Markets can be classified into two basic stages in terms of maturity. The first stage is manual. These markets are characterized by the need for human capital for setting prices, transacting and settling. This can look like anything from buying a house to buying tomatoes at a farmers market. Both trades require conversations for price discovery and execution, and a labor-intensive process for settlement (balancing tomatoes while scrounging for cash).
The second stage is fully digital. Examples include buying crypto on Coinbase, a stock on a retail trading app or tomatoes on Instacart. Pricing is transparent, execution is instant and settlement happens seamlessly.
Over the last decade, we’ve seen many manual markets go fully digital – Uber for taxis, Airbnb for rentals, LinkedIn for networking, Tinder for romance, etc. In capital markets, the thinning of traders from the floor of the NYSE is a quintessential illustration of this dynamic. Much of the human hustle and bustle that once defined the industry has been replaced by the machines.
The Inflecting Market
Today, the bond market is at a fascinating inflection point. Electronic volume is on the verge of crossing 50%, finally eclipsing traditional voice trading. Algorithms have proliferated to quote odd-lot tickets, eliminating human traders from the workflow. Buy-side firms are in turn developing their own automation to trade through execution platforms. The classic exchange of a trader chatting with a dealer, asking for a quote and saying thanks is rapidly being replaced by an automated protocol. In a world where tools like ChatGPT can have sophisticated philosophical conversations, this shouldn’t come as a surprise.
What may be surprising, however, is that every time a bond trade is executed, there are still substantial costs borne by both parties to process and settle the trade. How much? We estimate a single bond trade may cost as high as $10 in aggregate for two brokers over the post-trade lifecycle. Not too long ago, each bond was an actual certificate (with coupons to cut out!) that had to be physically transferred between vaults, but today they are simply updates in a database. Why is it so expensive to update a few digital ledgers to transfer ownership?
After the Trade
Post-trade fees are driven by three primary sources: regulatory reporting requirements, clearing costs and settlement costs. Whereas in equities these costs are negligibly small, enabling retail to participate en masse, in fixed income these costs are about two orders of magnitude larger. At OpenYield, in pursuit of reducing costs in every trade, we’ve been examining and questioning their sources. The most compelling explanation we’ve found is one I loathe telling my kids: “That’s just the way it is.”
The reporting, clearing and settlement fee schedules in fixed income are designed around assumptions that trades are infrequent and large. Once upon a time, equity trades were similar, but at some point the markets became overwhelmingly electronic and per-trade costs transitioned to a mil (%) calculation, making every incremental share negligibly cheap to trade. Post-trade fees for fixed income, on the other hand, generally remain flat – meaning that every trade, whether it’s a single bond or 1,000 bonds, carries the same fee. This model was preferential to participants in a world where every transaction was sizable and negotiated between institutions and frenetic traders yelling into phones. But that world is rapidly ceding ground to the new wave of algos quoting small-sized bonds automatically.
Yet despite advances in the market to facilitate electronic trading, the post-trade fee schedules haven’t been reformed. Fixed fees on small tickets are punitive. Clearing brokers regularly charge their introducing counterparts $5 dollars a ticket. A combined $10 for a buyer and a seller is an entire point (1% in bond talk) of value eroded from a $1,000 bond trade. This fact is even more bewildering considering corporate bonds and munis now clear and settle using the same pipes as equities (clearing via NSCC and settlement via DTCC). Treasuries flow over entirely different pipes at the Fed and have their own high ticket fees.
Peeling Back the Onion
Some parts of the fees can be attributed to “money out the door” to regulators and utilities. Reporting fees are particularly onerous for micro-lot corporates – the minimum TRACE fee1 is $0.475 for any <200k face ticket, which must be reported by the buyer, seller and marketplace. That means if you buy a single bond at your broker, this fee might be paid 4x over to the SEC by the daisy chain of reportable transactions: dealer, ATS, your broker and again for your broker allocating to your account with a commission. Another high exogenous fee is charged by the NSCC2 for clearing corp/muni bond trades: $0.85 per side. These exogenous fees can only be reduced by industry-wide pressures.
The bulk of the fees, however, stem from the convention of allocating the significant costs of running a clearing operation down to single tickets. While single tickets are mostly automatically processed, they might require manual operational processes to confirm, resolve mismatches and fails – especially when there’s still so much volume trading over manual execution protocols, which are prone to errors. Significant upfront investments in technology, personnel and capital are required for clearing firms to operate. Clearing fee schedules are generally blunt instruments that don’t accurately represent unit costs, which consequently benefit big trades and hurt small trades.
A New Way Forward
This market, like all markets, will change – but only in response to market forces. We believe these forces are coming in the form of heightened retail demand for bonds, applying pressure to clearing broker fee schedules, clearinghouse fees and regulatory reporting fees. New tech-enabled entrants to the clearing space, such as Clear Street, have an opportunity to reengineer operational processes versus incumbents who have less flexibility due to legacy overhead and fees imposed by outsourced core technology vendors. We think it’s unlikely traditional blockchains are our savior, as the key issue is not around enabling decentralization, but rather improving centralized processes. The growing narrative around digital assets and tokenization will help shine a light on the gaps in fixed income’s critical market pipes.
Most bond market participants are oblivious to this unglamorous underbelly of the market. Their obsession is with size, winning the largest customers with the largest trades. At OpenYield, we take just the opposite approach, focusing on the smallest tickets possible. We firmly believe that if you can change the unit economics for these tickets, then you can change the entire structure. Even the word “liquidity” screams the metaphor: for water to flow, it can’t be made of blocks. Execution algos can proliferate, but not if they result in death by a thousand cuts in the form of post-trade fees.
With the growing share of machines quoting markets, the pre-trade space has crossed the chasm. Now, the post-trade space must follow suit. While the current framework might have made sense in a more analog world, the fees are now far outdated and disproportionately impact the lowest end of retail investors faced with onerous commissions. Fortunately, fixed income is receiving renewed attention as brokers enter the market, discover and challenge the status quo. Investors will vote with their feet and fuel the continued reduction of friction in capital markets. A cheap and transparent bond market is what retail investors deserve. The gears are already in motion. Once broadly reformed, these speedbumps that exist in the form of post-trade costs will dissipate and the market can make its full transition to the digital age.