I recently sat down with my colleague, Kristian Kerr of DailyFX, to interview him about his time as a former research analyst and portfolio manager at FX Concepts, previously one of the largest currency hedge funds in the world.
Kristian covers a lot of ground in this fascinating interview and it’s one you don’t want to miss.
The interview is available in audio only format or via the transcript below.
B: Hello Everyone. This is Bryan Fletcher, Product Manager for Algorithmic Trading at a major FX broker and I’m joined by Kristian Kerr of DailyFX.
K: Hi Bryan, good to be here.
B: Thanks for joining me.
B: Alright, well Kristian joins me today and I’ve been very interested in talking to Kristian about his experience working at FX Concepts. So Kristian, could you just give us a little background on what FX Concepts is and did and your experience there.
K: Yeah. FX Concepts was one of the largest currency focused hedge funds in the world. They did everything from overlay to absolute return strategies, but you know at their height, they were about a $14 Billion dollar fund. So they were kind of one of the few hedge funds out there that focused on the FX business.
B: So, what did you do specifically there? You said you were there for about 7 years?
K: Correct. I started as a research analyst covering G10 currencies and emerging market currencies and then after some time there worked my way up to proprietary trader/portfolio manager, so taking on risk on behalf of the firm.
B: Very nice. So, what can you tell me about the strategies that FX Concepts used and how did their model work? And how did you fit into that?
K: Yeah, it was primarily a systematic fund. You know, I would say about 90% of the trading was model based, systems based. There was a discretionary overlay from traders, portfolio managers, and the CIO team and that could consist of anything from trying to help on execution to overweighting a position/underweighting a position to trading instruments that you’re not getting a model signal in. It was basically just kind of a hybrid strategy to try to add alpha to the performance of the fund, which I think worked pretty well.
B: So you bring up a couple of interesting points. First, I want to ask you more about the models used by FX Concepts. Can you talk a little bit about the inputs used - price, volume, sentiment? What types of inputs were used in their models?
K: They looked at anything. They had some pretty brilliant quants that came up with them. The initial model I believe was actually cyclical. So it was based on cycle elements that John Taylor came up with, but they evolved you know with the market and developed some pretty complex type strategies. I guess their most successful model was focused primarily on the emerging markets and that’s where they had some of their best returns and they were doing some interesting things with respect to indexing and making baskets out of currencies, but you know, one of the first real funds to aggressively trade carry as well.
So it was a whole host of strategies. It’s kind of difficult to boil it down to, oh they were a trend following fund, or a counter trend following type of fund. That’s not really how it worked, but on the discretionary side they would look at all of the things you mentioned. On the systematic side, it depended on what they wanted to achieve with the strategy, but it was varied in terms of the different systems that were running at any given time.
B: So in your view, you said the discretionary overlay was there to add alpha. Can you talk about that a little bit and where do you think that comes from? What can a human do better than an algo system?
K: It’s a decision you have to make, right. That’s always a debate that we have as traders. Whether you are systematic- a lot of guys will say let the system run and have as little human input and action as possible because the whole point of a system is you want that lack of emotion in your trading, but there’s also times in the market where you know, where you can tell that the noise, that the system might not be coming at the best entry point. Things like that.
Where that little bit of human discretion…I’m not saying when you run a system you should have it 50/50 what have you, but kind of like a minor percentage overlay does help, I think, in terms of things like execution, entry points, and timing. All of those types of things, I think, are things where you can get a little bit of an extra edge if done right. Where I think you get into problems is when you try to over-optimize and the human discretionary component becomes more and more. That’s where I think you run into trouble. There’s a fine line, basically, when you’re doing this type of thing. You have to be very diligent and careful that you don’t cross the line and become too active on the discretionary side.
B: We were talking earlier about the HFT type strategies that FX Concepts employed. Can you expand a little bit about what we were discussing earlier? What models did they use and what are some of the challenges of a fund that size implementing an HFT type strategy?
K: Yeah. I mean they were relatively late to the game on high frequency stuff. It really had already started to take off, so you know, I wouldn’t say it was something they were really known for. Also, the big issue with running an HFT type strategy as a large sized fund is that it’s not scalable. The big issue is after running a bunch of tests and running through a few different strategies, they ended up finding a decent high frequency trading strategy. What ended up happening is you couldn’t run it at decent enough size to make an impact on the broader fund P/L.
I think that’s where you run into those type of issues where you can very successful trading strategies on the algo front, but if you’re running any decent size of money, it becomes very hard to kind of make that impact broader in terms of your returns if you’re running $14 billion dollars or whatever. That’s kind of the good and the bad of it. I also think that running smaller size allows you some opportunities you might not get elsewhere when you are moving the market by your size, so it’s kind of a give and take in terms of what you want to get from it.
B: FX Concepts, from what I read, was in business for 32 years and then they closed up shop not too long ago. I’m sure you’ve had a lot of time to reflect on the reasons that led to that. Can you talk about that a little bit? What do you think ultimately led to FX Concepts closing their doors?
K: Yeah, I’ve had a lot of time to think about it. You know, as I’ve said, I’m very much kind of an Austrian Economist or of that bent, where I just don’t like intervention from authorities in the market. If you’re going to run a free market system, let the market run itself, right? So I think what was very detrimental to [FX] Concepts and a lot of the other currency focused funds that ended up shutting down around this same time, it wasn’t that they made bad trades, they weren’t a Long Term Capital [Management] that really just got too big and suffered from hubris and ended up failing because they made bad trading decisions.
They were in an environment where central banking authorities and other policy makers decided to suppress volatility in the aftermath of the global financial crisis. You know vis a vis QE, vis a vis the suspension of mark to market. So we had suppression in vol[alitity]. Currency volatility fell to its lowest levels ever if you look at CVIX a couple of years ago. What ends up happening, you have the markets not moving, so pensions and those types of vehicles that invest in these products didn’t see a need to be in FX overlay or FX type of alpha seeking strategies. It just didn’t make any sense when you have a realized vol[atility] of 5% in G10. So I think that is what really undid them. Just the fact that vol[atility] just collapsed and stayed that way for a while.
Of course the interesting add on or PS to this story is that was almost - once all those funds started going out of business was almost the exact low of currency vol[atility]. You know, so it kind of goes to that idea of a Minksy moment. Stability begets instability which is what we ended up seeing and now there is a void in the market for these type of big FX hedge fund players. There’s really only a few big ones left, but I think that was really what undid them was the fact that no one saw the need for currency management anymore because vol[atility] stayed so low for so long. Obviously what ended up happening 6 months later, you had the BOJ come in doubling down with Abenomics. You saw USDJPY take off. A few months later, the SNB undid the peg in EURCHF and we had Swiss Franc move 30% in a matter of minutes.
We went from very, very extremely low levels in vol[atility] to extreme levels of volatility. There’s a lesson to be learned there I think. When things get that way you almost have to start to think contrarian because the market almost becomes too accepting of thinking that regime is going to last forever.
B: So, we were talking earlier about some scalability issues with low vol[atility]. In terms of a retail trader, should there be a fear of low vol[atility], if you’re a retail trader? What’s different that a retail trader would face as opposed to FX Concepts?
K: Yeah, not so much the vol[atility]. I would say on the retail trading side, the big difference between a big-sized trader that is going to run a system like [FX] Concepts. I mean, in some emerging markets, they were 70% of the daily volume. When you’re running that size, when you have that big of a footprint in the market, it’s a very different process that you go through because you will essentially push prices up and down going in and out of the market. Which I think ultimately it’s an advantage to being a smaller trader because you can dart in and out.
You’re that small fast boat that can go in and out of places and that’s what you’re trying to take advantage of. So, I think that’s the big difference. When you’re trading less size and I think a big thing now that’s a lot different is that the FX market has changed a lot in the last 10 years. When I first started at FX Concepts in the mid-2000s, I mean we were doing 95% of our volume voice, over the phone. When I left, 90% of the volume was being done electronically.
There was this gigantic shift in the terms of the way the FX market works and I think that really is a huge change or benefit or potential catalyst for the retail algo trader. They’ll be allowed to take advantage of some of the changes in the micro-structure of the market that frankly didn’t exist a decade ago. So I think that’s another big advantage I would say. It’s just basically the expansion of the market. It was basically very antiquated and now it’s very modern. I think that should help someone of smaller size get access to things they wouldn’t have a decade ago.
B: When the change happened from Voice to Electronic, did that impact P/L at FX Concepts in any measureable way and did that change up your strategies in any way as well?
K: Yeah, that’s a good question. Theoretically, moving to electronic, you’re supposed to improve execution costs, but you have to ask yourself at what cost, right? You know, I always tell people this, for example when I started there, Christmas time would come around and from December 1st to December 25th, you’d walk in that office and there was literally every single space in that office would be covered by either some sort of cheese plate, cookie plate, wine bottles from banks that were giving us their thank you for the year because you are trading voice through them. When I left, we maybe got 1 bottle of wine. What I’m trying to illustrate with that point is that when you’re dealing on the voice side, you’re dealing with a relationship business, you’re talking to these people.
Who wants that type of service? It’s probably more the global macro, Soros type of fund that wants access to information. Is the guy running a system, does he need that? Probably not because his inputs are all technical, or primarily price based, you know, market focused. Not so much on flow of information. So it’s a different game, but you have to ask yourself – do you want to have both? Have a hybrid model. I know a lot of funds will still give a lot of business over voice because they want to have that liquidity access in case something happens on the electronic side.
In case you get an SNB type event, you still have relationships you can use if the electronic side goes down. That’s another reason why you’ll see that, but also they want the information. If you’re running a purely systematic fund, if that type of an event happens, you probably just shut down. You probably pull the plug for those few minutes. It’s kind of things you have to ask yourself. I don’t think there’s really a real answer, Bryan.
It kind of depends on your make up, the type of strategy you’re running and what you feel more comfortable with, but you know, I know just as many funds out there that just trade electronically and do very, very well and they don’t want, you know, it’s almost the things they talk about, the flow information ends up being noise. It all depends on what you’re trying to get from your liquidity providers.
B: I’m going to change gears a little bit. Speaking from my own experience, when I got my start trading and was trying to build an algorithmic trading strategy, I always felt like I was on the outside looking in and geez, if I just worked for one of these big players, I would know so much and it would make it so much easier to find profitable trading strategies. What are some of the takeaways that you have from working at one of the largest hedge funds out there and speaking to someone in my shoes who doesn’t have that experience, how big of an advantage is that?
K: I think we’ve kind of hit on the same theme. There’s pros and cons to size. The more money you manage, the better access you have to talent, things like that. Working with very, very smart people who can come up with solutions very easily to things, but you also give up a lot too. It becomes much more of a process – investment committees, those kinds of things. Again it’s a give and take. You want size, you don’t want size. It’s the same thing. I mean, I hear what you’re saying, but in the end, I would say I worked there 7 years and you think these big institutions have, somehow don’t suffer the same emotional pitfalls an individual trader does, or doesn’t fall under the same kind of dilemmas and traps and it’s all the same, but on a different scale.
Tudor Jones talked a lot about this in that famous PBS video, but when you’re dealing with money, in the end, the reason why I trade the way I do is that I do believe a lot of the market is about greed and fear and that’s kind of what drives us and that emotion. Just because you’re managing $14 billion dollars opposed to $1 million dollars doesn’t mean you’re not going to fall under the same pitfalls. So I would say it’s actually pretty similar, you just add a few zeroes on the ends on the trades that you’re looking at.
B: On that size of a fund, you talked about the transition from voice to electronic, how advanced were the execution algorithms for something like that.
K: It shifted with the market, right? When we first started, we were one of the first ones to use aggregation mainly because we had so many counterparty lines, so we were able to do that. We were one of the first ones to actually take advantage of some of the aggregation software. So it went from being almost we were taking advantage of the banks because we had almost a deeper knowledge of liquidity than they did to where they got more sophisticated as they started to figure out the electronic side of the business as high frequency guys left equities and started to come to FX.
The buy side went from being they’re the one in control, to, towards the end, basically banks could tell how much was behind your order, so then that game starts of basically you’ve got to splice up things and try to trick them and you get into that whole predatory flow and that’s the way I left. That’s how the markets had gotten into. Nowadays, everyone is trying to read how much is there and the market shifts almost instantaneously. It went from being very, very simple to taking advantage of to have that turned around on us when the banks basically figured it out, so to speak.
B: A couple more questions from me. What question should I have asked you that I didn’t, and what would the answer be?
K: Where’s the Euro going (laughter) or when is the Euro going to parity I would say. No, I think you hit on some really good topics here. I would just stress, and I kind of touched on it earlier, I think nowadays, given where the market has been and I’m talking FX, just this gigantic shift from being one of the most antiquated markets to being one of the more advanced now in such a short period of time that it obviously brings a lot of opportunities for kind of the retail algo trader that didn’t really exist.
There’s a lot of access here that just didn’t exist a few short years ago. I think that’s thing I would say. It’s an exciting time to be involved in this market, especially with where we’re heading. If you do think we’re in some sort of phase in terms of economic cycles, volatility, hopefully, will be here to stay. Theoretically, those using algos should be able to take advantage of it.
B: Well, good deal. Tell us, where do you want people to find you, connect with you, contact you and what are you focusing on next?
K: Yeah, I’m on DailyFX.com.
B: Great interview, Kristian. Thanks for your time.
K: Alright. Thanks, Bryan
Unique experiences within this interview and past performances do not guarantee future results! Past performance is no indication of future results.