Exchanges are upgrading their infrastructure to deliver speed, performance, capacity and fast market data dissemination.

As algorithmic trading has become mainstream in the US, and is growing fast in Europe, hedge funds and quant traders have crowded these markets and are now looking to apply their models elsewhere. Exchanges in Asia-Pacific are keen to attract these high volume players, and to this end, they are gearing up to become algorithmic trading friendly venues.

Asia-Pacific has lagged the US and Europe because the liquidity pools in the region are scattered and fragmented, and competition between exchanges is less intense. While the Asia-Pacific exchanges are eager to attract algorithmic traders, the challenge lies in understanding their requirements and anticipating the level and extent of their participation.

Algorithms can be divided into three categories. The first type is an arbitrage approach whereby algorithms are used to take advantage of risk-free arbitrage opportunities in the markets. The second and most common type is the execution algorithm. This is when orders are broken into smaller pieces to minimize the price impact on the market. Often they are designed to seek out liquidity and be routed to the venue that offers best execution. The third type is the statistical algorithm that identifies patterns in the market that will yield investment opportunities.

In the US, execution algorithms are rapidly taking over the use of traditional DMA. Considering the existence of many liquidity pools, smart order routing tools are frequently deployed to obtain best execution. “Hedge funds and proprietary trading desks are continuously seeking new markets in which to apply their arbitrage and statistical algorithms,” says Stellan Råberg, Senior Adviser at OMX. “Currently, hedge funds in the US and Europe are fine-tuning their algorithms to work under different circumstances in several markets in Asia-Pacific.”

There is evidence that algorithmic trading has arrived at Singapore Exchange (SGX). Over the past year, trading volumes in SGX’s domestic equities contract, the MSCI Singapore Index, grew by 81 percent and the average daily value of transactions in the cash equities market doubled. “We know there is some algo trading coming through as traders arbitrage the domestic derivatives contract and components in the underlying cash markets,” says Chew Sutat, Executive Vice President and Head of Development at SGX. “But based on the increase in the number of inquiries we have received from customers, we believe this is just a fraction of the amount of algo trading that could come to our marketplace.”

To entice algorithmic traders, SGX is strengthening its position as an Asian gateway by developing a high quality marketplace that provides access to pan-Asian products and listings. It is also introducing a series of algo-friendly features and services to enhance the market microstructure and support these customers. Equity tick sizes and spreads have been reduced, for instance, and proximity hosting services are being introduced. The exchange is considering new order types, facilitating co-location, DMA and high-speed trading access. It is improving its infrastructure to decrease latency and increase throughput. Systems are being put in place to disseminate real-time market data and supply customers with historical data for back testing purposes. Finally, appropriate fee structures that can bring the benefits of additional liquidity to the entire marketplace from high volume algorithmic trading activities are being assessed.

SGX actively monitors and upgrades its system capacity to ensure it has headroom above all its previous peaks. Like most exchanges, the latency and throughput of SGX’s systems were put to the test during the heavy, volatile trading in the summer of 2007. To this end, the exchange is already in the midst of implementing major infrastructure upgrades. It has a clear technology roadmap to upgrade its data engine, front-end trading capability and bandwidth capacity to ensure throughput and latency can be progressively improved. “We want to stay ahead of the game and make the marketplace in Singapore more attractive for algorithmic traders,” says Chew. “It’s not something that we can do overnight, but we have a roadmap, we have a plan and we’ve already started to roll out some initiatives to capitalize on this phenomenon.”

Similarly, the Australian Securities Exchange (ASX) has taken steps to attract and retain algorithmic trading business by creating a favorable environment for it. In particular, the exchange changed its pricing model so these customers are charged based on the value of the trades, not the number of trades. The plan worked, and ASX has experienced radical spikes in volume driven by algorithmic trading. In October 2006, its peak number of trades per day was about 200,000. By August 2007, that number climbed to 575,000 trades.

By upgrading its technology, ASX has been able to keep up with the demands of heavy trading and stay ahead of the capacity curve. “As a rule of thumb ASX prefers to have 50 percent headroom above our previous peak and the capability to move to 100 percent within three months, so ASX currently has a major capacity program underway with OMX,” says Jeff Olsson, Group Executive, Technology at ASX. “We’ve also reduced the average latency from 80 milliseconds to 20 milliseconds.”

As hedge funds and quant traders look to deploy their models in Asia- Pacific, exchanges are revising their product offerings and pricing as well as upgrading their infrastructure to deliver speed, performance, capacity and fast market data dissemination. Considering the potential for much higher trading volumes, Asia-Pacific exchanges that stay ahead of the curve stand to reap a significant return on their investment in technology.