Jim Greco on HFT, Crypto and Starting An Exchange
Welcome to another issue of Front Month, a newsletter covering the biggest stories in exchanges & market structure every Friday. If you have questions or feedback, please reply to this email or find me on Twitter. If you like this newsletter and want to follow the exchange industry with me, please hit the Subscribe button below & be sure to share with friends & colleagues:
Today’s writeup marks the beginning of a new series of posts I’m excited to add to the newsletter - interviews with some of market structure’s best & brightest. The number of people who’ve subscribed to Front Month has grown tremendously, surpassing 5,000 free readers as of a few days ago. Among them are experts in the fields of exchanges, high frequency trading and the market plumbing of almost every asset on the planet. I’ve spent the last two years sharing many of my opinions - it’s time I hear from some of you!
Today’s interview is with Jim Greco, a former software developer at Getco, founder of fixed income platform Direct Match and crypto prop trader. We had a great conversation about his career & topics like high frequency trading, the state of crypto, DeFi & fixed income markets. Please go give Jim a follow on Twitter & thank him for sharing his insights!
We’ll start with the basics – where’d you go to school, what were your career interests at the time & what drew you to work at a Wall Street trading firm?
I was an electrical engineer at the University of Florida & moved to Chicago shortly after graduating. I always had an interest in the stock market, particularly when everything crashed after the global financial crisis. Watching real-time updates as markets were going crazy was fascinating to me, and I knew I wanted to be somehow involved in that. I entered the industry as a software developer which was well-timed because markets were at that point going fully electronic, particularly in US equities. Chicago became one of the hotbeds for this electronic trading industry – guys coming out of the pits & developing their own firms.
Through a friend of mine I got an interview at Getco, a firm I had never heard of & I’m sure most people outside of financial circles had never heard of. After a grueling interview process - 17 and a half hours – I finally got a job offer to work on their US equities team, which at the time was a dark pool & an algo routing platform.
What was your role within Getco’s equities team? What did a normal day there look like & what were some of the challenges of that business?
Getco was one of the few high frequency trading firms that also ran a dark pool, where the pool’s liquidity was provided by Getco’s proprietary trading desk. I worked on the team that ran the ATS operation – the matching engine, interfacing with our customers, etc, while a separate team acted as the market maker.
Working there taught me just how difficult running a successful customer-facing business is. Institutional equity traders have 50-something dark pools to choose from & there are only so many levers you can pull to make sure they keep trading on your venue. Some of those levers include greater liquidity, greater price improvement, paying customers for order flow, and so on. Each customer has their own preference for one specific lever or another, so our job is designing a system that optimizes for a global market & the specific needs of our customers at the same time.
Because we were such a small team – I was one of four developers working on the dark pool at the time – you’re forced to build a jack-of-all-trades skillset which applied to the junior guys like me as well as the senior members of the firm. After working on equities for a while I volunteered to work on Getco’s expansion into fixed income, which involved building a trading platform similar to what we’d built for equities. The platform would allow institutional clients – primarily banks – to directly access our liquidity in US Treasuries.
As a developer, what were some of the differences between your work in equities & fixed income? How did the strategy & your day to day responsibilities change?
The job really was completely different. Equities regulation is much more nuanced, equities trading technology was at a much higher level, and the underlying products were much simpler than fixed income. I took some basic liquidity provision algorithms from our equities team & ended up rewriting 90% of it when applying those algorithms to US Treasuries. This speaks to one of the first points I learned very early on – there’s not as much horizontal scale to trading technology as you’d expect. Each of these asset classes are so different from each other that the amount you can carry over from one to the other is much smaller than you’d think.
Another point I learned on the fixed income side was that in order to keep customers happy, our firm had to make losing trades sometimes. This meant providing more liquidity than we were comfortable with to preserve relationships that would end up being very profitable over the long term.
At its peak Getco’s fixed income business had a huge presence in the market, accounting for ~18% of trading on platforms like BrokerTec, eSpeed & CME. This business – which may include some of the algos I wrote – still exists in some form today as Virtu Fixed Income.
In late 2012, Getco announced a major deal to merge with Knight Capital for $1.4 billion. How did the merger impact your role at the firm?
The Knight Capital deal marked a dramatic shift in Getco’s focus away from proprietary trading in fixed income and US equities. Knight was a giant customer facing business in the equities world with ~1,400 employees when they were bought. I saw the deal as a chance to move on & take on a new challenge.
I had built a good base of algo trading experience at Getco and I wanted to get a better understanding of the fixed income customers who worked with the firm. One of our clients at Getco was Jefferies, a relatively new primary dealer in US Treasuries. They brought me on to build out their electronic execution capabilities & expand their trading business. It was a good chance for me to learn how banks think about the world relative to a firm like Getco & deepen my understanding of fixed income market structure.
As you’d expect, banks think about the world in a completely different way from quant funds or algorithmic traders. Jefferies traded with customers including hedge funds, pensions, and other asset managers, and focused heavily on building strong relationships with those customers & understanding the macro environment to gain an advantage.
Can you talk more about how Jefferies used Getco’s products to do business? Why did Jefferies need to trade with Getco in the first place?
Good question. Banks like Jefferies traded with Getco as a hedging mechanism – if a pension fund wanted to buy US Treasuries for example, they would call up Jefferies to fulfill their order via an RFQ (request for quote). Jefferies would take the other side of their trade & offload its exposure by trading on a number of electronic venues, whether it be CME, BrokerTec, eSpeed, OTC venues like Getco or even other banks. Jefferies was paid to take on the risk of fulfilling customer orders & worked to manage this risk with other links in the fixed income trading chain.
After I started there Jefferies began trading electronically on Tradeweb & Bloomberg, which were the two main RFQ venues where customers would submit orders. Within weeks Jefferies had improved market share & received good feedback from customers on their liquidity experience. We still had a lot of work ahead of us to build a sustainable electronic business, but I believe my time with Jefferies had so far been a success.
It seems like you were adding value to Jefferies’ trading business & your efforts to expand their market share were going well, but after a few months you left the bank to launch a startup. What caused you to make that decision?
Working at Jefferies helped me learn how market structure was changing in the fixed income space. Jefferies was, at its core, a middle-man – it would quote customers prices by taking data from BrokerTec & eSpeed, add a small mark-up, and then immediately trade its exposure away through other platforms. To me, it felt like there was no way the world was going to work like this in five years’ time. In the electronic, post Dodd-Frank era, I thought for sure that banks would be out of the liquidity provision business in liquid markets like US Treasuries.
So with this idea in mind I left the bank in late 2014 to found a platform called Direct Match, which was basically an exchange for US Treasuries. The idea was to create something that looked just like equities – a place where HFT firms, banks and buyside institutions all traded in the same orderbook.
How did you go about building & launching the exchange? How hard was it to get a company like this off the ground?
When I first set out to build Direct Match, I knew I wanted to build this platform but had no idea how to do it. So I set out learning what I had to do & began fundraising to build a broker-dealer. I was a technology guy originally, so I knew how important Y Combinator was to the Silicon Valley ecosystem. This is at a time when companies like Robinhood were successfully raising money, so capital market startups were beginning to get some VC attention. We told a pretty good story about how antiquated the existing system was and how we could kick out the intermediaries and build a moat around the business. We were accepted into Y Combinator, I moved out to California & our team went through their full program. Afterwards we ended up raising ~$10 million across two rounds to build out the business & hire more staff.
The entire Y Combinator & startup launch process was extremely eye-opening. I went from an individual contributor at Getco & a manager of a small team at Jefferies to a CEO with much broader responsibilities. This included fundraising, PR, compliance, and more generally just selling the public on my idea for this exchange. The experience helped me understand the industry not only from the market structure side but a broader, all-encompassing business perspective as well.
Can you help us understand Direct Match’s competitive challenges? What were your main obstacles to success?
Direct Match had two primary groups of competitors. You had the established inter-dealer brokers that had all the sell-side firms signed up to trade on them, and then you had bilateral trading platforms like LiquidityEdge. LiquidityEdge & Direct Match were trying to disrupt the market in two completely different ways – LE was advocating for a direct counterparty system of trading like the FX markets, and Direct Match was working towards a true all-to-all way of trading like equities.
Another challenging part of launching a fixed income exchange is clearing. In equities, if you want to set up an ATS you don’t need that much capital because you’re not taking principal risk. In fixed income, the platform is the one standing in the middle of every trade. This adds a lot of complexity to trading because each side has to report & confirm that information separately, and if the exchange isn’t a FICC member they have to take counterparty risk until trades settle the next day. We ran into a lot of difficulty getting a clearing relationship set up – we had some partners pull out at the last minute, other partners that were too expensive, and others who weren’t able to handle the regulatory complexity we needed to navigate. It got to the point where our exchange couldn’t launch without raising a lot more money, and we decided to liquidate & close the business at that point.
Now a few years removed from Direct Match’s liquidation, would you have done anything differently? Was your prediction about the industry right?
I have no lingering regrets about the way Direct Match went – after we closed the industry ended up moving away from the all-to-all model we were pitching & chose other ways of making markets. If we had stayed in business it would have cost us & our investors a lot more time & energy only to tread water with a product the market wasn’t attracted to.
What did your career aspirations look like after winding down Direct Match?
So backing up a bit, in 2014 when I was beginning to build Direct Match, I had a fateful introduction to the Winklevoss twins through a friend, and a partner in my new firm, who was working with them to build Gemini. I talked to them about building their exchange, which was in pre-launch mode at the time, and about the crypto space more broadly. As far as Bitcoin was concerned, I thought “ehh, this thing will never work”, which was a dumb move looking back! But we stayed in touch & I helped them understand how to best work with & attract market makers to trade on their exchange.
In 2015 a spot on Gemini’s board opened up & they offered it to me. That’s when I started to really learn about crypto as a user & as a market structure practitioner. After Direct Match ended in 2017, I found myself in the middle of complete crypto euphoria. Bitcoin was approaching $20K and the entire financial world was now talking about the blockchain. I was playing around with another friend, also a partner in my new firm, on a few crypto startup ideas when 2018’s bust hit & it seemed like the entire industry was dead. Most new projects & trading volume quickly evaporated, and there was a time when I thought for sure that the industry wasn’t coming back. I did some consulting work in the meantime & became interested in crypto again towards the end of 2020.
What brought me & my business partners back this time around was the growing crypto futures trading complex that had formed at FTX, Binance and CME. What we found was that price action really started to be driven by futures which lined up well with the trading background we had cultivated. We started building some models & trading technology, and before we knew it we had a small crypto prop trading firm up and running. We still had our day jobs & the business was running at small scale until 2021, when GTS asked us to come on as their inaugural crypto trading team & launch what became Radkl.
From your perspective, how does crypto market structure compare to equities & fixed income? How does your role in crypto compare to your previous roles in traditional finance?
In some ways, crypto is a more advanced market than equities or fixed income. Crypto markets never had open outcry pits or legacy systems housed in data centers – everything sits in the cloud & the exchanges are all operated by ex-Silicon Valley types and less so the legacy Wall Street types. It’s a great model because it allows new trading firms to get started very quickly. You don’t have to negotiate with a bunch of vendors about getting data center access or colocating near the global exchange hubs. Everything’s on AWS which makes connecting to trade much easier.
So the technology side of crypto trading is great. What’s not great is the operational side, which is riddled with inefficiencies. Each crypto exchange is its own island with a brokerage, a matching engine, an OTC desk & a whole margining system under one roof. Sam Bankman Fried talks about how this makes trading much more efficient, which is true if you’re just using that one platform. When you’re a proprietary trading firm however, the value is in connecting these different exchanges together, and that’s where it becomes extremely complicated. Now instead of managing one margin account with a prime broker, you have to manage a separate margin account with each exchange where you’re doing business. Every new hub you add to this network exponentially increases the operational complexity. There are some solutions that are being built to tackle this issue, but it’s still a very immature ecosystem. You still really have to be a market plumber to be able to trade on these exchanges.
How much does retail participation matter to the crypto market?
Retail flow is very important in the crypto market. Crypto is really the only market where retail was active first before institutional interest came in. Retail still makes up a very, very heavy percentage of price takers in crypto. In US equities I think they say retail makes up ~15% of the market – in crypto it has to be well more than double that. There are still very few institutional players that participate in the crypto markets.
What are your opinions on the state of DeFi today? What do you see as a present or future real world use case for these protocols?
DeFi is a super interesting space to watch & be active in today. It almost goes back to my original vision for Direct Match, where I wanted to disintermediate banks & other central counterparties in the US Treasury market. DeFi is disintermediating all the centralized crypto exchange platforms & is allowing all types of people to participate in the marketplace without relying on an intermediary. It’s really revolutionary in that way. DEXs like Uniswap are now larger than Coinbase. dYdX is now larger than futures on FTX. If you would have asked me if that was possible a year ago I would’ve laughed.
When it comes to markets, DeFi has clearly found its use case. The ability to transact on a permissionless basis provides real value. However, I think crypto in general struggles with connecting back to the real world. I think Joe Weisenthal puts it well when he asks “is there any use case beyond just getting more coins?”, and I think that’s still a real struggle. The narrative about what even Bitcoin is used for has changed multiple times – it was originally a gold type of asset, and if price action is any indicator that’s turned out to not be the case. We’re still figuring out as an industry what real world use cases are going to be. One thing I can say that has caught on in this sense is NFTs. This is something that people really enjoy & becomes an easy on-ramp for broader crypto adoption. NFTs may become the medium through which a real world crypto use case does appear.
While we’re on the topic of crypto, I have to get your take on the Ethereum Layer 1 wars. Are you a fan of one particular blockchain today & do you think one blockchain will end up with most of the ecosystem share long term?
It’s a good question, and I don’t have a great answer for you. The industry changes so fast that it becomes hard to make long term predictions. All the Layer 1 chains have reached capacity limits in the past and they do it in different ways. Ethereum transactions get super expensive to the point where minting a cheap NFT becomes prohibitive. Solana just went through their own fiasco the other day, where every time the network is under stress you can’t do anything on it. It’s almost worse than Ethereum because with ETH you can at least pay to get a transaction done. Other Layer 1’s like Avalanche, which is an Ethereum clone, were thought to be the real solution, but when Avalanche got really popular it also got too expensive. It does feel like there will have to be more specialized chains for different processes, but whether any of these chains will meet the challenge, I don’t really know.
It was recently made public that you left Radkl to launch your own hedge fund. What drove that decision and what kinds of strategies will your fund be focused on?
I’ve always been very entrepreneurial, and so have my partners. We were excited to take a new venture on & explore the crypto markets for ourselves. Our investment manager is launching in March with a focus on mid-frequency relative value strategies. We’ve witnessed the birth of a new asset class in crypto – our team has been in finance for a long time & we’ve never seen that. We’re super excited about the knowledge we can take from our traditional finance & crypto-native experience to create something unique and special.
Thank you for reading this issue of Front Month. Word of mouth is the #1 way others find this newsletter - If you liked this week’s content, please consider sharing with friends & colleagues. Questions & feedback can be sent via email or Twitter.
Disclaimer: I am not a financial advisor. Nothing on this site or in the Front Month newsletter should be considered investment advice. Any discussion about future results or projections may not pan out as expected. Do your own research & speak to a licensed professional before making any investment decisions. As of the publishing of this newsletter, I am long ICE, CME, TW, NDAQ and VIRT. I am also long Solana.