Quick Takes: 2023 SEF volumes and share in SOFR swaps

Ali Curi: Hi everyone and welcome to ION Markets Quick Takes. I’m Ali Curi and every week along with my guests Amir Khwaja and Chris Barnes, we take a quick dive into the headlines on the Clarus blog. Let’s get started.
Hi Amir.
Hi Chris.

Amir Khwaja: Hi Ali.

Chris Barnes: Hey Ali, how are you doing?

Ali Curi: I’m doing great. Welcome back to Quick Takes.
Amir, let’s start with you. What’s your Quick Take for this week, which headline from the Clarus FT blog, would you like to discuss?

Amir Khwaja: Sure Ali, thank you. So I’m doing “SEF volumes and share in SOFR swaps 2023.” I wrote this a few weeks ago, and really I look at market share using our SDRView product of U.S. Dollar SOFR Swaps.

And the way we do that is, I think for over a year now, we now have a platform ID of the SEF venues in the SDR transaction data. So we can categorize these venues as dealer to dealer or dealer to client. And I’ve really looked at the volume in 2023 of surface swaps on dealer to dealer venues and their market share.

So for example, if you look at that blog, it has lots of figures and stats. We talk about, also we enhance the underlying raw data in SDRs to really figure out what is trading from a price forming share of wallet for the SEF venues point of view, right? So the most common trade in SOFR swaps for the D2D sets are spreadovers.

So these are traded as a spreadover to on the run U.S. Treasuries. And I just brought you a few numbers. So in the last year, there were 30,000 spreadover trades done on D2D venues. And ICAP is a market leader with almost half the volume. The next most common trade type is butterfly trades. They have three legs followed by curve switches.

And again, different venues have different market share on those products. So really the blog looks at volumes, market share, both in trade count terms and in DV01 terms. And DV01 terms are a better proxy to the share of wallet that these brokerages or set venues are getting from their customers.

And then on the other side, you have dealer to customer venues. So the way trading happens is that clients trade on D2C venues and then dealers hedge on the D2D venues, that kind of flow amongst themselves. So obviously there’s much higher volume on the D2C venues. So that’s really TradeWeb and Bloomberg dominate that space.

For example, they were over 100,000 vanilla outright top trades on those venues. There’s also lots of forward trades and IMM data trades, et cetera, that clients do. So I think clients typically are taking more outright positions and dealers are hedging through spread overs or with curves and flyers.

So really the blog looks at those stats and data. Trade counts, DV01s by trading venue categorized by D2D and D2C. In the D2C space, both Bloomberg and Tradeweb dominate depending on the type of product that’s traded, so the share varies. And there’s lots of charts that are taken from a SDRView product. And I think that’s really a quick summary of what that blog showed. I was gonna ask Chris if you had any questions.

Chris Barnes: Thanks, Amir. The big thing that jumps out to me is when you look at the difference between D2D and D2C. D2D is dominated by spread overs, right?

A swap versus a U.S. treasury. And the charts suggest that clients just don’t trade those. It’s tiny. Do we think that’s an accurate representation of what clients are doing, or do clients split up a treasury trade versus a swap trade?

Amir Khwaja: Yeah, good question, Chris. Yeah, I think it’s hard to know.

Clearly, there’s some noise in the data, so it may be possible. But I do feel, the majority of client trades will be outright, even in the dollar market, I would assume. What do we see? So it shows 7,000 spreadover trades compared to 100,000 outrights.
Ballpark, that ought to be correct. But you’re right, but if they’re doing reporting in a way that’s unlikely because they should still be doing a package, the identification.

Chris Barnes: To me, the data basically says dealers trade spreadovers, clients don’t, and that is a conscious decision that dealers have made to manage their risk in U. S. markets. And the reason I’m going through that is that there was a recent article on risk, which has a very clever headline called “CLOBbered.” which is about CLOBs, Central Limit Order Books in U. S. treasury trading. And it’s talking about how toxic flows leaked into the Brokertech, CME-owned platform and have really caused a change in market share for U.S.

Treasury trading. And it’s interesting because you typically think of as dealers having that type of edge that either information edge or market makers edge, or a technological edge that allows them to be the leaders of the market. And what that risk article hints at, is that dealers are being picked off in U.S. Treasury flows because high frequency traders and non-bank market makers are so much quicker, are so much better informed in terms of the information that they can get from electronic trading that dealers feel threatened. Dealers feel that these high frequency flows are toxic, and so they are staying away from certain venues.

Now, if you try and tie that in with what this picture is painting of the swaps market, you would think that dealers are very active in U.S. treasury cash markets because all of their inter dealer trading is in spreadovers. But I guess all of those spreadover trades are voice broked. It’s not an electronic transaction whereby they’re breaking up a swap versus a treasury and they’re streaming the prices separately and they’re trying to get filled at the best price. This is actually traded as a pure voice broked package, right? Someone’s saying, “I’m minus eight, bet yours.” That is how it’s traded.

So these volumes are completely outside of those order books that risk is talking about.

Amir Khwaja: Yes, I think I agree. And I think even those number of trades, 30,000 will be less than the number of trades done by high frequency firms in order books in U. S. treasuries, right?

Chris Barnes: Exactly. We’re talking about very large, notional, relatively infrequently traded compared to a treasury order book, which is by the nanosecond, right?

It just strikes to me putting those two articles together, that if there was going to be an area of SEF- land that could still be digitized and threatened by an electronic change, then maybe dealer to dealer spreadovers could potentially be threatened in time.

Amir Khwaja: Interesting. Yeah. One for future discussion, yeah.

Chris Barnes: And when you look at the client flows. That I guess is where the dealers are strong and are providing a very consistent access for clients, whether it’s via streamed or RFQ. I guess this doesn’t tell us how those trades are executed, whether they’re voice, whether they’re part of a central limit order book, whether they’re streaming prices.

Amir Khwaja: No, that’s correct, yes. So we have the platform ID of the venue, but not the training protocol. And they’re allowed to use different training protocols.

Chris Barnes: Makes sense. Thanks, Amir. Ali, that was the main questions I had today.

Ali Curi: Thank you, gentlemen. And Amir, please share with us again the title of your blog post.

Amir Khwaja: Sure. It’s “SEF volumes and share in SOFR swaps in 2023.”

Ali Curi: Great. That works. Amir Khwaja, Chris Barnes, thank you both for sharing your Quick Takes.
Let’s do it again next week.

Amir Khwaja: Thanks, Ali.

Chris Barnes: Thanks, Ali. Looking forward to it.

Ali Curi: And that’s our episode for today. You can read more about these topics on the Clarus blog, and you can follow ION Markets on X and on LinkedIn.
Thank you for joining us.