Asia is a vast, diverse market, where algorithmic trading is only just starting to take off. There are a number of reasons for both this and the fact that it looks like it will be a slow burn for algorithms in the region. One reason for this is that each country in the Asian market has its own way of doing things. These include highly diverse communication technology, widely varying bid/offer spreads, and regulatory issues such as the banning of DMA trading.
Sang Lee, managing partner at Aite Group, comments: "These reasons make it very difficult to create a homogenous view of Asia as a market, because each country has different standards and ways of operating."
Technology and standards
As far as technological standards go, Asia is a hodgepodge. There is no one dominant standard in use in Asia, which makes it harder to start an algorithmic revolution in this region, more difficult to maintain it, and therefore tough to increase growth.
FIX adoption in Asia is sporadic at best, which creates a significant barrier to algorithmic trade growth. Lee comments: "I do believe there are other options and other players out there that can create standards for people to work with here. The low usage of FIX is a barrier, but it's nothing we can't overcome. This is proven through the movement of big players from the US and European markets to Asia."
Yet FIX is extremely important for the spread of algorithmic trading in Asia, Hani Shalabi, executive director of direct execution encompassing DMA and algorithmic trading at UBS, says. He claims that even though DMA and algorithmic numbers may be lower in Asia than in the US and Europe, FIX use is quite high for working orders. However, he adds that not everyone is using FIX version 4.4. "The reason for this may be that the features offered by newer versions of FIX are not necessarily something that everybody needs, so people won't spend the money upgrading when they can make do."
According to Christine To, vice president head of Asian equity trading at T Rowe Price, FIX connectivity in Asia is only around 35%. She adds: "It will filter down, but the cost is an issue. We deal with some smaller brokers in local markets and they just refuse to set themselves up with FIX because of the cost."
Christine To, T. Rowe Price
Another issue Asia has in implementing algorithmic trading is its spreads. Many Asian stocks have very wide spreads - for example NTT DoCoMo has a spread of 60bp. In markets such as Singapore the average dealing spread for stocks is around 70bp, so while the actual display liquidity in the market can be large, it can be extremely difficult to trade there because of the equally large spreads. Some stocks also have low turnover and high volatility - a less than ideal combination for algorithmic trading!
Despite the obstacles, various parties throughout the industry
are discussing how the popularity of algorithms in the US and
Europe might be replicated in Asia. "I would say Asian markets
are in their infancy in algorithmic trading," says Aite's Lee.
"Some are more ahead than others, such as Australia, Hong Kong
and Singapore, and Korea and Japan. The local domestic players in
these regions are not so advanced in algorithmic trading
strategies. The Asian market is just starting."
While there are a small number of large buy side institutions that are as sophisticated in their use of algorithms as their counterparts in the US and Europe, the vast majority of Asian clients are less aware of algorithmic trading and its benefits and will need significant encouragement and support before adopting. Gyan Newman, regional head of ecommerce, for Asia Pacific at Barclays Capital, states: "We are spending a lot of time talking to our clients about what algos are capable of, and what is just hype."
Newman says that voice-driven, manual order requests still dominate flow in exchange-traded asset classes throughout Asia. She claims her Asia-based clients prefer to leave orders at an execution desk. To work with that style of trading efficiently, Barclays Capital is using its BARX algorithm engines internally to handle more manual orders concurrently, which also allows the company to tweak its algorithms for the Asian markets.
Yet Asia's position at the back of the algorithmic trading global market is almost proving to be an advantage. What happens to algorithmic trading in Europe and the US today is serving as a guide to where Asia will go.
Andrew Freyre-Sanders, JP Morgan
One of the driving forces in the growth of algorithm use in Asia is increasing competitive pressure from global players. Local banks are feeling squeezed, as they can see global banks moving into the region and tweaking their existing algorithms to suit the Asian market. This is pushing local and regional banks to develop their own algorithms or to buy in expertise as quickly as possible, with Nomura's purchase of Instinet being a case in point.
Shalabi states: "The reason a lot of the big banks are moving to Asia is because without an Asian offering you're missing the fastest growing piece of the global algorithmic offering. The knowledge that someone will offer algorithmic trading to clients in Asia means that all the big banks are rapidly moving into this market place. Additionally the big banks have the resources to invest in human resources and any area that needs attention; we had all these algorithms already in the US so we just had to ship them over quickly and adapt them to this market."
UBS is also trying to increase the use of its algorithms in Asia is by having local quantitative analysis teams on the ground. These teams look at all trades each day in a local area to see if algorithms are achieving best execution.
Exchange rules and regulations vary drastically among Asian countries, making it much harder to develop single algorithms that serve multiple markets. Barclays Capital is one of those companies that have created a dedicated development team focused on enabling a core suite of algorithms across Asian markets, so as to shield clients from these complexities as much as possible.
Three firms in particular have been early algorithmic movers in Asia, namely Credit Suisse, JP Morgan and Goldman Sachs. These larger companies have already spent a lot of money and time building and developing their algorithms on their home turf, so they are now looking to leverage that investment across Asia by adapting them to local market conditions.
John Feng, principal consultant at Greenwich Associates, comments: "There is clearly a very intense race going on in algorithmic trading. It is a story all the major brokers are very focused on right now. On the sell side, this is one of the key areas for brokers to differentiate offerings from those of competitors. On the buy side, there is a growing focus on trading execution quality, where trading desks are no longer viewed as order takers, but as a way to achieve best execution and add to performance. This isn't just about commission rates, but about achieving optimal overall execution results."
John Feng, Greenwich Associates
As mentioned above, simply exporting existing algorithms unaltered from the US and Europe to Asia is just not feasible. Many standard algorithms from Europe and the US are intended for use in relatively liquid markets and are thus inappropriate for many Asian stocks. Furthermore, once algorithms have been adjusted to suit the market, they still have to be implemented and maintained within the framework of an OMS or EMS and the technology cost of that is significant, especially for smaller players. "My experience here is that it's more challenging to trade algorithmically in Asia than elsewhere in the world," says Freyre-Sanders.
There has been a lot of innovation over the last four years in Asia in terms of work on algorithms. However, many Asian institutions still use VWAP as an important benchmark for measuring execution levels, so VWAP-based algorithms may do well in this region over coming years as the market assimilates algorithmic trading.
Yet Shalabi says the problem of less liquidity in the market means that some names will not trade well in VWAP. "Algorithms that can look at the specifics of this market will do well, and those that will prove more successful will be the ones that are better at trading more difficult names. Algorithms imported from the US and Europe will not do well as this market is so different. Algorithms need to be developed locally in the Asian market, with someone on the ground tweaking them. Clients will notice which algorithms perform the best among brokers based on comparative benchmarks."
This is a highly specialised market; an example of its peculiarities is in Japan, where the structure of the market is very different to the US and Europe. Here, a higher percentage of volume is traded during auction periods and around the lunchtime breaks.
Anomalies such as the break time rush mean algorithms need to be developed locally in order to be effective. Freyre-Sanders comments: "The core VWAP and inline with volume (POV) properly adjusted to Asia will still be the bread and butter algorithms here. But more reactive, liquidity searching algorithms should also provide value. Western algorithms will be able to work in these less liquid markets, but they will just have to be more proactive in terms of strategy to be able to work with bigger tick stock. I think the initial focus will be on single stock, then gradually to basket algorithms. We see demand for algorithms on futures already though."
However, Lee says that many are moving away from VWAP because everyone is using it. "Why is it a good thing to maintain mediocrity?" he asks.
Quantitative hedge funds' quest for arbitrage opportunities is also likely to push the use of algorithmic trading in Asia, as it becomes harder for funds to find such opportunities in the West. Logically, hedge funds from the US and Europe will migrate to Asia as opportunities in their own markets are arbitraged away.
Gyan Newman, Barclays Capital
Feng says Greenwich Associates has seen diversification of investment as a reason for hedge funds moving to Asia. He states: "There has been an increase in the number of hedge funds as well as assets under their management globally. Hedge funds operate by looking for inefficiencies in the market, which are increasingly difficult to find when you have many smart people with considerable capital all looking for them. For hedge funds moving into alternative investment strategies, geographic diversification is a natural dimension to explore."
Asia is a collection of markets where many stocks are thinly traded. This creates opportunities. In addition, Asia has consistently performed well from a market return perspective over the last three to five years."
Also, Feng adds: "Market lessons from recent years illustrated the risk of a single-strategy approach. A case in point is the convertibles market over the last few years, which experienced a downturn in 2005 that led many funds to either diversify into other strategies or shut down."
Rajeev Baddepudi, hedge fund analyst at Eurekahedge, comments that the hedge fund industry is becoming more mainstream, which will only increase its development. The main investors in hedge funds currently are at the top of the market, with wealth managers and private banking firms, such as superannuation funds in Australia and large investors in Japan, leading the way. But that investor base is widening. Smaller investors are becoming interested in these hedge funds and are trying to understand how they are able to make so much money, he states.
Apart from hedge funds from the West trading Asian markets with algorithms, intra-regional hedge fund growth may also boost activity. According to Eurekahedge, as of September 2006 there were around 950 Asian hedge funds, including those in Japan and Australia. This is an increase of 23.9% in the 10 months since December 2005, and a 29% increase overall according to a report from Asian hedge fund analysis firm GFIA. While Eurekahedge's database shows there are 388 funds on its Asian performance index, GFIA's data points closer to the 450 mark. Of the 735 funds in the region, 306, or 46%, are funds with over USD50 million under management, and with over a 12 months history, and therefore potential investments for international allocators. That is a rise of 3% in 10 months.
Peter Douglas, GFIA
Japan is the first place that many think of when contemplating a move to Asia. Greenwich Associates estimates that among larger institutions trading Japanese equities, 12% of trading volume is done electronically
(figures taken for mid 2006). The research firm expects the trend of e-trading adoption to accelerate in Japan, growing to 18% by mid-2007 among larger funds, mostly driven by the increased use of algorithmic
trading, not DMA. Feng states: "In the short term, institutions foresee the key driver to be algorithmic trading, as opposed to nonalgorithmic trading."
Japan is doing well in algorithmic trading because of its liquidity, says Shalabi. He says Japan is a powerhouse right now and will continue to go forward. Peter Douglas, principal at GFIA disagrees. He comments: "This sounds odd; most of the shorter term traders I talk to avoid trading Japan as trading execution is too slow because the market infrastructure is very outdated."
Shalabi sees Korea as likely to become a DMA and algorithmic powerhouse of the future. Korea is as an extremely liquid market, and Taiwan is another possibility, Shalabi states, although he says the periodic auction system does keep some investors away. In Hong Kong the exchange is trying to increase volumes by narrowing spreads, but in Singapore little is being done in this respect.
Feng comments: "The long term future of algorithmic trading and which algorithms will prevail in Asia remains to be seen. A lot of these algorithms are still relatively new
even in the US and Europe, so most institutions may not have enough data points to assess the true performance of these algorithms."
On how quickly algorithmic trading will grow in Asia, To states: "I think algorithmic trading will grow fast. All the investment bankers are launching algorithmic tools and it's been widely promoted. People are hearing about algorithmic trading, they're talking about it and they will try it out and learn to use it like we did; we had algorithmic trading for eight months and didn't really use it, but when we did we loved it. Having said that, there is always a need to talk to traders. The Asian market is very fragmented so there will always be a need for full-service trading. There are some trades we wouldn't even think of putting into an algorithmic engine as it just wouldn't work."
As to what will drive the spread of algorithmic trading in Asia, To says one area that is lacking is training from providers of algorithms. She says: "A lot of people do not understand algorithms and even do not know what an algorithm is. Training people on which trades should go into which algorithms and how to cleverly use algorithms is vital if people are to start to feel more comfortable with algorithmic trading in Asia."
Case Study: Algo user
T Rowe Price's Asia has had algorithmic capabilities for some time, but was initially reluctant to use them, according to Christine To, vice president head of Asian equity trading at the firm. "We've been using algorithmic trading for a while and have had access to it for a year, but we've only actively used it in the last four or five months," she says. "The reason for this is we trade regionally in Asia but not all markets here have algorithmic trading, plus we were just getting comfortable with it."
To says T Rowe Price uses algorithmic and DMA tools when trading Japan, Hong Kong, Australia and Singapore. She said the main benefits her dealing desk has gained are the fact that algorithmic trading is cheaper, with commission rates that are only a fraction of those otherwise paid. Additionally, she says algorithmic trading provides anonymity and complete control of the trade. Yet To is conscious that algorithms are not an easy way to trade: "Some people use algorithmic trading only because they think it's a cheap way of trading; I see some of my competitors putting everything into an algorithmic engine and just leaving it.
We don't just put an order into an algorithmic engine and let it run. We might put just 20% of the order into the engine and watch it to see how it does. I slice the order to see how the price performs and how the market moves and this enables me to completely control an order."
T Rowe Price acquires its algorithms from a number of providers on the basis of what individual traders feel is most appropriate for their trades. These providers include Credit Suisse, Citicorp, and the FIX-based Bloomberg Trade Book, on which To comments: "Bloomberg has cornered a market because it's not a merchant bank, so people feel more comfortable using it because it's a third party."
Japan: Strong liquidity