Exchanges Q3 – all in the data

A few months ago, I would have been writing notes, updating models, calling clients and so on to celebrate the Q3 results from Deutsche B?rse (“DB1”) and LSEG (London Stock Exchange Group) , who helpfully announced within a few hours of each other earlier this week, earning the heartfelt gratitude of analysts everywhere.? Or something.? Now, my friend and ex colleague Hubert Lam is doing this, so if you want to talk about share prices, call Hubert.

Exchanges have become less exchangy

I wanted to step back and ask how the Q3s fit into the broader story about exchanges (something I touched on in my post “is the LSE an exchange?” (spoiler alert – the answer is “yes, but LSEG is a different matter).? The big change in the exchange landscape over the past few years is that exchanges have tried to become less exchange-like.? They have tended to buy businesses which have some overlap with the exchange business, but which diversify away from it.

A stereotypical exchange provides a platform for investors to trade some form of security.? The original exchanges traded cash products (equities, bonds and so on), with derivative markets growing in the wake of cash markets.? Side by side with exchanges (providing access to liquidity in listed markets) you have a range of "almost exchanges" like the all to all FX platforms and power trading platforms (DB1 is active in both of these).

These are good businesses.? They provide an obviously useful function, they are typically oligopolistic (FX isn’t, power is an evolving market), they have deep regulatory moats, they are mainly fixed cost businesses (so have operating leverage) and they generate a lot of cash as they don’t have regulatory capital needs.

The problem with exchanges is that they typically charge a fraction of the value transacted.? Many exchanges have fiercely complex tariffs, and have a membership component, but as volumes go up, so do revenues and vice versa.? The problem with this is that if everybody decides that they can’t be bothered to trade German equities today, DB1’s revenues for the day will suffer.? And sometimes people decide to trade a lot of products, and sometimes they don’t.

Post trade

There are two antidotes to this problem.? Firstly, a lot of post trade businesses are less cyclical.? Custody, for example, typically charges over assets, which are sticky, not volumes.? Certain forms of clearing also have an asset based component (LCH’s IRS business does, for instance).? However, a lot of post trade has a similar revenue model to trading.? For instance, Eurex, DB1’s derivatives powerhouse, doesn’t charge separately for trading and clearing.

Data

The more radical approach is to stress the “exhaust” from trading – data.? Most exchanges sell trading data from their own platforms – see my various posts on the “consolidated tape” for a discussion of this.? It is a short step to move from this to selling other financial data, and this is a step which the European exchanges have all made to varying degrees.?

LSEG has made the most extreme shift - in Q3, around two thirds of revenues came from data.? However, DB1 has been going in the same direction.? Like LSEG, it has been focusing M&A on data-like businesses, such as ISS, the ESG specialist, Axioma, an analytics/factor business and, most recently, SimCorp.? SimCorp isn’t exactly a data business (it sells software to asset managers) but like these, it has a recurring revenue model.? About 15% of DB1’s Q3 revenues were badged as data, but SimCorp will increase that – it generates slightly more revenue than DB1’s existing data business.? These assets are being bundled together as the “Investment Management solutions” segment.

The reason for this strategic move is pretty clear from the chart below.? The blue line is DB1’s overall revenue growth, ex M&A.? The green line is the growth in LSEG’s “Annual Subscription Value” (“ASV”), essentially a measure of subscription business in force at the quarter end.? Lastly, we’ve added in DB1’s secular growth.? This is a measure which the company itself calculates, and investors tend to have mixed feelings about it.? Some feel that it’s a slightly hand-wavey measure, and to me, the boundary between structural and cyclical isn’t always clear.? But I think as a concept it makes sense, and I’ve tended to take it face value.? One last caveat – the LSEG figures are adjusted for FX, DB1’s aren’t (I think they would be sensible to provide FX adjusted data).

So, after all that, here’s the chart.

The big lesson: more data = more stability

There are a few obvious lessons.

DB1’s overall growth has been volatile.? One important reason for this is the net interest income in its Clearstream segment.? This is a function of excess balances (anyone who wants to know what these are only has to ask me...) times short rates, as the balances are deposited with central banks (the largest piece being USD).? This was €13m in Q1 21 (when the chart starts) and €169m in Q3, as central bank rates have risen.? This revenue is for all intents and purposes pure profit. ?One way of looking at DB1’s Q3 is that NII accounted for well nigh all Q3’s YoY revenue growth. ?

However, there are other reasons for the volatility. ?The most obvious is simply that cash and derivative trading volumes are volatile.? For instance, in Q3 23, Eurex’ revenues fell by 10% year on year.? These are a fairly direct function of volumes, although yield also fell a bit.

Turning to LSEG, its mainstream exchange businesses weren’t much better.? Equity revenues fell by 9% (this is a mixture of trading and listing fees – listing is relatively higher quality than trading) and FX fell by 3%.? LSEG owns a majority of TradeWeb, which appears under “Fixed Income, Derivatives & Other”, and this managed to grow even in the dull environment.? TradeWeb continue to benefit from FICC volumes moving towards all to all platforms from bilateral trading.

TradeWeb helped LSEG in Q3, as it has done for many previous quarters.? The big reason why it showed very solid growth in a cyclically difficult quarter, though, was largely the weight of data businesses.

ASV growth trending upwards – self help

The last point I’d make is that although the data businesses are typically steadier than execution revenues, LSEG’s ASV growth has increased steadily over this period, whilst DB1’s structural growth has been pretty consistent.? A key plank of LSEG’s acquisition of Refinitiv (the old Thomson Reuters data business) was that it would be able to boost Refinitiv’s growth – this had been a bit lacking in the past.? This seems to have happened. ?

The most striking part of LSEG’s revenues for a few quarters has been the performance of the so-called “Trading and Banking” segment.? This is largely the desktop business.? The brand LSEG inherited, Eikon, wasn’t especially loved.? We talked a bit about desktop in last week’s post, but one of LSEG’s signal achievements over the past couple of years has been turning this division from shrinking (-0.3% in FY 21), to growth (2.3% in Q3 23).? It’s still a relatively slow grower, but that kind of turn-around in what is the data business’s largest single segment makes a big difference to the overall outcome.

Importantly, there ought to be more to come here, as LSEG’s tie-up with Microsoft should be very helpful for the desktop business.? The desktop is likely in future to include much more capable AI (Microsoft has some impressive assets here) as well as slick integration with the Office products and with Teams – last week’s post highlighted how important messaging has been to desktops.

So, Q3 is just one quarter, but it shows clearly why exchanges are looking to diversify, and to increase their exposure to stable revenues.? The market will pay more for this stability than the same overall growth with a sawtooth growth profile.

Next week, I really will return to DLT, unless waylaid by news.