The Gateway to Algorithmic and Automated Trading

Charles Huang: One in a billion

Published in Automated Trader Magazine Issue 27 Q4 2012

He was at university by the age of 14. A decade later he became the youngest-ever senior vice president at Prudential Securities. Skip ahead a few years and he was buying a Chinese commodities exchange, soon to be transformed into a high-tech trading operation. Charles Huang is clearly someone who knows how to do things quickly. Now, as chief executive of Bond Trust, he is staking out new territory on the Chinese mainland as one of the few local investment companies to employ algorithmic trading techniques. Adam Cox hears all about his remarkable story and what he thinks of the trading scene in China.

Charles Huang

Charles Huang

Adam: Why don't we start with the origins of Bond Trust. It doesn't have the typical back story of a fund in that it actually began as an exchange. Could you tell me a little about that?

Charles: Definitely. The history is not as straightforward as you typically encounter for an asset management or global hedge fund entity. Essentially, I was working for Deutsche Bank until the year 2000. At Deutsche Bank I was running a securitisation finance book for Asia. After I left Deutsche Bank I did quite a few things, including investing in a few internet companies and education companies. So, after a few years I decided to start a boutique, an advisory firm; that was 2002. I started Bond Trust because my background was in fixed income, trading, research and origination. I was advising a few property developers in China on how to raise money from outside China at the time. Over the years, we did quite a good business on the advisory side. But as the company grew, we figured out we wanted to get into mainland China, because Bond Trust at that time was based in Hong Kong. It's an SFC (Securities and Futures Commission) licensed company in Hong Kong.

I was looking for opportunities - back in 2004, 2005 - for how to get into China. There's a lot of regulation and it's very difficult for a foreign company to get into China directly. I'm actually a native of China myself. So I bought a company which was a metals exchange in China before, and we established an office in China that way. As we got into China, we looked at the onshore markets and decided the asset management business would be the main business we wanted to focus on. We looked at how to utilise international expertise and technology within China, and quantitative trading and asset management was really the area we pushed for. We learned quite a bit over the last few years, in terms of the landscape of the market. That's the history regarding Bond Trust and how it got into China.

Adam: Interesting. So it was actually easier to buy an existing company, even though it wasn't an asset management company, than to start a new one. Is that a regulatory issue?

Charles: It wasn't a regulatory issue per se. But it is more difficult to start a business in China, and if it doesn't have a lot of history then you typically cannot do the asset management business easily.

Adam: So Bond Trust is a mainland China-based business, it's not based in Hong Kong?

Charles: Bond Trust is an SFC-licensed company in Hong Kong. We have an affiliated company, Bond Trust Consulting, which is in China. By law we cannot have an international company directly own an onshore company in regulated businesses. As I'm Chinese myself, I own the onshore entity.

Adam: In 2010 you announced a new alpha generation platform. Could you tell us a bit about what that platform does?

Charles: We developed some strategies ourselves over the years in China, then we decided how to make it into a more efficient backtesting, forward-testing and real time execution platform. We developed the script language which enables us to do that, and then later on we actually combined the platform with Deltix, which is a US-based company.

The quantitative trading based in China is very new. It's almost like a new frontier. People in China typically do not believe in quantitative strategies or automated trading at all - because many traders make their money by, sadly, inside information or manipulation of the stocks. So they say quantitative strategies don't really work; very few investors employ quantitative strategies. I had to go pretty much from the ground up, by getting the market data, cleaning the data, having a backtesting platform, trying to get an execution connection with the exchange, and with the broker's platform. All these things pretty much needed to be done from scratch, because there's not really much infrastructure in China in this space. We spent almost a year and a half doing this.

Adam: What you're doing is somewhat rare in mainland China. When you mention strategies, are they multiple asset classes? What kind of strategies are we talking about?

Charles: In China, the things you can trade are rather limited. There are not really options at all. You effectively cannot borrow stocks. And also stocks are traded on T+1 so you cannot buy and sell in one day. Also the transaction costs are very high. On the selling side for stocks there's a 10 basis point stamp duty. The brokerage commission each side is five basis points.

Adam: So with all that, there's no chance for HFT.

Charles: The only place where you can really do HFT is the commodity exchanges. There are three commodity exchanges in China. One is in northern Dalian, there's central Zhengzhou and there's Shanghai. There's also another financial futures exchange, which is stock index trading - that's about one year old now. So with these four futures exchanges, potentially you can utilise HFT strategies. However, the regulators don't really like to see it. They limit the number of times you can trade in one day, so effectively you cannot turn over too much. There are people trying out HFT a little bit, but it's not really on a large scale, not like ultra-high frequency. And on the market data side, it's very slow. For stocks, level two data only comes every two or three seconds, and there's no order ID either. There are a lot of impediments in terms of fast moving markets or trading costs, and so all of this is preventing HFT strategies.

Typically, what we trade will have to go anywhere from a few days to about a month. We have developed a strategy which arbitrages between stock indexes and the underlying stocks and also another one is ETFs with underlying stocks. We also have a lot of strategies which are more traditionally like factor model, alpha model strategies.

Adam: What sort of alpha generation are we talking about here?

Charles: We're talking about traditional factor models which analyse the value, the growth, momentum or mean reversion, the size of the company and then we sometimes incorporate some local news as well.

In China, 70 percent or 80 percent of trading is done by retail. Only 20 percent is done by institutional. And all these traders are influenced by information, by news flows. A lot of times you can't just simply play by the value or growth factors. You've got to really figure out what events are happening and how to adjust a position based on the events. Those sorts of things have to factor into the traditional alpha strategies.

Adam: How does the presence of tens of millions of retail investors change the trading landscape? Does it make a difference in the way you develop your models?

Charles: I think there are advantages and disadvantages. You've got to really know the markets well. In the first place, people tend to be extreme. When people are bullish, your taxi driver, your maid - they're trading stocks. And the market develops bubbles. That's what happened back in 2007. When things go bad, nobody wants to touch it. Everybody's dumping it [the market]. Very much a retail mentality. Things tend to be more extreme in the China market.

We can't really expect things will be rational. On the good side, retail traders are relatively slow and they tend to be more emotional. So sometimes you can probably take advantage of that and trade against them and make money.

Adam: You're referring here to opportunities for arbitraging against a stock index?

Charles: More on the news side. The news sort of filters, or propagates. It goes from the source, to the analyst or insiders, and then it goes to the institutional guy, and finally it goes to the retail guy. By the time the retail guy gets it, it must be late. So they trade on the wrong side of the equation there. You've got to figure out whether the news has made it to the retail side already. If it's on the retail side already, the opportunity's probably gone.

Adam: How do you figure out where you are in the news cycle? Do you have some automated method for doing that?

Charles: We have a database, a data warehouse, and we combine judgement with an automatic news ciphering mechanism to generate signals.

Adam: Now, as a China-based company you're able to trade A shares and not just B shares.

Charles: The B shares to a large extent are irrelevant. It's probably one millionth of the volume. They're talking about phasing out the B share market in China. B shares will probably be converted into A shares, or H shares (mainland companies listed in Hong Kong). So basically it's A shares and stock index futures in Shanghai.

Adam: There's very little algorithmic trading in China. Would you describe what you do as algorithmic or is it more a matter of using some automated systems for generating signals?

Charles: We tend to be a combination of signal generation and incremental algorithms like VWAP or similar strategies for execution. But as you pointed out, rightly, even institutional investors are not using algorithms. So even that 20-30 percent, the institutional investors, which are mainly mutual funds, they have trading rooms typically with, like, five traders in a trading room manually typing orders into the market. That shows how people in China, they don't believe in algorithmic trading. I talk to a lot of mutual fund managers in China. I say, 'Hey, why don't you look at our tools? You can cut costs, you can probably increase efficiency'.

First, they think they can do better than algorithms. And second, they have to have a lot of layers of control, so it becomes a manual process. I believe over time that number is going to change dramatically because with trading execution, human beings on average definitely are not as good as computers. I do see more JV types of mutual funds, they're starting to test algos already.

Adam: Does this give you guys an advantage at the moment, the fact that you're more automated than many firms?

Charles: You have to understand the dynamic market structure as well. It's not automatically going to give you an advantage. But you try to be a little faster, probably try to anticipate the movements a little bit more than other people and save a little cost on the trading side. So that's what we're trying to achieve. But it's by no means easy. We ourselves are learning through the process as well.

Charles Huang

Adam: What sort of timeframe do you see for the market changing in this respect?

Charles: We are in the vanguard of the market. We are trying to advocate the advantage of algo execution and quantitative investments. We actually talk to most of the mutual funds in China. Most of the mutual funds only have a very small quant department and their IT technology is outsourced. They typically cannot change very fast; I don't think the pace will be very fast. People are relatively cautious. The trading costs are very high, so saving like half a basis point or basis points for them doesn't really matter that much. I mean, they don't really feel that much pressure to save costs.

You've got to see the major competitors start adopting algos. And the regulators, they're very important. If the regulators approve the adoption of algos, and a few key guys adopt them, it can be very fast, this sort of change. But before that happens, I think people will just be testing it.

Adam: What's the actual legal status in terms of using algorithms in China? On one hand, it's taking place. But it seems a little grey as far as the regulatory approach.

Charles: Fund managers cannot execute orders. They have to pass orders to the middle office, to traders. And the traders, their only job is to put orders in the market and follow all the rules.

The regulation does not say you cannot use algos. But the checks and balances they require basically make algos impractical. It's not that they say you cannot do algos; there's nothing against VWAP.

The regulators sooner or later are going to be more market-driven. They've got to think that basically that's how you improve market efficiency. Instead of saying every order has to be checked by a person - what's the purpose of checking that order? What are they trying to prevent from happening?

Adam: Having seen events like the Flash Crash in 2010, and then various technology glitches and most recently Knight, I can understand why there would be a certain degree of caution.

Charles: The US and international markets are very liberal, very market-driven. And China, on the other hand, is very tightly regulated.

For example, all the stocks have limit ups and limit downs of 10 percent. Some of the smaller stocks have five percent limits. So you won't have a crash or meltdown to zero or one cent in one day. That will not happen. But the regulators are cautious. They're looking at the international markets for lessons too. That's why they're afraid that if people use HFT strategies to make money, they're essentially taking money away from other people. When they approve QFII quotas - qualified foreign institutional investor quotas - they tend to want to attract long-term investors, like insurance companies. They're not giving quotas to hedge funds yet. They don't want too much volatility in that sense.

Adam: Perhaps we could talk about the trading infrastructure. You spoke about how market data was slow. Why is that, is it changing, and how does it affect you now?

Charles: There are two stock exchanges in China. One's in Shanghai and one's in Shenzhen, and all the trades have to be cleared inside the exchange. You can't do internalisation. Like in the US, if you're a brokerage firm, you have buy orders and sell orders, you can internalise; you don't even have to go to the exchange. In China, every single order has to go to the exchange to be cleared. As a result, everybody jams into the exchanges. The exchanges don't really have a lot of pressure to increase their speed for disseminating information or reducing the costs. They're making a lot of money, so they don't really have any incentive. In China, there's no competition. They're improving all the time, but there's no competitive forces there to try to force them to have better market feeds or to execute trades faster. It's a slow evolution process.

For us, the shortest strategy is three to five days. And if you trade too often, the cost of trading is going to eat away your earnings. Round trip is at least 20 basis points, and that's not counting the market impact. To count market impact, it's more like 40 basis points.

Adam: What sort of timeframes do you typically have for your strategies?

Charles: On the index arbitrage, it depends more on the spread. That's typically a few days to about a week or two. One week is a typical time horizon for that strategy. For alpha strategies, our baseline is probably about three weeks or so. It depends on the information flow. Sometimes it's a little shorter, sometimes a little longer. But typically not more than about a month.

Adam: In more developed markets in the West, arbitrage can be less than a second. But here you can have arbitrage opportunities lasting well over a day.

Charles: If you're doing the index versus the stocks, where the spread widens out or comes in, that takes days and weeks to converge. Trading costs are pretty high and the number of people trading that market to take advantage is not very high. It's not like the developed markets where things happen and everybody jumps in.

Adam: Back to the subject of the exchanges and the lack of competition. Is there any chance of new competition coming in there?

Charles: That's impossible. They're basically part of the government enterprise.

Adam: As far as what you're doing, are you aware of other firms doing similar stuff?

Charles: The bigger ones are probably in the order of 100-200 billion renminbi, which is about $20-30 billion. They have not really come to the point where they want to invest a lot in quant strategies, but they have recently hired people from overseas. Like China Southern Fund, which is in Shenzhen, they hired a guy from Barclays. And another fund in Shanghai hired another Barclays person.

But when they say quantitative funds, they're not really using algos to trade. They probably use some models to generate signals. So that's one category of people who are testing out a little bit more on the quantitative signal generation approach.

And then there's another category which, like us, are private. There are probably more than 1,000. Most of them don't use quantitative strategies at all. Probably less than one percent.

Then there are a few people who traditionally trade commodities - cotton, lead, zinc. These people used to speculate before. Now they are realising they need more automated tools. So those traders are gradually adapting more to automated trading in that space. In the commodities futures space, more people have a need for automated trading, while in the stock space people are testing it out but are a lot more cautious.

Adam: The technology they're using, are they mostly buying it from the West or is it built in-house?

Charles: For the backtesting, forward-testing and execution platform, it's not necessary to reinvent the whole wheel. I think more of the Western firms will come into China to deploy the technology in China. And hopefully the more that Chinese firms go into that direction, they'll need to realise that they need that technology. Then they'll probably ask, 'Do you develop it yourself or adopt from developed markets?'. More likely than not, they will want to adopt from developed markets.

Adam: On the macro front, there's been a never-ending obsession in the West with the level of the yuan. Does that affect you, or how do you play that?

Charles: I think the international monetary flows are more important than the macro base. But what's even more important in China is the retail traders' actions. They are more on the driver's side in terms of determining market prices or how the market goes to extremes. In terms of the macro yuan level, obviously the Western governments are very obsessed about it, trying to push China to get the renminbi to appreciate even more. But from what we see, the market forces aren't really there.

Adam: We've been talking about your activity on the mainland. You also have Bond Trust in Hong Kong. Do you trade there as well?

Charles: No, we just have an advisory firm in Hong Kong. We tend to act more as an onshore firm, trading the A shares rather than trying to access the onshore markets from an offshore vehicle.

We're a young firm. We're not as big and developed as a lot of international large companies. If we want to compete with the big boys then we probably have less of a chance. But going onshore, by utilising technology and knowing the local market better, it gives us an edge.

Adam: Finally, about yourself. You began university at 14, so you had an early start! And then you went off to MIT. I know there are people who start university at an early age, but that strikes me as particularly young.

Charles: In China, back in the '80s, there was a push for science and technology. A lot of kids wanted to become scientists at that time, including me. There's a special [programme] for the gifted young, about 50 of us a year are going to university anywhere from 13 to 15.

Adam: So 50 people out of more than a billion, that's a pretty select group.

Charles: It's not many. Probably 30-40 of us went to the US to study, at probably every major university, including MIT where I went, and Yale, Harvard, Princeton, Chicago, Berkeley.

Adam: I once read about how Chinese Premier Wen Jiabao liked the idea of Chinese students going off to the West and then, like sea turtles, swimming back to China. That's kind of what you did. But others prefer to stay in the US or Europe and make their money there. Why did you want to come back?

Charles: The opportunity. I always wanted to come back. I worked in New York City for five years at Prudential Securities and basically worked up the management ladder. I see China just at the beginning, waking up and providing a lot of high-growth opportunities for people like us. After five years at Prudential, I was already trying to come back to China. I approached a few firms and eventually got a job at Deutsche Bank, which sent me back to Hong Kong to run my group. After that I started my own company.

Adam: What's the long-term ambition?

Charles: Because we're partners, we've got a reasonable amount of money ourselves. Our goal is, really, first trying to manage our own money well. That's what we've been trying to do for the last couple of years. Second, obviously we want to make it a successful business. But we are probably more patient than other people in the sense that we want to make sure our strategy is right. Once we have good strategies, then obviously we can open up to our friends or even to the market to manage money for other people.

I don't expect we'll open up to outside investors for maybe a year or two years' time. We want to make sure we can manage our money well first.

Adam: What kind of results have you been having?

Charles: We have two types of funds. One is more of an absolute return type, more on the arbitrage side, and another one is more like a beta-to-index type. So on the arb side, we achieved a low teens return, which is reasonable. And on the other side, basically 10 plus points over the index. Not exciting yet. We're trying to figure out how to ramp up that side.

Shenzhen City