bits of finance. It is not entirely obvious why that would be the case, and you could imagine a world where the puzzle-solvers flocked to, you know, cancer research or sending rockets to Mars or historical linguistics or theology. Also finance has always had a lot of competition from the tech industry, and in particular these days from generative artificial intelligence. 

But there is something about finance as a meta-layer on economic activity that makes it a good general source of puzzles: If you are good at building rockets, you will build good rockets, but if you are good at capital allocation you can find the most lucrative possible puzzles, whether they are in rocket-building or biotechnology or internet advertising, and solve those. AI has a similar meta attraction: If you are good at building smart computers, you can find the most lucrative possible puzzles and have your computer solve those. We talked yesterday about having your computer solve puzzles to steal crypto, a particularly pleasing application.

Anyway the point is that now if a young person in the US is good at solving puzzles, someone will eventually come up to her at a chess tournament or a Math Olympiad or an MIT Puzzle Hunt or a historical linguistics convention or even, like, a SpaceX internship, and will quietly take her aside and hand her a bag of money and a Rubik’s dodecahedron and whisper “you come with us now.” And off she will go to join her fellows at quantitative hedge funds and proprietary trading firms, where she will spend her days solving market-structure puzzles and her evenings playing poker and the occasional weekend doing this:

Working together, they figured out how to punch a dummy at the right speed to play a series of TV news clips that correlated with the dates of famous boxing matches. Nine boxes held objects related to specific boxers along with a scale. Weighing the correct pairs of boxes provided the coordinates of words in the newspaper they’d been given at the beginning of the scavenger hunt, directing them to the final spot on their quest.

That is from Rainier Harris’s story in Bloomberg Markets Magazine about Midnight Madness, the big financial-industry puzzle hunt in Manhattan. We have talked about Midnight Madness (and its predecessor, Compass) around here before; I have played a few times (not this year, sadly), and this year’s puzzle creators were on the Money Stuff podcast a few months ago. The anecdote I quoted is about a team sponsored by Raymond Iwanowski of Secor Asset Management LP and including Daniel Aisen of Proof Trading Inc.; they came in third. Is a puzzle hunt like finance? Sure why not:

 “Quantitative people like to solve puzzles,” Iwanowski says.

It also turns out that the scavenger hunt, with all its intricacies, has a lot in common with investing. Just like the Ouija board puzzle that ensnared the Midnight Marauders, being successful in finance often comes down to figuring out which problems are worth solving. “Really good investors are really good at separating the signal and the noise,” Iwanowski says.

Presumably really good _____s are really good at separating the signal and the noise, for lots of professions that could go in that blank, but investing is a particularly lucrative place to apply that skill. 

We talked the other day about URL guessing, a simple way to occasionally get material nonpublic information: A company or government agency regularly publishes some data on a web page, the URL of the web page has the same format each time, last quarter’s URL was “company.com/data_q3.html” or whatever, so you go ahead and guess that next quarter will be “company.com/data_q4.html.” And then you type that into your browser an hour before the data is supposed to be announced, and occasionally you get lucky, the data has been uploaded before the official announcement, and you get to see it first.

A slight variation on that is: Some company or agency has a website, it uploads media assets (pictures etc.) to that website to get ready to put out some press release, and those assets are publicly findable even before the release goes out. You set up a web scraper, it scrapes the site, it finds the pictures before they are announced, and those pictures give you early access to material nonpublic information.

We talked about this approach in October, because someone reportedly used it to (1) predict the winner of the Nobel Peace Prize hours before the announcement and (2) make tens of thousands of dollars trading on Polymarket. Apparently the same thing happened with the year-end Spotify rankings. Prediction market trader Fhantom tweeted:

Just like Nobel site, The Spotify Newsroom blog is run on WordPress. If you know where to look, you can query when they upload new images to the media library. Anyone who studied previous years would know that Spotify uploads the Wrapped graphics there before the blog post going live.

So if you’re watching that feed, you can see tomorrow’s “big reveal” today, and trade on it before any public announcement.

All year, Spotify’s streaming numbers are public, so most of the prediction markets on the event were already trading above 90%, but one ranking was a big surprise.

Drake was on track to finish as the #3 artist globally, ahead of the Weeknd by over 2 billion streams. On the leaked graphic, he suddenly shows up at #4. This is likely tied to Spotify’s recent lawsuit over botted streams. It looks like they filtered out a huge chunk of Drake’s streams.

The market on Drake was trading >95% yesterday morning and then began to dump and this image hit the site and traders discovered it.

It is in its way a tiny triumph of market efficiency, a micro-scale case of the power of financial markets to attract puzzle-solvers and incorporate information into prices. Spotify’s public data suggested that Drake was the #3 artist this year, but you could make, you know, tens of thousands of dollars by finding out that that data was wrong, and someone did. The world — I mean, the people watching Polymarket Spotify markets anyway — knew that the Weeknd beat out Drake hours before it was officially announced. Is this good? Is it socially useful? Man, I don’t know, don’t ask me, but I guess the market signals tell you that it is.

Anthropic taps IPO lawyers as it races OpenAI to go public. Prediction Market Kalshi Hits $11 Billion Valuation in New Funding Round. Native American Tribes Fight Sports Betting Rivals. Brussels floats ‘emergency’ powers to raise €210bn from Russian assets. Serial-Defaulter Argentina Preps Return to Foreign Bond Market. Harvard’s Big Wager on Bitcoin Came Right Before the Bust. The 26-Minute, 51% Wipeout That Deepened Trumps’ Crypto Woes. Dell’s $6.25 Billion Gift Marks New Path for Billionaire Charity. The New, In-Demand Job Skill: Being a TikTok Influencer for Your Company. Baby slop doo doo di doo di doo. Raccoon goes on drunken rampage in Virginia liquor store and passes out on bathroom floor.

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!