If you are a participant in the USD credit markets, then here is a question: what percentage of trades are electronic? 20%, 30% or 50%? I have always believed that the number was in the range of 20% to 30%. I thought an interesting data visual would be a treemap of the market share for each electronic platform in the market. This little data visual project took much longer than I anticipated and completely changed my view of how platforms report market share.
In order to calculate market share, I was going to use the reported volumes of each electronic platform and the total volume figures for TRACE in high grade and high yield.
That was until I applied the trade size ranges in the Trace capped volume report.
The Trace capped volume report for January 2020 led me down a long path of questions about the definition of "the market" for USD credit and the true market share for electronic trading platforms. The following is an attempt to break down the USD Credit total market into two distinct addressable markets: Electronic and Voice. Then apply reported electronic trading volumes in January to evaluate market share for the Electronic Market. Hopefully, these visuals will lead to a conversation about greater transparency with electronic trading volumes.
Please don't ask me why I did this.....what I thought would be a simple post with visuals turned into an agonizing process of building visuals from my own calculated data.
TAM
I first heard of TAM when I was reading Super Pumped by Mike Issac. Bill Gurley put together a post about the size of Ubers TAM that counters an article by NYU professor Aswath Damodaran (Both are good reads). The book and the posts made me think about the market share and revenue opportunities for electronic trading platforms in USD Credit.
I thought an interesting way to look at the total market for USD High Grade and High Yield electronic trading venues would be combining the trade size bands from the capped trading report and the total TRACE monthly volumes. For January, the total market for USD High Grade and High Yield was approximately $729 to $735 billion. High Grade was approximately $505 billion & High Yield was approximately $224 billion. Using the trade size bands in the capped trading report, I built money bucket visuals:
USD HIGH GRADE
USD HIGH YIELD
These are my estimates and my calculations were pretty rudimentary. The lack of standards for volume reporting makes this a really fun exercise! For example - MarketAxess posted TRACE a volume estimate for January of $507 Billion; Tradeweb posted for January of $520 Billion. The monthly report on the FINRA volume report doesn't have a breakdown for High Grade or High Yield just CORP and had a volume of 826 Billion and include 144a's. FINRA's regulatory notice 19-12 has a good explanation of current dissemination protocols. After visualizing the trade size bands as money buckets, I noticed that four individual markets comprised of the total market. However, when I thought about trade size (estimating that the average trade size for fully electronic execution was between 500k and 1million) there were really only two markets: Electronic and Voice. The Electronic Market defined as trades between 0 - 5 million and the Voice Market defined as all trades above 5 million. So I built the interactive treemap to break down the two markets.
If you look at the Electronic Market for High Grade, you will see that there is still quite a bit of room for opportunity for any of the electronic trading platforms. You can also see that MarketAxess is clearly the leader.
Let's look at MarketAxess's market share using the Electronic Market instead of the total market.
MarketAxess posted an eye-popping 100 billion in volume in High Grade for the month of January and it would be reasonable to conclude that the majority of trades were under 5 million in size. For this exercise, let's assume that all 100 billion were in the 0 - 5 million trade size range.
MarketAxess stated in its Q4 2019 earnings presentation a 19% market share for electronic trading in High Grade. However, if MarketAxess restated its market share using the Electronic market then its market share would be in the area of 30%. That is an enormous swing in market share!
If you add the volumes from Tradeweb and estimates for ICE, Bloomberg, Trumid and other platforms, then about 50% or more of the 0 - 5 million size range in High Grade is electronic.
If electronic platforms used the Electronic Market instead of the total market, then market participants would have a better picture of market share. An accurate market share metric can be really handy when a firm needs to evaluate the costs associated with each platform. Look at what Citibank did with their FX platforms.