One thing that sell-side research analysts do is write reports: long detailed reports situating a company in its industry and giving a buy or sell recommendation, earnings previews pointing out what to watch for, quick updates for news, etc. These reports get published to clients and are historically the most salient and noticeable and regulated thing that the analysts do.

Another thing that sell-side research analysts do is one-on-one phone calls with buy-side investors. An analyst publishes a report, investors read it, and they call her to ask questions like “have you thought about the risk that AI poses to earnings?” or “could you give me some detail about how you got to those growth assumptions?” or “when you met with the chief executive officer last month was he unfeasibly tanned?” The investors might get more out of the phone calls than they do out of the reports; certainly they get something different. Every client can read the same reports, but the clients who ask the best questions get the best answers on the phone calls.

A third thing that sell-side research analysts do, I guess, is record little videos of themselves talking about their reports? Like you write a report about the payments industry and then someone on the video team comes to you and is like “hey some of our clients prefer watching two-minute videos rather than reading reports; could you summarize the highlights of your report in a video?” And then you are like “what, no, I don’t believe you, surely our sophisticated hedge fund clients actually want to see the details of the model, not just watch me talk awkwardly into a camera?” And the video people are like “I know it seems implausible, but it’s 2025, get with the program you dinosaur.” I don’t know if any of this is true. I might be projecting.

I assume that eventually artificial intelligence models will be better than the research analysts at producing the research reports, which will be an existential problem for the research analysts.

Will the AI models also be better than the research analysts at doing one-on-one interactions with buy-side investors? Ehh, maybe; that’s a more complicated question. An AI chatbot might be substantively better at answering detailed follow-up questions from an investor; it might have more information and more patience and better analysis. On the other hand the human element might be important. The (human) investors might like the human analyst and want to talk to her. Also some of the value of the human analyst is that she talks to lots of investors and can be a clearinghouse for their views; that is an implicit function of a human analyst, and an AI will have to either make it explicit (by saying “be warned that I may use what you tell me to inform my conversations with other investors” [1] ) or not do it.

When will the AI models be better than the research analysts at recording dutiful video versions of themselves reading their research reports into a camera? The Financial Times reports that the future is now:

UBS has started using artificial intelligence to turn its analysts into avatars, sending videos of the simulated bankers to clients in a move the lender said would free up staff to focus on more productive tasks.

The Zurich-based bank is using OpenAI and Synthesia models to create AI-generated scripts and avatars of its analysts following increased client demand for research in video format, UBS said.

“It’s not a parlour trick,” Scott Solomon, head of global research technology at UBS’s investment bank, told the Financial Times. …

To create a video, UBS analysts can use a language model to analyse their reports and generate a script. They can then review the script before it is turned into a lifelike video using their avatar. ...

Solomon said that UBS had seen greater demand from clients for video content in recent years, amid a rise in the popularity of short-form video apps such as TikTok.

“Think about how we, in our consumer lives, consume so much more video content now than we did five years ago,” he said. “We publish about 50,000 documents a year, [but video production] has been fixed at about 1,000 a year, because that’s basically our studio capacity. But the number of views on those videos has gone up dramatically.” 

In like a month you’ll be able to watch a dead-eyed AI avatar of me droning a summary of this column into a camera, which I guess is what you want. The good news is that I won’t have to do it!

We talked yesterday about Moody’s Ratings’ decision to downgrade the credit rating of United States government debt from Aaa to Aa1. For a long time US government debt was the safest sort of debt, with perfect AAA/Aaa ratings, but now all three of the biggest US ratings agencies — Moody’s, Standard & Poor’s and Fitch — give the US their second-highest rating (AA+ or Aa1).

The main point I wanted to make was that, while Moody’s does of course have some specialized expertise in evaluating the creditworthiness of sovereign borrowers, it’s not, you know, that much. A lot of people pay attention to the fiscal situation of the US government! Many of them are, in various ways, professionals at it; there are economists and politicians and research analysts and investment managers and pundits with a professional interest in evaluating the creditworthiness of the US, and if you turn on a television at any time of day or night you can hear their opinions about it. It is possible that the Moody’s team is the single best evaluator of the creditworthiness of the US government — that Moody’s ratings are “correct” in the way that, I don’t know, Ray Dalio’s or Ken Griffin’s or Larry Summers’ or Stephanie Kelton’s or  Scott Bessent’s or Elon Musk’s opinions are not — but it would honestly be pretty weird.

Instead, Moody’s decision is salient for two reasons. One is that it is sort of a symbolic lagging indicator; Moody’s downgrade arguably suggests a broadening market consensus that the US is less creditworthy than it used to be. Moody’s main product is its credibility, and arguably it wouldn’t go out and do a downgrade that would get a lot of attention unless it thought that the market more or less agreed with it.

But I think the more important reason is that, in most of the world of debt, a downgrade by Moody’s is in some sense binding. Various investors are constrained by various rules to only hold debt with certain ratings, only recognized ratings agencies can assign those ratings, and so when a ratings agency lowers its rating on an issuer, that can force some investors to sell the debt and thus raise the issuer’s borrowing costs. This is not (just) a matter of reasoned analysis and specialized expertise; it is a matter of legal and contractual requirements. Investment-grade bond fund managers don’t read a Moody’s downgrade report and think “yes this company does seem to be too risky for me, so I will sell its bonds.” They see that Moody’s downgraded the company to Ba1, they are only allowed to hold bonds rated Baa3 or better, so they sell the bonds. 

And, I argued, that could be true of US government debt, but it mostly isn’t. You could imagine a lot of investors having mandates or contracts or regulatory requirements saying “you can only hold debt rated Aaa by Moody’s,” and having to dump US Treasuries on the downgrade, but empirically that seems mostly not to be true. Instead, most of the relevant regimes are more like “you can only hold debt with a rating of at least ____, but also you can hold US Treasuries no matter what, those are Treasuries.” So the downgrade doesn’t have the same effect on US government debt that it might have on a more normal issuer.

But that is just mostly true, and if you look hard enough surely you can find some investor that might be required to sell US government debt on a downgrade. Bloomberg’s Greg Ritchie and Echo Wong found Hong Kong pension managers:

Funds operating under the city’s HK$1.3 trillion ($166 billion) Mandatory Provident Fund system are only allowed to invest over 10% of their assets in Treasuries if the US has a AAA or equivalent rating from an approved agency. After last week’s cut by Moody’s, the only remaining such score is from Japan’s Rating & Investment Information Inc.

The Hong Kong Investment Funds Association has raised managers’ concerns to the Mandatory Provident Fund Schemes Authority and the Financial Services and the Treasury Bureau, said the people, asking not to be named because the information is private. The association recommended that authorities make an exception for US Treasuries, by allowing funds to invest in the assets even if they are rated one notch below AAA, the people said.

The situation underscores the risks of the US falling foul of the unusually strict investment mandates governed by Hong Kong laws. The bulk of global investors do not require the top-tier rating to invest freely in US Treasuries — a factor which minimizes the risk of forced sales.

Note that the fund association suggested the obvious solution here, which is not “diversify away from US Treasuries, which Moody’s has alerted us are riskier than we thought” but rather “change the rules to allow us to hold Treasuries no matter what their rating is.”

One thing that I like to say about private credit is that we are in a weird, probably brief in-between place where a company or private equity sponsor that wants to borrow money will have to have two sets of meetings. It will meet with investment bankers, who will say “here are the terms that we could get you on an underwritten bond offering or a syndicated loan.” Then it will meet with private credit lenders, who will say “here are the terms that we would give you on a direct loan.” And then it will decide on one or the other or occasionally a mix. And that is obviously inefficient and bad customer service, and in the long run investment bankers will just have to show up at their meetings and say “here are the terms that we could get you on a bond deal, or syndicated loan, or private credit loan.” There just has to be a third column in the pitchbook. 

For boring reasons, putting that third column in the pitchbook appears to require some reorganization of investment banks, so that is happening. We talked in January about Goldman Sachs Group Inc. combining its bankers who advise on regular credit (bonds, broadly syndicated loans) with its private credit teams, and now here’s HSBC:

The London-based lender said in a statement Friday that it was creating a new Capital Markets and Advisory group to house all of its disparate worldwide financing and investment banking activities under a single management structure, confirming an earlier Bloomberg News report. …

HSBC is among lenders seeking to step up offerings in private credit, a $1.6 trillion global asset class that’s luring more and more players as demand rises from borrowers seeking safer options amid the chaos set off by the Trump administration’s trade policies. Among attractions is higher management fees. …

Under HSBC’s new set-up, the bank’s financing solutions units, which includes its debt capital markets, leveraged and acquisition finance, and private credit operations, will sit alongside its corporate finance and strategic advisory businesses.

Adam Bagshaw, global head of investment banking, will take charge of CMA with a mandate to grow the lender’s private credit business, as well as consolidating its position as one of the leading finance businesses in Asia and the Middle East.

Right now, private credit has a certain novelty, but in the long run corporate lending tends toward a certain fungibility.

Similarly: One important selling point of private credit, to borrowers, is that if things go wrong for a borrower, private credit lenders will be more relationship-focused, and more willing to work with the borrower to fix things, than public lenders would be. Is that true? I mean:

• probably, but
• we are in a big private credit boom and let’s see what happens as defaults increase.

So Bloomberg’s Edward Clark reports:

Private credit funds are increasingly turning to restructuring advisers to sort out their problem loans, suggesting that while the hot asset class is still raking in capital, there are risks simmering under the surface.

“The amount of work we’ve been doing with private credit has doubled in the last three to four years and we weren’t doing any ten years ago,” said David Morris, head of UK restructuring and chief operating officer of EMEA corporate finance at FTI Consulting. …

The opaque nature of private credit means that some distress can stay hidden. Fixes for struggling companies like covenant waivers and maturity extensions can happen behind closed doors. But consultants will come in when a situation is beyond this point, and if lender thinks a debt-for-equity swap is necessary to preserve value.

In a huge boom for private credit, the most important skill of a private credit manager is putting money to work in new deals. If the money dries up and the old deals start defaulting, the most important skill might be maximizing recovery.

Other people must have a more sophisticated understanding of memecoins than I do. As I understand it, the way a memecoin works is:

1. There’s some thing or person that people are paying attention to, a “meme.” A squirrel goes viral on the internet, say.
2. Someone — who may or may not be affiliated with the meme — launches a crypto token referencing the meme and markets it as being somehow connected to the meme. “The PNUT token is the token of Peanut the viral squirrel,” etc.
3. People are like “oh that meme is cool” and they buy the token as a way to express their fandom for the meme.
4. The token goes up as a lot of people pay attention and express that attention by buying tokens.
5. Squirrels do not go viral for that long! Eventually — in, like, a day? a week? — people stop paying attention to the meme, there are no buyers for the memecoin, and the price collapses to approximately zero.

You can have variations on this. A meme can go viral more than once, if the squirrel, you know, does a second thing. [2]  Some memetic things are pretty enduring. The silly-talking Shiba Inu dog “Doge” has been a meme since 2013, has a US government department named after him, and also the Doge cryptocurrency has taken on a life of its own even apart from the dog. Donald Trump has gotten a lot of attention for decades. 

But when we talk about, you know, Hawk Tuah Coin, or $MELANIA, I do not understand what other people see that I don’t. “Hawk Tuah” got the purest imaginable 15 minutes of fame — just one viral video on someone else’s YouTube — and it got turned into a memecoin. The memecoin went up, during the 15 minutes, and then down, afterwards. 

Yet I keep seeing people who are somehow surprised and disappointed that some memecoin was a “pump and dump,” that the insiders who promoted the coin made a lot of money while later buyers lost money. There’s nothing else! That’s what this whole thing is! The whole pitch is “buy this coin while everyone is paying attention to it”; the whole game is knowing if you are buying in Minute 13 or Minute 15. I am aware that some memecoins have lockups or vesting schedules for their promoters, so the promoters can’t sell all their coins and tank the price too quickly, but the natural lifespan of attention is gonna tank the price pretty quickly anyway, and if the promoters aren’t making any money then what are they doing?

Anyway Donald Trump’s meme token — “$TRUMP” or “Trumpcoin” — is a somewhat unusual memecoin in that Trump has been getting attention for decades, and he does keep doing things. But the essential analysis is the same, just repeated. He does a thing, the token goes up, people sell it, it goes down. We have talked a couple of times recently about the latest thing, which is that he announced he would have a dinner party where he invited whoever bought the most Trumpcoin. “The democratization of … bribery,” I called it.

There was, as it were, a record date for the dinner: Whoever owns the most Trumpcoins as of May 12 gets to go to the dinner. It is now May 20. If you bought enough Trumpcoins to be on the list on May 12, you get to go to the dinner. You don’t need them anymore. The buying pressure to get to the dinner pumped the price, and now, you know. The Financial Times reports:

Traders who won a ticket to a banquet with Donald Trump by entering a contest to buy large amounts of the president’s memecoin may have netted multimillion-dollar profits, a Financial Times analysis has revealed.

The competition, announced on April 23, offered dinner to people who held the largest quantities of the $TRUMP memecoin over a specified window in a publicly visible crypto wallet. The contest helped push the price of the token from $9.26 to $15.33 when it was announced.

Many traders started moving tokens out of their wallets as the window closed on May 12 and their place at the dinner at the Trump National Golf Club in Virginia was confirmed — despite the contest website’s exhortation to “Hold Big. Hold Strong. Hold $TRUMP”.

The FT has identified that 16 of the 25 winning “VIP” accounts, which have been awarded seats at “an unforgettable Gala DINNER with the President” plus an “exclusive Reception . . . with YOUR FAVORITE PRESIDENT” and a VIP tour, now have no Trump coins left in their public wallets. The event is due to be held on May 22.

Of course? What else would anyone be doing? Do you think Trump cares? Do you think if you show up and say “lol I dumped all my coins on May 13,” he’ll be like “I can’t believe you do not have long-term confidence in my meme token”? Come on. Do you think the Trump-connected promoters of the coin were selling it into the demand for the dinner? What are we doing here?

Obviously it’s bad for the president of the United States to be doing this nonsense but what are you gonna do. Elsewhere: “A Tiny Bank in Trump Tower Is Enriching the President’s Sons.”

America’s Fiscal Situation Threatens the Good Mood on Wall Street. KKR Says Bonds’ Role as Portfolio ‘ Shock Absorbers’ Is Eroding. Nvidia Pushes Further Into Cloud With GPU Marketplace. GOP’s Tax on Foundations Takes Aim at Billionaire Philanthropy. Crypto Billionaire Accused of Defrauding Creditors, Propelling Industry’s 2022 Collapse. Microsoft-Backed Builder.ai to Enter Insolvency Proceedings. Blackstone Infrastructure to Buy TXNM Energy for $5.7 Billion. Austin’s Reign as a Tech Hub Might Be Coming to an End. Levi Strauss Will Sell Dockers to Authentic Brands for $311 Million. Qatar’s $524 Billion QIA Warns Private Credit Is Getting Crowded. Qatar PM Defends Plane Gift to Trump, Calling It a ‘Normal’ Transaction. “VC-funded millennial slop bowl.”

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!