The Markets ConversatION Podcast

Quick Takes: SOFR swap SEF volumes – May 2024

June 28, 2024 | Duration: 11 minutes

Speakers: Amir Khwaja and Chris Barnes


In this episode, Amir Khwaja will explore SOFR swaps SEF volumes for May 2024, examining key trends and insights in one of the largest products traded in the U.S. rates market.

Amir provides a detailed analysis of the trading volumes, market share, and competition among SEFs, highlighting new entries and significant trends. Chris brings a unique perspective on the numbers, offering a deeper understanding of the dynamics within the dealer-to-dealer and dealer-to-client markets.


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 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 ClarusFT blog would you like to discuss?

Amir Khwaja: Sure, Ali. Yeah, so it’s called “SOFR swaps SEF volumes, May, 2024,” every few months, I look at the volume of surface swaps, which is one of the largest products that trades in the U.S. rates market.

I look at the volume, so in this blog, looking at the volume between Jan and May 2024, looking at that, I divide the market into two bits, the D2D market where dealers transact with each other, either for their house accounts or to hedge client activity and the data-to-client market where dealers face clients. In our SDRView product we collect all that data, transform the data to what actually transacts in the market in a price forming way.

So in the dealer market, the major trade is spread over U.S. treasuries and we look at market share, both in terms of the number of trades that trade in a period, and in DV01 risk terms. And we look at market share, number of trades, volume, et cetera. The stats are pretty similar from when I last did this blog, so I kept them to lead in spreadovers, tradition in flyers, et cetera. So that share is pretty much similar. We do have a SEF this time for first time in quite a while. So RTX is a new SEF. So they’ve done some transactions. It’s always good to see more competition for the incumbent five D2D venues.

So they’re starting to trade, small trades now, but it’s meaningful. So that’s interesting. And then, we have types of trades spread over as butterflies, curves, CCP switches, and we look at volumes in each of those in trade count and

DV01 terms. And then we, we switch focus to the dealer decline market, which is really dominated by Bloomberg and TradeWeb with their different venues. So that tends to be a larger market in terms of trade count in DV01 than the dealer market. There’s far more clients, trading more trades. It’s also a more variant of what trades. So not only do you have spot starting swaps as curves or flys or outrights, but you also have a lot more forwards. Forwards, IMMs, MAC, so a wider variety of sorts trade on different dates, et cetera. And again, the market share tends to be pretty similar. It’s close to parity in some products Bloomberg dominates in some Tradeweb dominates, it just depends. But they both have, large market shares depending on the niche that we’re looking at or the type of trade we’re looking at.

So that’s a blog, so SDR product, I’ve got six charts in there. All the charts come from our SDRView product. They’re easy to produce. I just go in, log in, click on filters, pick a date range, export the chart in HTML, and display to a screenshot. So I think, straightforward to produce, and I think I’m doing it pretty regularly. Occasionally I’ve noticed, because a lot of numbers that I transcribe manually, on this time I did get two or three emails with minor corrections of things I hadn’t got quite correct. So I do wonder I should, obviously there’s quality control, but there’s a smoother way to transcribe from the shares and numbers.

‘Cause I’m exporting percentage shares and things that I calculate. It’s also good to see that people comment because it proves they are reading it. Even if I have to make small corrections, that’s useful. And it’s, this is one of my normal things. So at least quarterly, I’m looking at year to date volumes or quarterly volume or monthly volume for products. So I think that’s really the blog and I was going to ask Chris if he had any questions.

Chris Barnes: I sure do, in a bit of a geeky way, some of the numbers really jumped out at me. I think it’s important to point out these numbers cover five months of trading, right? So I looked at it and I was like, spreadover! Spreadovers are like really liquid. Let’s see how liquid they are exactly.

So blog states 13,000 trades, right? I reckon that was done over a hundred and ten trading days. If you do the maths, that’s a hundred and twenty trades a day. Splitting that across five venues, it lets you ignore market share, or let’s just say it was a level playing field. You’re expected to get 24 trades a day.

You’ve got, let’s say, 10 blokes, or girls, sat in a room facing each other from 7am to 7pm. With four phones each covering anything between five and 10 clients each, right? And they’re fighting over 24 trades a day. It’s like less than a trade an hour. And so you can really see how that dynamic can work in clients’ favors there, right?

If a client has a price and wants to put a price into that order book, you really want people fighting to win that trade. If you know that’s the only price that you’re going to hear, and that’s the only likelihood that it will actually trade for an hour, and the only way that you earn money is to actually transact to trade is it’s not salary, it’s all commission, right? You put yourselves in front of, in the shoes of those voice brokers and competing against your colleagues for a trade an hour is a pretty visceral experience, I think. And so then you extend it, right? And people are always talking about brokerage. Is it fair the amount of brokerage which is paid when it’s a commoditized product?

We’re all interested in liquidity of the market. And a lot of the reasons that this data is interesting as well, is because in many ways it makes the revenue of brokerages very transparent. So we’re talking about a total size of wallet there of $640 million of DV01 traded across 110 days. Divide that by the 13,000 trades, right?

Your average trade size is $50K of DV01, that sounds about right, ballpark. That sounds exactly right. What are you going to earn in brokerage on a $50K ticket? Let’s say that brokerage is a 20th of a basis point. You might be earning $2,000 for a single ticket, right? That sounds like a lot. And to a normal person, that does sound a lot. But when you wind it back and you’re fighting across five venues, you’ve got between 10 and 20 brokers fighting against each other for that ticket. And there’s one ticket an hour. You’ve got a dynamic where people are really motivated to earn that $2,000 dollars in that hour, because that might be their only chance for like five days or something.

And then you look at how ICAP are the market leader with a 45 percent market share, right? You imagine being a broker in one of those other shops, those numbers can be orders of magnitude lower. And so I do think when market participants are reading these type of blogs, that is the type of lens they’re lending this, right?

But then you flip it to the DTC side. And on the dealer to client SEFs, I thought it was really interesting that you just put the compression bit at the very end with no comment. It’s just like the numbers are massive, and the numbers are absolutely, insanely huge, like a hundred thousand trades, right?

And those, and okay, these are probably still charged at $25 a ticket, irrespective of size, et cetera. It’s more of a workflow solution. But you really are, there’s an argument that we should be talking about this as an all-to-all market, but you really are talking about two completely different beasts.

You cannot compare a D2D SEF with a D2C SEF.

Amir Khwaja: Yeah, great points, Chris. Yeah, I think great points. As voice broking on commodity products is a tough business, right? I think it’s no question. And we can see that in the share prices of the publicly traded voice brokers, versus the share prices of the electronic firms.

And a lot of the client activity is more automated and more electronics, certainly on the compression side, and I think I think that lends to the profit. Competition is good. Clearly it’s very tough, but then it’s interesting that we have a new SEF competing in the D2D market.

So they’ve obviously taken a view that there is profit to be had in a very competitive market. So I think, either you’re waiting for the really large trade on the day it happens. You make a lot more than $2,000. We’re talking about infrequent days with massive volume trades, that you can make up the profit from an average days or…, but I think it’s a tough business. Or you’re really looking at other products that are higher margin, high margin brokerage. Otherwise, the future has to be electronic, right? Much less people based. And I think, the investors in the share prices of the IDBs that are public, and the venues that are public can demonstrate that, right?

Chris Barnes: Makes sense. Final question, now that the SEFs are split between D2D and D2C on the SDR tickets, does that mean in theory on the D2C platforms, we could mark up whether a client is a buyer or a seller, ’cause they’ll be crossing the bid offer each time, or do we still run into the problem that there’s the CCP basis, which makes it pretty much impossible to know whether it’s a buy or whether it’s a CME cleared swap?

Amir Khwaja: That’s a good point, Chris. I think we could. We haven’t looked at it in much depth. Obviously, I think the CCP basis is noise in that one surface swaps, but I think we could account for that. And we could get a sense of how much the volume is payers versus receivers on certain tenors. And that would be interesting.

Chris Barnes: That would be valuable, I think.

Amir Khwaja: That’d be valuable. So we could say which part of the curve was being paid or received on the client venues. Yeah, I believe that is possible, we would have to spend time getting rid of all the noise to achieve that correctly.

Chris Barnes: And we have to stress it would be a subset of trades, right? Because it would be comparing an order book, let’s say on a D2D SEF in spot starting software swaps with what is just traded on a D2C SEF. I do think with this added transparent it’s increasingly possible now.

Amir Khwaja: Great. Thanks, Chris.

Chris Barnes: That was everything I had.

Amir Khwaja: Back to you, Ali.

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

Amir Khwaja: Sure, Ali. It’s called “SOFR swap SEF Volumes May 2024.” Not the most catchy title, but I repeat it on a regular basis.

Ali Curi: It gets the job done. Thank you, 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. I look forward to it. Always a pleasure.

Amir Khwaja: 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.