Money Stuff: Banks Made a Lot of Money on Stocks

Here are the three main ways for a big bank to make money from stock trading:

1. The bank’s clients want to trade stocks, so the bank helps them out, buying the stocks the clients want to sell, selling the stocks the clients want to buy and capturing some spread. This is called “market making” or “intermediation,” and tends to do best when markets are volatile. For one thing, when markets are volatile, clients are buying and selling a lot. For another thing, when markets are volatile, the bank — if it is good — will often be buying low and selling high, as customers panic and sell low and then get enthusiastic and buy high.
2. The bank’s clients want to borrow against their stocks, to buy more stocks, and the bank lends them money to do that. This is called “prime brokerage” or “equities financing” or “margin lending,” and has become increasingly important to banks. It tends to do best when markets are going up. For one thing, when markets are going up, people get more enthusiastic and want to borrow to buy more stock. For another thing, when markets are going up, the banks’ collateral gets better: If you lend $100 against $200 of stock and the stock goes to $300, you’re in great shape; if the stock goes to $90 you’re in trouble.
3. The bank’s corporate clients want to raise money by selling stock — in initial public offerings, or in follow-on equity offerings — so the bank helps them out, finding buyers for their stock and collecting a fee. This is called “equity capital markets,” or “ECM,” and banks tend to count it separately from the other two categories. (ECM is usually part of investment banking fee revenue, while market making and financing are part of the equities trading division.) This business also tends to do best when markets are going up and reasonably calm: Companies like to sell stock high and don’t like surprises, so they tend to sell stock when “the window is open” and confidence is high.

Those are the important conceptual buckets. Another bucket is prop trading — the bank buys stocks that it thinks will go up, and they go up — which is largely disfavored these days. A fifth bucket is derivatives trading — the clients want to hedge their stocks or do other bets by trading derivatives — which is mostly a subset of market making. [1]

These things to some extent offset one another. When markets are extremely volatile, you can make a killing with smart trades but can lose a lot of money on bad margin loans. When markets are placid and rising, you can get a lot of ECM fees but your traders might be bored. Occasionally markets are volatile but also up a lot; then you make a lot of money everywhere. Bloomberg’s Sridhar Natarajan reports:

Five of the biggest US banks reported second-quarter results Tuesday with their divisions handling equities shattering one record after another. JPMorgan Chase & Co. reeled in $6 billion from stocks alone, a personal best, while Goldman Sachs Group Inc. hauled in $7.42 billion, a new industry high. Underscoring the trend, a 45% jump at Citigroup Inc. was seen by shareholders as too shallow.

The biggest trading desks are getting inundated by clients’ constant repositioning as the artificial intelligence boom shrugs off any note of caution. Even as the US military wades deeper into a war on Iran and rising affordability concerns pinch Main Street, Wall Street’s equities desks keep coming out as winners.

“The markets are booming right now,” JPMorgan Chief Executive Officer Jamie Dimon said after posting record numbers. “It’s getting as close to as good as it gets. We just don’t know how long it’s going to last.” …

That’s aided by wide-open capital markets, as sovereign funds and mom-and-pop investors gorge on new equity offerings, capped off by the record listing for SpaceX — the tweets-to-rockets empire that made Elon Musk the world’s first trillionaire.

On Goldman’s earnings call, [2]  Chief Financial Officer Denis Coleman said:

The activity for our equities business has been very broad-based. It's across intermediation, and financing. In intermediation, it's across cash activities and derivative activities. And there’s a dynamic in the marketplace right now that where we’re seeing a single-name equity dispersion relative to index and we're seeing that dispersion relative to index while the market is going up. And that sort of concoction is very supportive for the activity that clients are undertaking to manage their own portfolios and their own returns and it is causing them to come to us to assist them in managing that dynamic.

That is: The market for individual stocks is quite varied and volatile; some stocks do much better than others (“single-name equity dispersion”), clients want to trade a lot to make sure they own the good stocks and not the bad ones, so Goldman has many opportunities to buy low and sell high. But the stock market overall is going up a lot, so clients mostly want to buy stocks, which is good for financing revenue and also for doing stock offerings.

A lot of this is, of course, AI. Enthusiasm for AI drives the stocks of big AI-linked companies higher; worries about AI disruption causes dispersion among AI winners and losers; AI financing needs drive huge stock offerings. Just last month, SpaceX did an $86 billion initial public offering for SpaceX and Alphabet did about an $85 billion stock offering; SK Hynix did a $26.5 billion offering this month, and Samsung Electronics might be next. [3]  At-the-market equity offerings are hot.

Two themes we have talked about recently are:

1. The re-equitization of the US stock market. It used to be that big public companies generated lots of cash, didn’t need all of it and paid a lot of it back to shareholders; the stock market shrank over time. But now big companies need vast amounts of cash to build out AI, and they are raising it by selling stock into the public stock market. Instead of being net sellers, US public shareholders are now net buyers.
2. The increase in stock market leverage. Investors (retail customers in margin accounts, hedge funds, leveraged exchange-traded funds) are borrowing more money to buy stocks.

You could argue that those themes are related: If companies want to issue more stock, and if all the investors who want stocks already own stocks, the investors might have to borrow money to buy more stocks. [4]  And if companies want to finance themselves with more stock and less debt, banks that lend them money will have money sitting idle. If the banks loan that money to the investors who want to buy stocks, the accounts sort of balance. [5] I wrote a couple of weeks ago:

That sounds like a job for the financial industry: shifting risk around, giving more risk to people who want it. The financial industry is good at putting stuff in boxes and issuing tranches of those boxes. If you have too much equity, you put the equity in a box and issue debt and equity against the box. … You have taken the given mix of debt and equity and used it to create more debt and riskier equity.

For corporate finance reasons, ompanies are creating a lot more equity, which might be hard for markets to absorb in the short run. But the financial industry transmutes some of it back into debt, making it easier to absorb. When big companies want to raise hundreds of billions of dollars of equity all at once, it is the job of the financial industry to make that possible. That’s an expensive job, so equities revenue is up.

The main risks of banking are (1) a bank might make a loan to a customer who doesn’t pay it back and (2) a bank might take deposits from customers who all ask for their money back at once, causing a bank run. Banks know a lot about these risks, and bank regulators know a lot about these risks, and a lot of the purpose of bank regulation is to make sure that banks consider and manage and diversify these risks.

There are a lot of risks in the world, though, you can’t pay equal attention to all of them, and bank regulators are in part political creatures. If there is some group that you find distasteful, it will take you roughly half a second to think of reasons why lending to that group, or accepting deposits from that group, is a risk to the safety and soundness of a bank. Here:

1. Lending to oil and gas companies is risky, because as climate change gets worse the world will move away from oil and gas, leaving the companies’ reserves worthless.
2. Lending to gun companies is risky, because if those companies are held liable for mass shootings they will lose a lot of money and won’t be able to pay you back.
3. Taking deposits from crypto companies is risky, because if crypto prices drop those companies will all take their money back at once.
4. Lending to Donald Trump is risky, because he has some history of misrepresenting his finances and not repaying loans. 

None of these stories are wrong, exactly: There are all sorts of bad things that might happen in the world, and a prudent bank or regulator would want to think about them. A bank that made loans only to oil and gas companies, or that took deposits only from crypto companies, would want to think about these risks a lot, and be really sure that they won’t come true. A banking supervisor might reasonably say to that bank “hey consider diversifying a bit because you’ve got a lot of eggs in this one risky basket.”

But there are a million other hypothetical risks, too, and selecting one to focus on is — or might look like — a political choice. Perhaps a banking regulator might send banks a memo saying “hey, lending to gun companies is risky” solely because she is worried about the financial risk to banks’ balance sheets of lending to gun companies. But you might suspect her of having other motivations.

Anyway if the US government deports an immigrant, that immigrant might be less likely to pay back a bank loan, so:

The Trump administration is continuing its push to limit undocumented immigrants’ access to the U.S. banking system, urging banks to consider loans to them as potentially risky. ...

Under existing laws, banks have to maintain safe underwriting practices to protect against systemic defaults and their own failure. The latest guidance doesn’t change those standards, but it spells out reasons why offering loans to unauthorized workers could be risky.

The risks include the possibility of deportation, or that an employer terminates workers after discovering they don’t have valid work authorization. That would likely affect the borrower’s ability to repay the loan, the agencies said.

Banks aren’t prohibited from offering accounts or loans to undocumented immigrants, but the changes undertaken by regulators following Trump’s executive order appear intended to make it harder for them to justify to regulators why they might be doing so. Such warnings tend to put a chill on bank activity.

The crackdown on serving undocumented immigrants comes at the same time that the Trump administration and regulators are also investigating banks for allegedly “debanking” clients for political and religious reasons.

Not this, though, this is not for political reasons; this is just about prudent loan underwriting.

In a leveraged buyout, you buy a company by borrowing against the company. A company is for sale for $100, you put up $30 of your own (and your investors’) money to buy it, and you have the company itself borrow the other $70 to cover the rest of the purchase price. People get miscellaneously angry about this sometimes, though it’s basically the same as buying a house with a mortgage. (You put up $20 of your own money and borrow $80 from a bank secured by the house, etc.) Eventually you sell the company for, like, $150, pay off the $70 of debt and keep $80 for yourself, a $50 profit.

Those numbers are made up. In the swashbuckling glory days of early LBOs, you might put up 10% of the money yourself and borrow 90%. These days, much lower levels of debt — maybe 50% your money, 50% borrowing — are more common. It is just about possible to imagine an LBO financed with 100% debt, though it mostly seems stupid. If the target company can borrow $100 to pay to its shareholders, then (1) its equity should be worth more than $100 and (2) what do the shareholders need you for? [6]

That said, Bloomberg’s Martin Braun reports on a water-company LBO in Connecticut:

The Aquarion Water Authority — a local agency created by Connecticut’s legislature — on Tuesday is planning a $2.4 billion bond sale, the biggest in the state’s history, to pay for the water company it’s buying from Eversource Energy, an investor-owned company.

The move marks a relatively rare takeover by a government agency, which in recent decades have often moved to sell off assets to raise cash or outsource services to for-profit companies. It also pushes the new agency into the role typically played by corporate raiders, with some of its 220,000 customers angered by its plans to push up rates 60% over the next decade to help cover the cost of its debt and pay for needed construction work. …

The buyout has been in the works since mid-2024, when a bill passed during a special session of the legislature included a provision to create an authority empowered to purchase Aquarion, which serves nearly 60 municipalities, including Greenwich, Stamford and Westport. …

In its filing with regulators, Aquarion said converting to a public entity would benefit ratepayers by allowing the utility to finance infrastructure using lower-cost tax-exempt bonds. And as a nonprofit, it won’t have to pay shareholder returns or corporate income taxes. It estimated customers will ultimately save $366 million over the first 10 years.

As far as I can tell, the $2.4 billion bond offering ($1.83 billion of A- senior bonds and $547.2 million of BBB+ subordinated bonds) accounts for approximately 100% of the $2.4 billion purchase price, which is weird for an LBO but sensible for a government project: The local water authority can’t issue equity, and doesn’t have piles of ready cash lying around; if it’s going to raise money it will issue bonds. Also it doesn’t want equity ownership in the classic sense; it’s not buying the water company hoping to flip it at a profit in five years. It’s buying it to deliver water to residents. Financing the buyout with 100% debt makes sense.

Of course most companies can’t borrow 100% of their value with investment-grade ratings, but publicly owned utilities are different. For one thing, utility cash flows are sort of bond-like anyway: You are securitizing the rates that you can charge customers in the future, people need water and you’re a monopoly. For another thing, Aquarion isn’t entirely borrowing against itself; the fact that it is publicly owned probably helps its credit. “We view positively the willingness of the representative policy board to apolitically approve raising rates when necessary,” S&P Global Ratings said in rating the bonds. 

I used to have a recurring bit around here titled “blockchain blockchain blockchain,” the point of which was that for a while just putting the word “blockchain” into some gibberish could capture a lot of attention. In particular, I sometimes suggested that, if you had some idea for improving back-office functionality at a bank, and you walked into the chief executive officer’s office and said “I have an idea for improving back-office functionality,” security goons would pick you up by the scruff of your neck and throw you out, because how dare you talk about back-office functionality to the CEO. Whereas if you began your pitch by saying “blockchain,” the CEO would be all ears, because people were really into the blockchain for a while.

Now they aren’t so much, but the Financial Times checks in on SoftBank’s Masayoshi Son:

“Those who dislike AI have essentially refused their own evolution,” he said. “Those who condemn AI are themselves spitting upwards.” ...

While there has been a public backlash over the perceived impacts of AI on employment, Son expressed his distaste for the hesitant — especially Japanese business leaders who failed to create new giants in the internet era.

“Modest presidents” who are not prepared to embrace AI to become number one in their industry within 15 years “should play the role of spouse”, he said.

“The most important thing for the president to say is ‘AI, AI, AI’.”

I feel like a lot of people have gotten the message that the most important thing in business is to say “AI, AI, AI” all the time. 

Here is a claim that a promoter of a US-listed special-purpose acquisition company that has agreed to merge with an AI data center is also involved in money laundering for a Spanish cocaine kingpin. Exactly how the SPAC promoter might have been involved in money laundering is not quite clear — “through a ‘complex financial instrument,’ according to investigators” — and as far as I can tell the SPAC was not itself involved. That said:

1. In the course of financial history, it is possible that, once or twice, something shady has happened at a SPAC;
2. I can’t really think of how a SPAC would be a great vehicle for money laundering (since the point of a SPAC is that the promoter gets his shares for free); but
3. I am hoping my readers will give me some ideas, because I would read, or possibly write, a novel about this. “A SPAC to turn my cocaine money into AI data centers” is just a very 2026 idea, and if these people aren’t doing it then maybe someone else is.

JPMorgan Bags Record Profit on $6 Billion Stock-Trading Haul. Data-Center Builders Are Racing to Offload Stakes Worth Billions. Can a Dozen Blue States Block the Paramount-Warner Merger? “In retrospect, existential risk has been the best capital-raising tool ever in the history of the world.” DeepSeek Is Preparing For IPO Filing as Soon as This Year. DeepSeek weighs new fundraising a month after closing first round. Buffett to Offload Berkshire Stake in 8 Years, Gates Snubbed. Altman-Backed Brinc Raises $125 Million to Expand US Drone Sales. Job Hunters Are Using AI to Cheat in Interviews, and Failing at the Office. Robotaxi Riders Are Falling Asleep, Sparking Frantic 911 Calls.

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!