Morning Coffee: How to make a fortune at Deutsche Bank in a down market. The firm where all the analysts are better than mediocre
When you’re negotiating with an investment banker to acquire an investment bank (for example, if you’re Deutsche Bank and you’re buying Numis), what kind of deal are you hoping for? The answer’s not necessarily as straightforward as one might think; in fact it’s surprisingly similar to a classic paradox. On the one hand, cheaper is better, so obviously you would want to underpay and get the acquisition as cheap as possible. On the other hand, any bank that would sell itself for less than fair value is probably a franchise you shouldn’t be acquiring at all. So maybe the successful deals are the ones where you overpay, in the short term at least.
Deutsche might be having exactly these kinds of mixed feelings right now; back in April the acquisition was described as “a vote of confidence” in the UK market. In one of its last trading updates before the deal completes, Numis has referred to a “deteriorating market backdrop” and described the state of the IPO market as “effective closure”. It looks as if this sale was pretty well-timed for the vendors (and for members of the management long term incentive plan)
But of course, it was not a secret back in April that London IPOs weren't exactly at the top of the cycle. Deutsche isn't commenting publicly on the deal, but “a person familiar with the matter” reminded the FT that “tough industry conditions are well known” and the trading update “does not change the assessment of the deal”. And in fact, the M&A revenues at Numis are up on last year, suggesting that the corporate client relationships which were the real rationale for the deal are still strong and capable of generating value in the longer term.
Sam Goldwyn famously said of a Hollywood actor that “we’re overpaying him, but he’s worth it”, and the same thing can be true in banking. Large and profitable European national champions can afford to have a planning horizon that stretches over multiple industry cycles, making it possible to strike prices which are both extremely generous to the bankers and investors getting the money, and commercially sensible as a cheaper way of picking up a franchise than building it organically. The people who might really be upset about the Numis trading update could be the other UK domestic boutiques, who will be taking exactly the same kind of revenue pain, but with second choice (or none) of potential buyout partners.
Elsewhere, Greg Jensen of Bridgewater has been testing how a custom version of ChatGPT that might be brought into the company’s famously rigorous investment process. Suggesting that the latest Large Language Models seem to perform at or about the 80th percentile on reasoning tests, he says that “all of a sudden if you have an 80th percentile investment associate, technologically, you have, you know, millions of them at once. And if you have the ability to control their hallucinations and their errors by having a rigorous statistical backdrop, you could do a tremendous amount at a rapid rate. And that's really what we're doing in our lab and proving out that process can work.”
Which sounds exciting, but … from a “rigorous statistical backdrop” point of view, if you add a million analysts at the 80th percentile, that’s not going to be the 80th percentile anymore. In fact, Bridgewater currently only has 1500 staff; if we moved to a world in which more than 99.9% of the investment professionals were different implementations of the same AI, it’s not obvious that percentile distribution would be meaningful at all.
And of course, Bridgewater wouldn’t be the only company in the world using AI bots in its investment process. In order to generate investment outperformance, you need to have an “edge” – that doesn’t just mean being objectively good, it means being better than everyone else in the market.
As veterans of high-frequency trading can testify, if the machine learning revolution ended up turning into a massively expensive technological arms race that ended up competing away nearly all of the benefits and just going to make a small number of techies very rich, it wouldn’t be the first time this had happened. It makes you realise why people buy index funds.
Possibly it’s a generational thing, maybe they just learned odd habits in the pandemic, but (allegedly) it seems that (some) members of Generation Z (maybe) are much happier to talk about their sex life in the office. (WSJ)
Former Morgan Stanley derivatives trader Judy Joo now runs a restaurant called “Seoul Kitchen”, and thinks it’s important to treat staff well “Unpaid breaks in the middle of the day with no place to go, physically fighting for a staff meal that’s just kitchen scraps, having one day off, working two days and then another day off — it’s just not a nice way to live your life” she says, although it sounds slightly better than being a junior banker. (Bloomberg)
While you’re waiting for an opportunity to arise to make big money off the SEC whistleblower program, get into the snitching habit by making four-figure sums out of noise complaints about bars and restaurants (NY Post)
Younger readers might not remember “slot machines”; they’re kind of like crypto trading but better regulated and you’re not allowed to pretend to own the funny pictures of monkeys. Like crypto, however, it’s apparently possible to lose hundreds of thousands of dollars on them, but make it all back by monetising your content for streaming. (WSJ)
Unlike some of the independent oil trading firms, Exxon pays its traders on a salary-and-bonus basis like engineers, rather than a share of revenues. They’re now announcing an intention to build a new “global trading” division, but if they aren’t paying up, it’s hard to see how. (Bloomberg)
Christian Channell, Goldman Sachs’ global head of product control and a partner since 2016, has gone to be CFO of a “cloud native” payments processing fintech. (Finextra)
Have a confidential story, tip, or comment you’d like to share? Contact: +44 7537 182250 (SMS, Whatsapp or voicemail). Telegram: @SarahButcher. Click here to fill in our anonymous form, or email firstname.lastname@example.org. Signal also available
Bear with us if you leave a comment at the bottom of this article: all our comments are moderated by human beings. Sometimes these humans might be asleep, or away from their desks, so it may take a while for your comment to appear. Eventually it will – unless it’s offensive or libelous (in which case it won’t.)