Jesse Greif, COO of OneChronos — Optimizing for execution quality in electronic trading markets
In today’s episode, Kailee Costello sits down with Jesse Greif, the COO of OneChronos. OneChronos pioneered the technology to run a unique kind of optimization in electronic trading markets; their approach draws from Nobel prize-winning auction techniques and solves for execution quality outcomes rather than just speed.
“We’ve created an environment to compete on the quality of liquidity and take speed out of the equation. The technology is available now to optimize for what the majority of users define as great trading execution quality … we’ve pioneered running this technology in the timescales and with the resiliency required for capital markets.”
In this episode, Kailee and Jesse discuss:
- How a traditional capital market works, and how this differs from the periodic auctions that OneChronos uses
Jesse: There’s not a ton of transparency in this space and the supply chain is somewhat convoluted to the folks that wouldn’t directly be in it and transacting in it. Effectively, the way that trading generally works now is an institutional trader has an order that they send to their bank or their broker. The bank or the broker usually divides that order, which is called a ‘parent order’, into many different child orders, and thoughtfully spaces that out over a prescribed amount of time. It delegates responsibility to something called a router — that router sends orders to many different trading venues, many different exchanges, and pools of liquidity. There are about 50 different trading venues in the US. Many listeners would likely be familiar with New York Stock Exchange, and NASDAQ, those are two of the approximately 50 that are out there.
The way that most exchanges work is that they have to make a really tough decision to say, “Okay, I have a buyer for 100 shares, I have two sellers who are both willing to sell 100 shares, how do I figure out who gets to be that seller?” … the way that this has evolved is that they’ve taken this concept of ‘first in first out’ as a fair way to allocate. So, the person that gets there first gets to be that seller in the event of a situation like that, which we call an imbalance. So how do we resolve that imbalance? We resolve it by, if multiple people are there at the same price, the person or the machine that’s there first gets to effect that sale.
Now, just in accounting, there are other methods — there’s ‘last in first out’, there’s ‘max loss’. But in the context of equity trading, the dominant format has been this ‘first in first out’ methodology.
‘First in first out’, over time, has caused some unique incentives to exist — the obvious one being an incentive to be fast, because you want to be the one that’s able to be selected in the case of that imbalanced scenario. So that has encouraged a lot of investment in being fast to instruct orders, being fast to cancel orders, being fast to have market data that you can react to very quickly, and being faster than your competitors. There are all sorts of other nuances, such as for example, being co-located within a data center so that the time of flight between your decision engine and the matching engine and where the actual software is located is very short. So literally, the time of flight through that optical fiber cable is very short.
This is an interesting incentive mechanism because it sounds pretty distant from investors and traders and users. Whoever are the end beneficiaries of these trades, whether they’re retirees or pensioners or limited partners, why did they have to be yanked into this ecosystem that really incentivizes being fast? It’s a longer conversation, but basically, the cost to be fast continues to go up.
How OneChronos is different is that we take speed out of the equation. We’ve created an environment to compete on the quality of liquidity. The technology is available now to literally optimize for what the majority of users define as great trading execution quality. We optimized for those outcomes, and take speed out of the equation. What ‘periodic auctions means is, we run auctions a few times per second. This sounds very fast; it takes about a third of a second to blink your eye. But in electronic trading, it’s certainly slow enough to aggregate liquidity and not have any reward mechanism for who was there first. So that’s the material difference between periodic auctions, rather than what we’ve kind of referred to here as a price-time-priority double-auction, which effectively says, “whenever there’s a contra order that can potentially match with you, that match will be created. To the extent that there’s some imbalance, the person that got there first is the one that gets to effect that trade.”
- How the Nobel Prize-winning ‘smart market’ concept is applied to enable traders to express indifference and substitutes in the context of equities trading
Jesse: OneChronos certainly didn’t invent math and didn’t invent decision science — the concept of optimizing for specific outcomes is something that’s really ingrained in many slower-stakes industries. For example, the electricity that powers the lights in this room, and the water that we’re drinking, go through optimization allocation methodologies that say things like “we’re going to provide a certain amount of power”, or people will bid for certain amounts of power across the network subject to certain constraints. The optimization goal, what’s called an objective function, is either maximizing something or minimizing something. So it could be maximizing, the revenue or social welfare across this ecosystem, or it could be minimizing costs or something to that effect. The challenge is that running optimizations can be computationally intensive, particularly when there are many widgets or many goods that are being transacted, and when there are many flexible constraints that can exist in the context of optimization. So in our case, we’ve brought this technology to operate in much, much faster timescales. The beauty of this is that not only can optimizations, you know, find great solutions for a predefined objective, but they can also create more efficient allocations of things by offering folks more ability to express their constraints.
To go through a trucking example: I’m a trucking company in New York, and I’m bidding to, carry cargo from New York to Los Angeles, I’m willing to bid $1.40 a mile to drive that freight from one coast to the other. But if I could get an allocation where I’m driving something back from Los Angeles to New York, rather than driving back with an empty truck, maybe I’d be willing to bid more aggressively. So I’m willing to do $1.40 one-way, but I’m willing to do $1.21 if I can do both ways. Imagine if I could submit two bids into this auction of “one way for $1.40” and “round trip for $1.21”. Those are two independent things, and I’m indifferent across those outcomes.
In trading, it’s a powerful tool because trading doesn’t allow for this type of behavior. It does in the sense of separate regimes in that I can attempt to do something, and if I can’t accomplish that action, I can attempt to do that afterward. But the market has changed, and the variables have changed, and this is kind of a nuanced thing but I’ve “leaked information” (I’ve told the market certain information about what I attempted to do the first time). So my second trial is really different and now kind of biased by my first attempt. Having the opportunity to express indifference and substitutes in the context of equities trading is a particularly powerful tool. These are things that traders and their executing algorithms already desired to do, but there aren’t tools at the exchange layer that allows for that type of flexibility or rather fidelity to be maintained throughout the stack.
An example in equities trading could be something like, “If I pay $10 for this I’m willing to buy 100 shares .. but if I can get it done for $9.50 I’m willing to buy 1000 shares. Or in the case of substitutability, “I’m happy to buy any of these momentum stocks, it doesn’t matter which one, I’m happy to buy any of them up to $10 million each.”
These are optional tools that are kind of the cream on top of a periodic auction, which, as we talked about, takes time, takes speed out of the equation, and solves for competing on quality. But, it allows folks to express specific constraints within the exchange matching engine, which is a novel piece of technology that hasn’t existed before. So in those other asset classes, whether it’s ad tech, logistics, or airplane flight allocations, these are things where this auction format is prevalently used. The technology is available now and we’ve pioneered it to run it in the timescales and the resiliency required for capital markets.
- The SEC’s proposal to increase competition for retail order execution
Jesse: I’d start by saying that our product was designed initially to be an institutional trading product. The beauty of it is if you survey a thousand of the largest institutional investors and traders and ask them what great execution quality means to them, the technology is available to encode that in an auction’s optimization.
Now, that said, the SEC came out with a piece of regulation not long ago that proposed a new market structure that effectively shakes up the interactions that retail trading activity might have with the rest of the market. Currently, that supply chain is relatively captive and happens in a very specific way that’s particularly segmented from how institutional activity is traded. The SEC’s proposal effectively creates a mechanism where that trading activity is much more open and can take place between all different types of market participants.
The unique thing that they proposed is a mechanism that functionally looks very much like OneChronos. So much so that after this rule was proposed, we got probably 50 phone calls the next day asking us about the kind of convenient and coincidental similarity between what’s been proposed and what we have. Without commenting on the rule itself, we’re really flattered by this regulatory body and the 160+ economists that work there saying that they think that periodic auctions that run in the timescales of 8–10 per second and that optimize for users to get price improved is the fair way to trade.
- The challenge of capturing both the supply and demand side of the market to launch a new trading venue
Jesse: Anyone who’s launching a new venue has this chicken and the egg problem. The short answer is you need a lot of order flow, and you need a lot of diversity of order flow. The way to do that is to spend a lot of time with the end user beneficiaries of these mechanisms. Our customers, technically, are the banks and brokers. But myself and our sales team spent most of our time with the end users, such as the pension funds, the sovereign wealth funds, the institutions, the hedge funds, the asset managers who are doing the trading. Those are the folks that are saying “Okay, either a human or a machine is saying we need to transition from Portfolio A to Portfolio B, and in order to do that we need to sell some stocks and buy some other stocks”. So there’s this whole workflow that ensues of them sending orders to their broker, and going back to what we discussed before, in terms of the broker thoughtfully taking this parent order and placing it in different parts of the market.
So we talk through our mission with these end users and say, “Look, this is what we’re solving for … if this is what you all say is great trading execution quality, the technology is here now to literally encode for that.” They ask us, “How do you do that so quickly, like 10 times per second?” Even within that, it’s only kind of a portion, about half, that we’re actually allocating to this optimization process. We have some kind of interesting techniques that we use, called probabilistic search techniques, coupled with offline machine learning or reinforcement learning elements that coach our optimization algorithms to find really fantastic starting points for future auctions.
We talked them through what our mission is, which is we’ve taken speed out of the equation, and we’ve fostered an environment to compete on the quality of liquidity. Most folks really subscribe to this message. So because of that, they speak with their brokers. And it’s not like a hard lobby, like “send all my flow to OneChronos”, it’s really like, “Look, if you haven’t plugged OneChronos into your trading and routing infrastructure yet, please consider it because we think they have potential to enhance our trading execution quality.”
The neat thing about it is, the electronic trading toolset that the banks have are systematic tools that don’t have human bias embedded in them. If a bank goes and plugs us in, and it turns out that OneChronos is really terrible, or people aren’t getting trades there, then we won’t get looks in the future. So there’s a kind of convenient element of “plug us in, and the router or the tools that exist in the bank will sort out if it’s actually doing what Jesse and team are saying it will do.”
That’s been our approach of spending a lot of time with tens and tens and tens of accounts, whether that’s here, or Boston, Chicago, San Francisco, or internationally. US equities is one of the largest and most liquid markets in the world, with $600 billion that trade every day, so even being a small portion of the market, or even creating a moderate amount of savings can add up to a good amount of dollars for end users.
- OneChronos’ partnership with the NSYE
Jesse: They’re certainly sponsors of innovation, and they’ve created a unique opportunity for members of their exchange to route orders directly to OneChronos. So for their customers who wish to, with very little friction, they can effectively apply what’s called a tag on the orders that they send, and that will shoot straight through to our venue, OneChronos.
The special thing about it is that it really lowers the barrier to entry for accessing our venue. This is an important part of your last question “How do you build liquidity in a venue?”. There are many folks who say “We’re interested, but we’re going to wait and see how this goes. We’re going to wait to see when you’re a bigger part of the market.” If everyone says that, then you have nothing. So a big part of this is how do you lower the barrier to entry, such that it’s very easy to onboard.
We’ve been very methodical, and this was part of our COVID story too — let’s inspect every element of the onboarding process, and where can we make it smoother and more efficient. We can certainly do that for folks connecting to us. But we can’t help the fact that when people connect to us directly, they have to sign a contract, they have to set up networking connectivity, and there’s testing and things that need to be done. So this route through the New York Stock Exchange is very helpful, insofar as it’s literally plug-and-play. It creates a very easy way to access us, and we’re pretty indifferent to whether that end user ultimately connects to us directly, or they continue to connect to us through NYSE.
Check out the Episode on the platform of your choice here: Spotify | Soundcloud | Apple Podcasts
About OneChronos
OneChronos is a technology company at the intersection of capital markets, machine learning and mechanism design, providing innovative execution venues to those in the electronic trading world.
The company was founded by executives from Goldman Sachs, Accenture, and top quantitative asset managers with a common vision to leverage domain expertise and emergent technologies to make electronic trading simpler, more transparent and more efficient for institutional investors and traders. OneChronos is a Y Combinator alum backed by top venture investors in fintech, digital marketplaces and A.I.
About Jesse Greif
Jesse Greif is the COO of OneChronos. Prior to joining OneChronos, Jesse spent 13 years at Goldman Sachs, where he most recently was a Vice President. He earned an MBA from The Wharton School and his Bachelor’s degree from Northeastern University.
About the Author
Kailee Costello is an MBA Candidate at The Wharton School, where she is part of the Wharton FinTech Podcast team. She’s most passionate about how FinTech is breaking down barriers to make financial products and services more accessible — particularly in the personal finance space. Don’t hesitate to reach out with questions, comments, feedback, and opportunities at kaileec@wharton.upenn.edu.
As always, for more FinTech insights and opportunities to collaborate, please find us below:
Wharton FinTech: Medium Blog | Twitter | Our Website | LinkedIn
Suggest a Podcast Guest: https://airtable.com/shrdbokQPxAJzgVh7
Hire Wharton FinTech MBAs: https://www.whartonfintech.org/recruiting
