The Gateway to Algorithmic and Automated Trading

Financial wanderlust

Published in Automated Trader Magazine Issue 31 Q4 2013

Andrew Auerbach began trading at university and soon realised he'd found his calling. Starting with simple long-only blue chip positions, his strategies became progressively more sophisticated. Now head of Auerbach Group, this young American fund manager is embarking on his most adventurous journey yet as he starts an algorithmic trading fund in Brazil.

“For me the intellectual pursuit was a really key motivating factor. There are obviously a lot of different strategies that work in the market, but for me what’s really interesting is to see new areas that haven’t been explored... ”

"For me the intellectual pursuit was a really key motivating factor. There are obviously a lot of different strategies that work in the market, but for me what's really interesting is to see new areas that haven't been explored... "

Adam: Let's begin with how you got your start in trading.

Andrew: I began trading when I was studying business. I started out with my own money, which I had earned from working summers and during high school. I did very well with that, expanded to family and friends of family, and then ultimately in the second semester of my freshman year I had consistently done well enough that I decided to withdraw from university and start a firm. That was before Harvard (when I went back to university).

The rule of thumb was buy blue chips and read the Wall Street Journal; a very fundamental style. My strategy was to be constantly looking for new opportunities, to not just go with what books would say. So ultimately I came from blue-chips long-only, to tech stocks on NASDAQ, and then to short-selling and trading on margin, and then to options in order to get leverage. Then on to option strategies where the strategy was using options as a vehicle for more precisely defining the risk. From there I got into automated trading as I started to see interesting things in the derivatives markets.

Adam: Do you feel there was an element of luck involved?

Andrew: There was absolutely an element of luck. I don't deny that. Very easily that element of luck could have been against me, and I don't know what I would be doing now - but something very different.

Adam: In your first trades, did you have price targets or any kind of system?

Andrew: I basically said, okay this exceeded my expectations, so I'm just going to cash it in and try something new, keep going and see what I do right, what I do wrong, how I could improve. For me the intellectual pursuit was a really key motivating factor. There are obviously a lot of different strategies that work in the market, but for me what's really interesting is to see new areas that haven't been explored, areas that have more challenges and opportunities.

Adam: What was the turning point where you decided this is what I want to do now?

Andrew: As my success gradually continued in trading, my motivation for school gradually declined. The original plan was that if I lost 30% of the initial capital that I began trading with I would call it quits, continue focusing on university and if I succeeded then I would cross that bridge when I got there.

Adam: Once you decided this was what you wanted to do, how did you go about getting the knowledge and information you needed for the more difficult aspects?

Andrew: It's mostly self-taught. There were research papers, books, and trial and error. Trial and error was definitely the most useful for me. I found there's a lot of good information out there. But in general there's also a lot of misleading and false information on how to make money in the stock market. So I found that trial and error was really the only one I could depend on to reliably learn.

Adam: What about the quantitative side and the higher math? Did you get into that area, and do you consider yourself a quant?

Andrew: I absolutely consider myself a quant. In terms of the higher math, some I knew from university and some was self-taught from reading books. And there were some problems where I realised I needed help. I have a network of friends and advisers who I can turn to for advice.

Adam: It's interesting because in this business there are several different areas where you need expertise. The macroeconomics and market knowledge is one. Then there's the quantitative analysis side, the mathematics. And then there's the technological aspect, being able to programme or understand the technology. Typically in a lot of the funds you'll see specialists in each of them. Ideally, you want to have all of them.

Andrew: The reason I think that I've been able to combine so many separate skills actually comes from learning from the example of my father, who himself implemented a career strategy of combining multiple different fields. He combined science with law, and then of course the demand for a person with a high degree of specialisation in different areas is much greater. Career-wise it's much more successful.

Adam: You've mentioned that you have some granted patents for financial technology. I'm curious, is that one of the reasons you've gone ahead and patented some of your work?

Andrew: Yes, I am the sole inventor of two granted US patents, a third pending US patent, and several granted and pending patents around the world. My father is in fact a patent attorney, and this is one of the reasons I have gone ahead and patented some of my work.

Adam: Okay, switching gears here. You're in the process of setting up a trading company in Brazil. Tell us where you are with that.

Andrew: We haven't started trading yet. We've now just got all the regulatory approvals and we're working on the final implementation of the infrastructure, which is a unique challenge given the developing world and especially in Brazil. If you want to do a cross-connect in the US or in London, it's ready in a couple of days. Or if you want a long haul connection, there's no problem. In Brazil it's more complicated.

Things move slower in Brazil. So from the time that you make the decision which path you want to take, which company you want to use, how much bandwidth you need, what type of requirements are needed, just to get a quote will often take a substantial amount of time. Even to get a one-week return on a quote request would be pretty impressive. So you're looking at a much more drawn-out process. Things work, they just work in the way of Brazil.

Andrew Auerbach

Andrew Auerbach at Bovespa during a recent visit

"Brazil is at a good crossroads between being developing and not overdeveloped. So the market has a lot of opportunity. It's not overcrowded like some markets."

Adam: What's your timeframe for launching the fund?

Andrew: We're targeting the fourth quarter of this year. We're also working on the design of the actual fund product vehicle, in terms of things like tax optimisation and the legal structure of how clients will participate in the funds.

Adam: Brazil had a discriminatory policy in the past, as far as foreign and local market players. But they've recently cut the transaction tax they had in place to zero.

Andrew: It's really good news, especially for getting ready to launch a fund in Brazil. Brazil may use that as a type of fiscal policy. It's not necessarily a permanent change, so, you might want to view that as a somewhat temporary measure from the government to deal with the economy and to enhance the supply of foreign capital coming to Brazil.

Adam: Either way, in one sense it seems an unusual time to be launching in Brazil given that the economy is not doing as well and some of the international enthusiasm for the BRICs has waned.

Andrew: Brazil is at a good crossroads between being developing and not overdeveloped. So the market has a lot of opportunity. It's not overcrowded like some markets. Western European and US markets can sometimes be overcrowded. But at the same time it's not so underdeveloped that it's impossible to conduct business.

In Brazil especially the market is just starting to get into medium and high performance infrastructure. Although Brazil recently adopted the new PUMA system and they're upgrading their matching engines and the exchange infrastructure, there's still a lot of work and room for improvement on broader technological infrastructure.

Adam: Do you feel, as someone who is coming in from the outside, that you have any edge given your experience trading in developed markets?

Andrew: There's a bifurcated market here in terms of domestic players and foreign players. So far, a lot of the domestic Brazilian players are using automation, whereas in the past that wasn't used so much in Brazil. But at the same time, the level of complexity is more basic; so you'll have some funds that are using essentially Excel spreadsheets with macros. It's a very different type of competitor.

And then there are international players. They've got a lot of the Chicago guys who are in Brazil and are starting to get accustomed to things. A lot of these guys have disadvantages though, that they're US firms doing business in Brazil, rather than local Brazilian entities.

We're trying to come in and get the best of both, technology and experience from existing high-frequency trading operations and more mature markets combined with a local presence and local business structure.

What we're looking for primarily is liquidity. That's the most important thing. We can help to add liquidity to the market, but there needs to be something, a seed to work with. That's probably the most important factor. The next one of course is that the markets have to be fully electronic and they have to have sufficient throughput and performance on the exchange side in order to handle these operations. You know, before the PUMA upgrade in Brazil, there were some studies and information from brokers that latency could vary anywhere from a couple of seconds at the best to more than 60 seconds in peak volume. When things are going wrong, imagine submitting orders and having 60 seconds without a confirmation and you don't know if you have exposure

Adam: You decided to not go in as a foreign player but to form a local company. Even though some of the disadvantages for foreign firms have been addressed, presumably there are still benefits to being locally based. Can you talk about that?

Andrew: The main advantage is an advantage of psychological presence in the market. For raising money domestically in Brazil, of course everybody's patriotic to where they're from, and people want to invest in a local company rather than a foreign company. My firm is not a foreign company masquerading as a local company. There are Brazilian partners who are Brazilian citizens, living in Brazil full time, speaking for the company, working full time, having equity in the company. We are also interested in finding the best Brazilians to join our team.

Adam: Have there been many others who have gone down the same route?

Andrew: There have been some that have gone down to Brazil for trading, but mostly for their own books. A lot of guys from Chicago. That's really where it is in terms of foreign people coming to Brazil and trading.

Adam: Is it mostly derivatives trading?

Andrew: It's a lot of derivatives trading, a lot of market making, so it's the familiar names that you would expect to be in this space around the world, in the US, in London.

Adam: The GETCOs?

Andrew: The GETCOs, the Towers… So they're trading for their own books and not really raising capital in Brazil as far as I know. Where I'm coming in is a different strategy of localisation and offering a high quality product and service to the local market, as well as an opportunity for non-Brazilian people to have a vehicle to enter the Brazilian market.

Adam: Do you have clients lined up?

Andrew: Yes, we have commitments from clients ready to go in. Some very important clients within the Brazilian market, as well as strong interest from outside of Brazil.

Adam: What about the size of the fund?

Andrew: In the beginning we're expecting an aggressively ramped launch, because of the nature of Brazil. It's not a highly developed market for automated trading. There's a high degree of scepticism within Brazil about automated trading because this is considered relatively new technology. So before people really come on strong with a lot of capital we need to convince them just to dip their toe in the water and make a first investment in this area. But if they're willing to dip their toe in the water and see that it's good, then they will be enthusiastic to come in for more.

Adam: Of the investment you're hoping to raise, what sort of split do you think it's going to be in terms of local versus foreign?

Andrew: We have a target that at least half of the money should be domestically from Brazil, in order to keep along with the theme of a Brazilian firm, that most of the ownership is Brazilian, and most of the assets under management should be Brazilian.

Adam: What do you plan to trade?

Andrew: Eventually we're expecting to expand into everything. In the beginning, we're going to be heavy on electronically traded equities and also on derivatives products, especially on currency and interest rates, which is the most liquid part of the derivatives market in Brazil.

Adam: How would you characterise the kind of trading you're going to be doing?

Andrew: It's going to be high frequency, low latency. We are going to be doing some stat-arb as well, but the core focus is on the crème de la crème, super latency sensitive trades. Because the market in Brazil is not as liquid and has lower volume than other major markets in the world, we're expecting that the hold time will be longer than in the US or elsewhere, but strategy and execution latency is still critical.

Adam: In that sense, you will be competing against some of the best.

Andrew: Absolutely.

Adam: And from your experience trading in the US, do you feel optimistic about that?

Andrew: I definitely feel we're up to the task, with our technology, knowledge and systems. I have a lot of friends, who are working with other firms, and of course we don't share any kind of secret information, but I have a good finger on the temperature: who the other guys are, how skilled they are. There are some immensely talented people in the competition. But at the same time, judging by their reaction to me and my team, I think that they also respect our capabilities to compete against them.

Adam: In terms of the technology on offer, there are a number of technological advances that are getting a lot of attention, for instance FPGAs. Is that an area you've explored?

Andrew: We have looked at FPGA, general purpose GPUs, a lot of advanced technologies. We've found that some technologies are very beneficial, but some technologies are highly marketed and overhyped. FPGA is a great technology but it has relatively limited applications in terms of what would be suitable to change to run on FPGA. And I think a lot of the commercial FPGA solution vendors push hard on the marketing and, while there certainly are some benefits in key areas, it's a little bit too much hype for my taste.

Adam: As far as your overall trading style, you're very much latency focused. But at the same time a lot of people are seeing more opportunity by going for lower-frequency trading with longer holding times.

Andrew: On the latency race to zero, sometimes people look at it and they say 'Okay, there's a diminishing marginal return on technology performance and we've paid twice as much money and we've only saved one percent latency, so it's not worth it anymore'. But it's important to keep in mind that it's a relative measure. There's a dollar on the table in these arbitrage situations. And the winner is whoever can reach their arm out first and grab that dollar. So, it doesn't matter whether you're doubling the speed at which you can grab dollars, or if you're just a fraction of a microsecond ahead of the next fastest guy, you just have to be among the lowest latency. A lot of times, people are saying 'the race to zero and diminishing marginal returns in technology performance'. But what I really think they should be measuring is the marginal return of an investment in technology's ability to be able to continue to operate profitably and take these 'cream of the crop' trade opportunities.

Adam: As if Brazil isn't enough, you've said that you're also interested in Russia?

Andrew: I'm actually currently living in Russia part of the time now. So I've moved my home base from Brazil to Russia in order to get to the next step. But, of course, I travel back and forth all the time.

I'm still evaluating the Russian market from a trading opportunity perspective. Russia is immensely rich from a recruitment perspective of human resources talent. As all of your readers will know, there's going to be Russians and former Soviet Republic citizens on every team of every firm that reads your organisation.

Andre Auerbach

Andrew Auerbach in Russia

Adam: When do you think you'll make a decision on whether to move forward with Russia?

Andrew: I'd say about a year after we launch in Brazil. I have been, and continue to do recruitment from Russia, and trading the Russian markets for trading is certainly a high priority target for us to consider. But we haven't made that decision yet.

Adam: If you do do Russia, that will be two of the four BRICs. Is there a grand plan to do them all?

Andrew: The reason why I've come into the BRICs - it's an interesting question, a lot of people know where Brazil is on the map but they've never been down there to actually visit for themselves. What I like about the BRICs is the crossroads between what we've discussed before, of being developed enough that there's a reasonable opportunity to trade, but not so developed that it's crowded out. The intersection is very interesting because the margins are so much more. But of course there are new unique problems: how to manage different cultures, different regions, different languages, and other problems that you do not know you don't know.

Adam: Can you give an example of the sort of problem you would never face in a place such as Chicago or New Jersey?

Andrew: Sure. Internet bandwidth is one example. You have a couple options for internet bandwidth in Brazil. If you wanted to get a dedicated internet link, you'd be paying thousands of dollars for a very small amount of bandwidth. In America, for thousands of dollars you get multi-gigabit per second connections. So of course, the physical infrastructure doesn't exist there. Other examples: it takes many months just to form a company in Brazil. Not to do anything, just to form the legal entity. Now they're changing the laws and they're targeting to have companies ready to run within five days of the time you submit an application. They see the need to improve and they're working on it. But, there are many issues like that.

Adam: You talked about how the technological improvements in Brazil make it much more viable. As a user, as a client of exchange technology, it's been an interesting couple of months. It seems like every time you look there's another glitch on another major exchange. What sort of impact does that have for you as a market participant? How much does it destroy confidence?

Andrew: The confidence is the main part. The exchanges do a very good job of being fair and even handed about it, but it would be really great if they removed some of their discretion, and made it more systematic. It's very difficult for algorithmic firms and for non-algorithmic firms to decide if they should stay in the market in a questionable, edge-case scenario. Because you don't know if your trade's going to get busted, you don't know if you're going to have unintended exposure. If the rules were very clear, I think it would help the market stabilise and recover much faster. And if they come close to the edge, I think that it would auto-balance them back, so that they don't fall off the edge, because people would have the confidence to say, 'Okay, we're one tick away from the threshold, but that's one tick, it's within the bounds, that's what we're doing and it's okay.' Instead, to have the thresholds with a little bit of discretion in enforcing that, I think it's definitely implemented with the right goals in mind. But I think it would actually be more effective to have specific, zero-tolerance rules for that.

Adam: The technology problems are happening with such regularity. Do you think that's just bad luck, the last couple of years?

Andrew: I would look at it as a couple of things. One is certainly that exchange competition is growing. You see with the merger of BATS and Direct Edge that's proposed… People are trying to consolidate to keep up margins and profit. So I think the high degree of competition in a fragmented market is driving exchanges to compete on technology, among other things. And that may mean exchanges push out and deploy systems - not prematurely, but before they would have as much testing. So you have a reduction in the amount, duration and quality of quality assurance testing before things go awry, in order to get that competitive edge. You know, 'We're going to have a new feature, we're going to improve the system, we're going to reduce latency' - Whatever they're doing, they're trying to get that out the door faster than the competition so that they can have that extra amount of edge.

Another thing that I would blame it on: the increased data volumes are going to lead to increased errors because of the law of large numbers. If you're dealing with the amount of information the exchanges do, then there is nearly 100% probability that you're going to eventually face some problems that always existed but may not have noticeably surfaced with a more limited quantity of data. One could buy lottery tickets every week and never see a winning ticket, but if you look at the larger data set of all buyers then you would see winner events that always existed but were effectively invisible with the smaller data set of one individual.

I think the enhanced data volume that's going through is just going to lead to more exotic, edge-case, corner-case scenarios. So the frequency of these flares that cause the exchanges problems, has been increasingly observed and I think it will continue to increase as computers get faster, things get better, more volume goes in. And there's nobody really to blame. I think it's just that the improvement in hardware systems has exceeded the rate of progress in the software development side and business development side of the exchanges.

Adam: As a user, do you want to see more activism from the regulatory side or do you feel the exchanges overall do a good job and they should be left to solve it themselves?

Andrew: I'm a fan of market capitalism. In some instances regulation could be helpful, but I don't believe that government intervention is the answer to most problems. I have a high degree of respect for the regulators at the SEC, the CFTC and their counterparts around the world. But I think the solution here is to have more awareness rather than trying to fix it. Of course it's in the exchanges best interest to not break down. I think that it's more effective to have awareness, to say, 'Okay, exchanges are not perfect, don't expect that if you send an order that it's absolutely going to get filled at a reasonable price.' Be more responsible for your own action and take charge, because if you put the responsibility on the regulators, I know they're going to try to do their best, and they have many tools at their disposal to do a good job, I think that ultimately you're going to have some people who are disappointed when regulatory efforts don't succeed because of many factors. Nobody can perceive the future, neither the exchanges nor the regulators, so I think that it would be better to have awareness that these are not perfect systems, they're designed by people. Of course they're going to have bugs like any other software or engineered system. There can be exotic circumstances that can cause problems. And be aware of that, be cautious and act appropriately. Don't go right up to the edge unless you're willing to fall off the edge. In my viewpoint politically and economically, that's a better way.