The classic description of a successful investment banking analyst is “detail-oriented.” When you arrive at an investment bank fresh out of college, you will be asked to prepare materials for client meetings and to build financial models in Excel. How can you demonstrate that you are good at the job? You just got there; you are unlikely to have any brilliant insights into the client’s needs. Pretty much you’re going to put together the materials your bosses tell you to, using pages from previous client pitches. You can do this sloppily in a way that embarrasses your bosses: Every investment banker has been in a client meeting where the pitchbook has the wrong logos because the analyst copied from a precedent without updating it. No one likes that! Or you can do it sloppily in a way that doesn’t embarrass your bosses, but that causes them to seethe with secret rage: Every investment banker has also seen a vice president get furious about a table with dollar signs aligned in two different ways. This rage is not substantively justified, but the VP is worried that if the dollar signs are misaligned then there might be a worse error somewhere in the book. Or you can just do it perfectly, and then the VP will like you.

Meanwhile, if you are building big financial models with a lot of moving pieces, they will take you hours to build, but they will also take a long time for your bosses to check. If you demonstrate that you don’t need checking — if you build your models right the first time, if you follow best practices in modeling and formatting, if everything looks clean, and also if your pitchbooks are error-free — then your bosses will trust you and appreciate your work. If you make them nervous — if your models have obvious mistakes, or if your pitchbook formatting is bad — then they will have to spend a lot of time checking your work and will not appreciate that.

Occasionally in investment banking you will run into analysts who are smart and creative and personable and financially savvy but not detail-oriented. They have good ideas and are good with clients, but their pitchbooks are full of dumb mistakes. Most of these people do not last long in investment banking: Their bosses dislike them, and they dislike being yelled at to reformat pitchbooks. They do their two years, or sometimes less, and then leave to become startup founders or food-truck operators or bloggers. A few of them survive, though: They start as incompetent junior bankers, and then they get yelled at a lot and feel bad and become semi-competent junior bankers, and if they last long enough they blossom into competent senior bankers, where the skill set is different and they don’t have to format pitchbooks.

Modern generative artificial intelligence models are, stereotypically, like that. They know a lot. They can do lots of things and calculate quickly and come up with ideas and put those ideas into competent prose; they can convey the impression that they know what they are doing. If you ask them to build a financial model to price and structure a leveraged buyout of a company, they’ll probably produce something that looks plausible and that is a good starting point. But they make mistakes, they “hallucinate,” and you can’t entirely trust them. You have to check the whole model yourself.

What you want, if you are an investment bank looking to deploy AI models, is to take an AI chatbot and have an investment banking vice president yell at it for a few years until it gets the formatting right. “Reinforcement learning,” I believe this is called. OpenAI is doing it:

OpenAI has more than 100 ex-investment bankers helping train its artificial intelligence on how to build financial models as it looks to replace the hours of grunt work performed by junior bankers across the industry.

The group, which includes former employees of JPMorgan Chase & Co., Morgan Stanley, and Goldman Sachs Group Inc., is part of a secretive project inside the startup that’s code named Mercury, according to documents seen by Bloomberg.

Participants are paid $150 per hour to write prompts and build financial models for a range of transaction types, including restructurings and initial public offerings, according to a person familiar with the effort. The company has also granted the contractors early access to the AI it’s creating that aims to replace entry-level tasks at investment banks. …

Participants are asked to create their models in Excel and they’re also expected to follow industry norms for formatting the models, including for areas like margin sizes and italicizing percentages.

Right, see, if you are getting paid $10 million a year to work on foundational issues in cutting-edge artificial intelligence research, and a customer comes to you and says “I’m impressed that your AI can build me a full LBO model in 10 seconds but the percentages are not italicized,” you will roll your eyes and not make it a priority to fix that. But if you hire a lot of ex-bankers to teach the AI to build LBO models, they will automatically get the formatting right. They have experienced a lot of reinforcement learning of their own.

There is something culturally very pleasing about this? For one thing, an essential part of investment banking analyst culture is complaining about the hourly rate. “Sure,” a 22-year-old analyst will say, “I get paid $200,000 a year right out of college, but when you consider that I work 110 hours a week I am practically getting minimum wage.” I mean, no, you’re not, but $200,000 really is $35 an hour if you work 110 hours a week 52 weeks a year. OpenAI will pay you $150 an hour for piecework! That’s like $300,000 a year for a 40-hour week, though I get the impression that this gig might not last a year.

Also, though: In some obvious sense the point of Project Mercury is to put investment banking analysts out of work. Once these ex-bankers train the OpenAI models to be really perfect at italicizing percentages, senior bankers will no longer need analysts to prepare pitchbooks or build models, and then … something. It is hard to know exactly what will happen to an industry build on an apprenticeship model when it no longer needs apprentices: Banks will still need senior bankers to show up and shake hands with the clients, and where will those senior bankers come from if not from the ranks of junior bankers? 

But of course most junior investment bankers don’t become senior bankers. Another essential part of investment banking analyst culture is leaving after two years. Once you’re out anyway, you might be perfectly happy to train a robot to replace junior bankers. 

I write a lot about what I call “abstract commodity space”:

Sometimes you want to buy nickel or aluminum or coffee or cocoa to make batteries or beer cans or cappuccino or chocolate bars, so you go to some supplier and negotiate a contract for the delivery of a useful amount of a particular grade of the commodity to your factory. Sometimes, though, you want to bet on the price of nickel or aluminum or coffee or cocoa, to hedge some risk to your business or just as a speculative bet. So you buy commodity futures, financial assets that reflect the price of a commodity but don’t require you to store it or worry about it spoiling.

The way these futures often work is that there are big warehouses full of the commodity, and people write futures contracts that essentially transfer the entitlements to the commodities in the warehouse, without ever having to take them out. Your futures represent a claim on some nickel or coffee in a warehouse in abstract commodity space, and you don’t have to think much about the physical properties of the actual thing. The warehouse system has put a layer of abstraction on the messy commodity business, and you can treat the commodity as just a number on your computer screen.

It turns out to be quite useful, in a lot of contexts, to be able to answer a question like “what is the global price of copper” with a single number. “The price of copper is $10,700 per ton,” you can say, which is useful for things like hedging your copper mine’s production or betting on macroeconomic trends or describing the long-run effect of Chinese industrialization or whatever.

This number is almost never right in any practical sense: If you need a ton of copper to make stuff at your factory, you will pay more, or occasionally less, than $10,700 to obtain it. The $10,700 number is the price of abstract copper, the price that you will pay to obtain a receipt entitling you to one ton of copper somewhere in the global network of abstract copper warehouses. Most buyers of abstract copper will have no occasion to convert it into useful copper. Some will, though, and if they do, there will be some costs of conversion. A few years ago there was a mini scandal involving long delays in converting abstract aluminum into useful aluminum: There was just a long waiting list to drive the aluminum out of the abstract warehouses, making it weirdly expensive to do the conversion.

One obvious cost is tariffs. In 2025, if you move anything into the US from pretty much anywhere else in the world, you will probably have to pay a tariff on it, which will increase its price. This includes copper. Abstract copper, though, arguably isn’t anywhere in the world: It is in abstract commodity space, which is not in any particular country. This is actually more or less true, as a legal matter. The Financial Times notes:

The London Metal Exchange ..., the world’s biggest base metals exchange, deals in copper stored worldwide in bonded warehouses before taxes become due, to deliver what chief executive Matthew Chamberlain described as a “clean, global price” exempt from tariff volatility.

The “clean, global price” is the abstract global price of copper. If you just want to buy abstract copper, the LME will sell it to you. If you then want to convert that into usable copper, you will have to take it out of the bonded warehouses, and then all bets are off: Depending on where it is and where you are, you might have to pay all sorts of tariffs. But if you are a financial user of abstract copper — if your goal is to hedge (clean) copper prices rather than to get copper into your factory — you just won’t do that and there’s no problem.

That FT article is about how the LME is gaining market share over Comex, a rival commodities exchange, because Comex is (1) less abstract and (2) behind the US tariff wall:

The US president’s on-again, off-again tariffs on imports of copper have triggered wild price swings on New York’s Comex exchange and continue to skew prices for future deliveries, analysts say, because contracts there include any duties or other taxes payable in the US. …

The result has been a big shift in business to London. Trading volumes this year at the LME — which has spent the past three years trying to recover from its calamitous nickel crisis of 2022 — have risen for its key base metals contracts. ...

At the same time, copper trading volumes at the smaller Comex, part of the CME Group, fell 34 per cent in the first nine months of this year from a year earlier, largely because of confusion over tariffs. …

Metal stored in LME warehouses in countries including the US is deemed to be in free-trade zones where goods are stored on a “duty unpaid” basis. Comex offers trading on a similar basis for some metals but not copper, which is stored in Comex warehouses on a “duty paid” basis, meaning all taxes must be paid before the metal enters its facilities.

As a result, traders rushed to import copper to Comex warehouses this year before the expected tariffs came in, with the sudden surge in demand driving the premium paid by Comex buyers over the LME price to a record high of about $3,000 a tonne.

The Comex price is a combination of (1) abstract copper prices plus (2) bets on tariffs. The LME price is a purer abstraction.

777 Partners is an investment firm that factors structured settlements. You get hit by a car, you sue, you get a settlement, the settlement provides you with $200,000 a year for 10 years. You decide that you don’t want $200,000 a year for 10 years; you want all the money now. Some company — JG Wentworth has the best jingle, but 777 Partners does it too — will give you, say, $1 million now in exchange for the right to your future payments. Or if you win the lottery, elect to get paid over time and then change your mind, a factoring company will buy your future payments from you. Other things like this: You are owed a series of future cash flows, you want cash now, a factoring company will make the trade.

The factoring company will do a lot of this, and will want to recycle its own capital. If it has $100 million and spends it all to buy structured settlements that will pay it $20 million a year, then that’s a good trade and it will want to do more of it. It can just wait 10 years, get paid back more than it spent, and then use the profits to buy more structured settlements, but that seems kind of inefficient. This trade is available now, and the going is good; the factoring company should buy more structured settlements now, but it can’t because it has spent all of its money. But it can raise more money. It can package the cash flows into securities, sell the securities to investors, and reinvest the proceeds. Or it can borrow against the cash flows: It can find a lender and say “hi I have $100 million worth of structured settlements, will you give me $90 million to buy more?” The lender will take the structured settlements as collateral; as the cash comes in, the lender will get the first $90 million and the factoring company will get whatever’s left over.

This is an unglamorous but potentially lucrative business. One reason you might go into an unglamorous but lucrative business is that, if you do it successfully for long enough, you will have enough money to get into glamorous but less lucrative businesses. For instance you can buy sports teams! Owning a sports team is not necessarily lucrative, but it is extremely fun and glamorous. And in fact 777 Partners became an owner of multiple European soccer teams and tried to buy Everton.

The traditional approach here is to do structured settlement factoring, make a lot of profits, get rich over time, and eventually have enough money to buy soccer teams. It is not the only approach. The factoring company is recycles its capital; it buys structured settlements and then borrows against them to buy more. Why not buy structured settlements and borrow against them to buy soccer teams? The factoring company is essentially in the business of translating future cash flows — its customers’ and its own — into cash now. If its future profits will pay for soccer teams, why not buy the soccer teams now?

This does not quite work, though, because the company has to borrow against its structured settlements to buy more structured settlements, to continue doing and growing its business. The money is needed in the business; it can’t just be spent on soccer teams. Unless! Unless it were to borrow against its assets twice: once to buy more assets for the business, and once to buy soccer teams. Just an efficient use of resources.

Obviously this is bad: The lenders want the structured settlements as collateral, and if you borrow against them twice then one lender doesn’t have collateral. But we have spent the last few years in a pretty chipper credit market and maybe no one will check.

Last year a lender called Leadenhall Capital Partners LLP sued 777 Partners and its founders, Josh Wander and Steven Pasko, claiming that it was double-pledging assets. From the complaint:

Leadenhall discovered that over 1,600 assets worth approximately $185 million, which 777 Partners had purportedly pledged to Leadenhall, had in fact been “double-pledged”—i.e., the same collateral had been pledged to both [another lender] Credigy and Leadenhall.

During conference calls in March and April 2023, recorded with Wander’s permission, Wander attempted to defuse the situation by acknowledging just the tip of the iceberg, referring to the double-pledged collateral as “embarrassing,” a “mistake,” and a problem that he vowed to resolve—and freely admitting that 777 Partners had breached the parties’ agreements while insisting that it had done so inadvertently. But, to conceal his broader scheme, Wander lied, assuring Leadenhall that the collateral shortfall in breach of the LSA was the result of a recording glitch that could and would be easily remedied. 

At the time I called this “an absolutely wild lawsuit,” but those were simpler times. Now I constantly find myself writing about incidents where companies allegedly double-pledged receivables, and this one seems pretty normal. “When you see one cockroach, there are probably more,” Jamie Dimon said last week, and he keeps being proven right. He was talking about Tricolor Holdings, but we have talked about similar problems at First Brands Group and Cantor Group, and it is becoming a trend. Why do lenders keep letting companies double-pledge receivables? Well, you know, the lenders have a lot of money to put to work, and they want receivables, and the companies that double-pledge their receivables can give them more receivables than the ones that don’t. Excessive enthusiasm for lending creates excessive loans. “Late-cycle accidents,” Marc Rowan called them.

Anyway last week Wander was charged with fraud by US federal prosecutors and the Securities and Exchange Commission. From the indictment:

By 2021, JOSHUA WANDER, the defendant, had directed significant spending on 777 Partners’ increasingly expensive acquisitions. … Around this time, rising interest rates and increased competition in the structured settlements market diminished the profitability of 777 Partners’ operations. As a result, 777 Partners’ cash and borrowing capacity were reduced. …

As 777 Partners’ borrowing capacity and excess cash dried up, JOSHUA WANDER, the defendant, directed employees … to use restricted funds from the firm’s structured settlements business to continue to finance its business expansion … When 777 Partners borrowed the funds, WANDER and others acting at his direction falsely represented that the funds would be used only for purposes expressly contemplated in the agreements with the lenders, namely, to buy structured settlements that would serve as collateral for the borrowing. For example, in or around May 2021, 777 Partners impermissibly used funds from Lender-2’s credit facility to cover expenses of the London Lions and Sevilla FC. 

The indictment also mentions the Leadenhall calls: 

On or about March 28, 2023, JOSHUA WANDER, the defendant, spoke by phone with representatives of Lender-1 about the double-pledged assets. During the call, which was recorded, WANDER falsely claimed that the double-pledging of assets was the result of a “screwup” caused by 777 Partners’ “antiquated” computer system. WANDER further claimed that he was “hopeful” the problem would be fixed by substituting “replacement deals” that 777 Partners had found in its files for the double-pledged collateral.

“Ehh we lost track of our assets but don’t worry, we’ll find you some more” really is the sort of thing that you say near the peak of a credit cycle.

One important feature of cryptocurrency is that you can hold arbitrary amounts of it yourself, electronically, without a custodian. Normal currency doesn’t work that way: You can keep $20 bills in your wallet, but if you have $20 million you will probably have to keep it in a bank or a money market fund or something. In a fundamental sense, dollars exist only as someone else’s debt, which might annoy you if you are paranoid or especially tidy-minded. “No,” you say, “I don’t want debt of Citigroup, I just want dollars,” but you can’t have them.

And, in a fundamental sense, crypto was meant to solve this. “A purely peer-to-peer version of electronic cash would allow online payments to be sent directly from one party to another without going through a financial institution,” says the first sentence of the Bitcoin white paper. Bitcoin, and cryptocurrency generally, is a form of electronic money that is not anyone else’s debt, which means that you can just have it. You can have a digital wallet containing your Bitcoins; nobody else needs to be involved.

And then the history of crypto since the Bitcoin white paper is largely about people building businesses that will hold your crypto for you, that will transform your nice atomic trust-free crypto into the debt of someone else. It is not, like, outrageously difficult to hold your crypto yourself; lots of people do it, and some of them are quite evangelical about it. (“Not your keys not your coins,” they say, to people who trust their crypto to others.) But it is somewhat more convenient to let someone else hold on to your crypto for you, and so lots of people cheerfully park their crypto at Mt. Gox or QuadrigaCX or Celsius or FTX. Or Coinbase; it’s not all bad.

In the long run, as crypto becomes more mainstream, more people are going to end up parking their crypto at big mainstream financial institutions. Some of those institutions (Coinbase?) might be companies that start out in crypto and then, with the rise of crypto, become big and mainstream and general-purpose. Others (Robinhood?) will be companies that start out in traditional finance, but with a youthful and tech-forward approach that makes them early to crypto. And some of them will be BlackRock. Bloomberg’s Emily Graffeo reports:

A new generation of [exchange-traded funds] is giving the crypto rich a novel way to fold their digital fortunes into the regulated financial system — without selling, and through funds run by big asset managers like BlackRock Inc.

A regulatory change this summer opened the door for large investors to hand their Bitcoin to an ETF in exchange for shares of the fund. It’s called an in-kind transaction and is used across most ETFs, but was only approved for Bitcoin products this July. The process is generally tax-neutral, whereby no cash changes hands and no sale is recorded. The result is that a volatile digital asset becomes a line item on a brokerage statement — easier to borrow against, pledge as collateral, or pass onto heirs. …

BlackRock has already facilitated more than $3 billion of these conversions, according to Robbie Mitchnick, its head of digital assets. … Large Bitcoin holders are waking up to “the convenience of being able to hold their exposure within their existing financial adviser or private-bank relationship,” among other reasons for converting, Mitchnick said. …

By exchanging their Bitcoin for ETF shares, investors can keep the same stake in the cryptocurrency while moving it into a form the financial system recognizes. Inside a brokerage account, that holding can be pledged as collateral, borrowed against or included in estate plans — things that are cumbersome, risky or impossible when assets sit in a private digital wallet. The ETF wrapper offers legitimacy and ease, turning what was once off-grid wealth into something banks and advisers can work with.

Yes the highest form of crypto wealth is obviously turning it into a stock.

I often say around here that every bad thing that a public company does, and every bad thing that happens to a public company, is also securities fraud. There are some important qualifiers to that, though. Most notably, the measure of “bad” is that the stock has to go down. The “everything is securities fraud” theory is that bad news comes out, the stock drops, and shareholders sue, saying “we didn’t know about the bad thing so we were defrauded.” And they sue for damages equal to, basically, the amount that the stock dropped. If the stock doesn’t drop the whole thing doesn’t work.

I mean you could try! You could say “sure the stock went up 2% when the bad news came out, but there was other unrelated good news, and if not for the bad news the stock would have been up 5%, so if you think about it we were still defrauded.” But that seems harder. [1]

Anyway, yesterday “a glitch with an obscure Amazon database disrupted life for millions of people across the U.S. as core internet services failed to function for an array of companies.” You might think that would be bad for Amazon. Maybe it will owe its customers credits under its service agreements, maybe people who were affected will sue for damages, maybe potential customers will have second thoughts about using Amazon Web Services and its revenue will go down. But in fact Amazon.com’s stock was up yesterday, and it was up some more this morning, so no harm no foul. 

One possible intuition here is “if the entire world shuts down due to an Amazon glitch, that drives home that Amazon is really important and I should buy the stock.” Though “the trading platform Rob