US Treasury Clearing is Broken
If Matt Levine’s shtick is that people are worried about bond market liquidity, mine is that people should be worried about US Treasury clearing. The US Treasury clearing landscape is broken. The bifurcated system reduces liquidity, allocates clearing costs inefficiently, and increases systematic risk. Together, these factors stifle innovation and entrench incumbents. In today’s newsletter, I unpack how we got here, the daunting costs in moving toward a centrally cleared market, and the challenges startup US Treasury venues face.
Bi-Lateral vs. Central Clearing
In bi-lateral markets, all aspects of a trade including legal, credit, market risk, and operational risk are coordinated directly between the two trading counterparties. If I sell $100M ten-year Treasury notes to you, I am depending on you to deliver the cash to me tomorrow when the trade settles. If you’re unable to come up with the cash for some reason, then I need to go back into the market to sell (or lend) the Treasury notes to someone else because I’ve almost certainly already committed the cash to another trade (it’s of no use sitting idle in a bank account!). A trading venue that facilitates bi-lateral trading (sitting in the middle to enable anonymity) might have thousands of offsetting commitments split among hundreds of counterparties each day.
In markets with central clearing, a well-capitalized central clearing counterparty steps in the middle of a trade, guaranteeing the terms of the trade even if one party defaults. Central clearing is great because you don’t have to worry about the credit of the counterparty or trading venue. Lehman Brothers going under in 2008? It didn’t matter to the stock market! The DTCC’s National Securities Clearing Corporation was the counterparty to all trades and guaranteed that they would be settled properly. Trading, while volatile, functioned normally. Meanwhile, liquidity dried up in markets without central clearing, like many over-the-counter derivatives, as banks looked to reduce their counterparty exposures. In large, systematically important markets, regulators have stepped in to mandate central clearing, as they did most recently with interest rate swaps.
The Bifurcated US Treasury Market
The US Treasury dealer-to-dealer market is unique in that has a mix of participants that centrally clear and bi-laterally clear. Dealers are members of the DTCC’s Fixed Income Clearing Corporation (“FICC”), and when they trade with one another, FICC guarantees the settlement of the trade. The system worked well during the financial crisis when FICC closed out over $190B in Lehman’s positions without impacting customers. As trading migrated from phones to electronic trading venues (e.g., BrokerTec and eSpeed), the venues themselves became members of FICC so that they could facilitate anonymous trading between members with relatively limited capital.
The rise of high-frequency trading firms (“HFT”) presented an opportunity and challenge to BrokerTec and eSpeed. HFT firms were clearly the future revenue drivers with their massive volumes, but there was no way FICC, which is owned by the banks, would allow them to become members (why let in your competition?). While FICC is technically required to be open to everyone, it has set its membership requirements so high that only large dealers can be members. Onerous membership requirements (e.g., $10M+ on deposit at FICC) and trading fees mean that membership is uneconomical, if not impossible, for most proprietary trading firms. BrokerTec and eSpeed, knowing FICC would never change its rules to be accommodative of HFT members, assumed the risks themselves and on-boarded them as bi-lateral trading partners.
Margin Call
The actual counterparty risks to venues by HFT firms are vastly overstated. However, supporting FICC members and non-FICC members subjects the venue to significant intra-day margin calls by FICC. If two FICC members trade with one another (e.g., JPMorgan vs. Goldman), then FICC knows both counterparties and sees the venue as flat. If two non-FICC members trade with one another (e.g., Virtu vs. KCG), then the platform has only the credit risk against the two counterparties. The trouble starts when a FICC member trades with a non-FICC member (Goldman vs. Virtu). The venue has no market risk, but FICC doesn’t see it that way. FICC only knows about one side of the trade: the side with the FICC member (Goldman). It pretends that the venue is taking a proprietary position that must be margined for.
The math of a margin call by FICC is rough due to the large notional size of Treasury trades. FICC doesn’t make the exact formula available in its 271-page(!) rulebook, but it’s approximately 1.5% of the net notional amount outstanding in ten-year equivalents. If FICC members buy only $1B of ten-year notes from non-FICC members then the trading venue needs to put up $15M in a margin call! That’s an absurd amount of money for a venue that’s just matching up counterparties. No startup has the revenues to justify that kind of capital on hand. Worse, FICC can, and does, change the rules on a whim. If it suddenly considers it to be a volatile trading period, it can double the amount of margin required. BrokerTec has seen margin calls of more than $100M as a result.
The Death of New Multi-Lateral Trading Venues
Multi-lateral venues have no choice but to support both FICC and non-FICC members because of how the market structure has shifted. Dealers aren’t liquidity providers in on the-run Treasuries, so including only FICC members would mean the platform would have no liquidity. You also can’t just build a market where there are only non-FICC members. HFT firms need liquidity consumers (i.e., dealers in US Treasuries) to cross the bid-offer spread. Even Tradeweb, owned by the dealers, knew this when they invited HFT firms onto their platform when it launched Dealerweb a few years ago.
New venues either need to be attached to a firm with a larger balance sheet or narrow the focus of their business model. Direct Match had a bank partner to do all the credit intermediation, but it backed out the same week Direct Match was to start trading. Open Door Trading’s “all-to-all” model requires that participants be FICC members and thus focuses on the off-the-run marketwhere dealers are still liquidity providers. LiquidityEdge is a bi-lateral platform, and any credit intermediation done by Cantor is for FICC trades only. According to people familiar with the matter, OpenBondX is still waiting on newcomer Wedbush to be ready before it can even start trading. New eSpeed, which we’ve discussed a few times before in Trading Places (to the point where you might suspect we have a secret crush), will leverage sister-firm Cantor to provide credit intermediation.
US Equities is the benchmark for market structure innovation with 13 exchanges and 50+ dark pools. Firms like IEX (regardless of my view on them) can start trading with $10M of capital and can seamlessly onboard any broker-dealer that already trades stocks. The lack of central clearing in US Treasuries has resulted in only three multi-lateral trading venues (with BrokerTec having 80%+ market share). New venues have focused on bi-lateral trading links, which operate largely in the dark and without accountability. Without central clearing to tie the market together and enable the portability of order flow, the future of US Treasury trading looks a lot more like F/X where a few large firms dominate because they can be selective about where they provide liquidity.
There are no easy solutions, and my view from the Direct Match RFI response remains largely unchanged. The economics of central clearing, as it is currently conducted, are cost prohibitive for HFT firms. This is because the fee structure of FICC presumes a business model that the firms do not employ, and a level of settlement risk they do not undertake. A clearing mandate under the current regime would massively increase the costs to HFT firms relative to the benefits conferred. This would force them to reduce their liquidity, which would in turn be felt immediately by investors in wider bid-offer spreads. Nonetheless, the official sector should explore ways in which the costs of central clearing can be distributed more evenly and reduced by more effectively pricing the settlement risk inherent in the marketplace, and promote greater access across the market structure.
