Despite being implicated in the biggest ever intraday drop in the Dow Jones industrial average, algorithmic trading – together with its faster and considerably more controversial progeny, high-frequency trading (HFT) – is making its way into emerging markets.
Ignoring concerns voiced by Western market participants, regulators and politicians, and unabashed by links to the so-called 'flash crash' in the US last May, exchanges across the globe are now attempting to attract automated trading flows themselves.
Investor desire for exposure to emerging economies is hardly a new phenomenon, and the strong fiscal fundamentals currently underpinning this demand look set to continue. Consequently, as capital flows into the developing world, international traders and asset managers are eager to deploy there the same sort of high-tech strategies they routinely make use of in the US and Europe.
For firms using such methods the appeal is clear. Aside from sluggish growth, US and European markets are rammed full of tough, sophisticated competitors. But exchanges that have embraced electronic trading more recently are likely to offer substantial arbitrage opportunities along with significantly less international rivalry.
It is not just investors who stand to benefit. Many of the larger emerging-market stocks are also listed internationally, easy to access electronically and frequently traded as a result. Domestic exchanges eye these volumes enviously and are keen to see as much of this flow as possible take place locally.
The exchanges have also observed the boost in liquidity, volume and efficiency that automated trading has produced in US and European markets. As a result, many are now upgrading their trading infrastructure and forging links with international brokers and connectivity providers in an attempt to attract new market participants.
These goals may be fairly universal among traders and exchanges, but specific market conditions can vary wildly in regulatory make-up and the infrastructure available for automated trading, says Stephane Loiseau, deputy global head of execution services with Société Générale.
Latin leaders
Mr Loiseau ranks Latin America as a tier above other emerging regions in embracing algorithmic and HFT activity. This is partly due to the close proximity of the US markets, which has produced a heightened awareness of the potential benefits of catering for automated trading and has also led to a number of large US firms becoming increasingly active in the region.
In particular, Brazil’s BM&FBovespa exchange has benefited from a close relationship with Chicago’s CME Group, in which it owns a 5% stake. Indeed, to many, BM&FBovespa is a fully mature market, pioneering access for the international electronic trading community. Nevertheless, many of its top names are still American Depositary Receipts that trade well in New York, and the exchange is keen to bring some of that volume back to Brazil.
“We see a lot of value in our market to having access to these new types of participant,” says Marcio Castro, BM&FBovespa’s chief technology officer. “We believe they will match well with other classes, such as hedge funds or institutional and retail investors.”
Catering for these new arrivals is demanding from a technological perspective, and the Brazilian exchange has been working towards increased capacity and reduced latency for some time. Over the coming year it will, in conjunction with CME Group, develop a super-fast electronic trading platform (based on the Chicago firm’s Globex trading system) for use in its cash equities and derivatives markets.
There is a real appetite from the international community to access these markets, says Jesper Alfredsson, vice-president of product management with trading technology provider Orc Software, which supplies direct market access (DMA) to, and co-location services with, BM&FBovespa. Mr Alfredsson notes an increase in interest from external players, including Chicago-based trading firms, over the past 18 months. Arbitrage spreading and volatility trading strategies have been particularly popular, he says.
Elsewhere in Latin America, Mexico’s Bolsa Mexicana de Valores has offered DMA and direct strategy access for some time. The electronic aspect has been somewhat slow to gain momentum, however, and the regulatory structure has caused issues in creating an efficient market, says Mr Loiseau. "In Mexico, the market structure can be more complicated. There's a combination of voice and electronic trading, and working out pricing can be onerous. There are still a lot of these kinds of issues to work through, which make it more difficult to create sophisticated algorithms to trade efficiently on behalf of our clients.”
Following Brazil's and Mexico’s examples, some of the smaller Latin American countries are beginning to equip themselves to deal with higher frequency flow too, including Colombia’s Bolsa de Valores de Colombia (BVC), which is now able to handle HFT on its equities and derivatives markets.
Uptake is currently hindered by a lack of volume, concedes BVC’s derivatives manager Felipe Trujillo. Nevertheless, he says some trading firms active on the exchange are working on automated strategies. Meanwhile, the exchange itself is attempting to ease legal obstacles for interested parties. "We try to open the door a little bit more for investors so that they might think about implementing basic algorithms in the Colombian market,” he says.
Further momentum should be added by plans to integrate the equity markets of Colombia, Chile and Peru, and Mr Trujillo expects the resultant increases in liquidity should allow foreign participants to successfully implement algorithmic and even HFT strategies.
Market disparities
The fragmented Asian markets have not embraced automated trading quite as readily, and the biggest economies in the continent are by no means the most advanced in terms of market infrastructure. For international firms, electronic trading is close to impossible on the exchanges of behemoths such as China or India without an established local presence to clear and settle domestically, says Frederic Ponzo, managing partner with capital market consultancy GreySpark. Even Japan’s Tokyo Stock Exchange does not yet offer remote membership.
Elsewhere in Asia, markets remain fragmented and isolated, and the large distances involved, together with patchy and in some cases pricy network infrastructure can further hinder international trading, says Mr Ponzo.
However, there is a structural willingness to open up to new electronic trading techniques, and markets in the region are becoming more attractive to algorithmic traders, thanks to narrowing spreads, increased daily activity and reductions in time to tick, says Mr Loiseau. He cautions, however, that regulatory and technological issues mean there is still some way to go before true HFT becomes a reality in Asian markets. This view is shared by Chew Sutat, head of corporate and market strategy at the Singapore Exchange (SGX), who describes pure HFT activity as a “non-starter” in much of Asia because of prohibitive access barriers.
Nonetheless, some the biggest moves in the region have come from SGX itself, which is in the process of implementing a new platform – dubbed Reach – intended to allow extremely low-latency access to the exchange for international traders.
Mr Sutat describes encouraging algorithmic and HFT activity as SGX's responsibility to both customers and issuers, as it should grow liquidity and improve price discovery. The Reach project is the latest in a series of changes pushed through over the past three years aimed at improving access to its market.
These moves have successfully attracted customers from Europe and the US to Singapore, says Mr Sutat, some of whom might be classed as high-frequency or algorithmic traders. SGX’s derivatives market has proven particularly popular, and the percentage of automated trading activity there climbed from 8% to 30% in the three years to 2010. Mr Sutat says the exchange’s co-location facilities already cater for many big-name trading firms, although the only one to have been made public is Getco, which has applied for membership in SGX’s securities market.
When some of Asia’s biggest exchanges have been slow to embrace HFT and low-latency trading activity, the less developed side of the market might be expected to be lagging behind significantly. But progress is being made there. Work is under way to construct a new platform allowing cross-border trading between south-east Asian stock exchanges. Meanwhile, the Stock Exchange of Thailand, which will be part of the project, is in the final stages of selecting a new electronic trading platform.
Similarly, Société Générale has been involved in algorithmic trading on the Indonesia Stock Exchange since the second half of 2010. Such activity was fairly new to the bourse and led to volumes it perhaps did not anticipate, but nevertheless stemmed from “a conscious decision from the market to attract algorithmic flow", says Mr Loiseau.
Little appetite
Brokers currently report comparatively little appetite for algorithmic trading activity in much of eastern Europe. Citigroup’s head of algorithmic products, Raj Nagella, says that of the 32 countries the firm trades in central and emerging Europe, it can only easily employ automated strategies from London via six local and two remote memberships. This is partly a result of liquidity patterns that make electronic projects impractical. “In those markets, it's important to have the right local broker relationships to get information on liquidity, and help execute the big blocks,” he says.
In Russia, however, the soon-to-merge Moscow Interbank Currency Exchange (Micex) and Russian Trading System (RTS) exchanges have been courting international electronic traders. As with Brazil, part of the reason for this is that a number of larger Russian stocks can already be traded electronically on other bourses. “We are doing our best to bring order flow back to Russia and to keep Russian stocks traded here,” says Nikolay Egorov, senior vice-president and member of Micex's executive board.
Sergey Zamolotskikh, RTS’s head of infrastructure, adds that algorithmic traders are among the exchange’s most important category of customers, estimating they make up more than half of its trading flow.
Both Russian exchanges are well positioned from a technological perspective, having invested heavily in the infrastructure needed to handle the large capacity of orders produced by automated trading strategies. And, indeed, international algo-based funds have been testing the waters in local instruments, where there are ample opportunities for arbitrage, says Yury Plechko, director at UniCredit Securities ZAO.
Micex data seems to confirm this; the ratio between orders and trades on the exchange’s securities market has increased from 2:1 just 18 months ago to 10:1 today, indicating dramatic increases in automated order flow, says Mr Egorov. However, he adds that coping with such an influx of electronic trading activity is currently one of Micex's biggest challenges.
For international investors, though, the Russian market structure can be problematic (Mr Loiseau notes that the settlement process is particularly cumbersome). The regulatory situation is also vague and there is currently no specific legislation concerning HFT in Russia. In fact, questions could be raised about the actual legal validity of the practice, says Mr Plechko, warning that a conservative legalistic interpretation of current rules could even class such trades as market manipulation. He adds, however, that regulators are well informed and fully aware of the two exchanges' investments in low-latency trading infrastructure.
Mr Egorov confirms that Russian regulators are keen to address the issue and are working with Micex and RTS to do so, despite some reservations. “HFT is very controversial and we all consider it as such," he says. "On one hand, it brings in the liquidity; on the other, it brings a lot of disturbance and uncertainty on the market.” He says legislation is currently in the draft stage and is expected to surface within the next six months to a year.
Turkey’s Istanbul Stock Exchange (ISE) is also beginning to attract algorithmic trading flows as a result of new trading rules implemented last year that allow intra-session order cancellation – a prerequisite for automated trading. As a result, Crédit Agricole Cheuvreux immediately deployed a range of algorithmic products.
A spokesperson for the Turkish exchange describes these changes as a result of efforts to harmonise market rules with international standards, rather than an attempt to attract automated trading, but he acknowledges the additional benefit of increases in electronic trading. “In spite of the fact that transactions based on advanced technology and speed are rather controversial, they substantially contribute to market liquidity,” the spokesperson says, describing ISE’s approach to algorithmic trading and HFT as “prudently optimistic”. There are currently no specific regulations in place in the Turkish market relating to such activity.
Regulatory barriers
The legal and regulatory situation in Russia and Turkey may be somewhat vague then, but financial watchdogs seem to be generally positive. Elsewhere, efforts to facilitate international electronic trading could easily be blocked by strict legislation, stemming from fears of the dangers of algorithmic trading or protectionist tendencies.
The exchanges themselves tend to be rather bullish about the potential risks. SGX’s Mr Sutat dismisses some of the more common concerns about HFT – its speed, absence of direct human involvement and arguable lack of contribution to the capital formation process – as being founded “more on fear than on fact”. BM&FBovespa’s Mr Castro even claims that many emerging markets are structurally less susceptible to the potential downsides of HFT by virtue of their unified nature. “In Brazil, for example, we have the benefit of being much less fragmented and a lot of the things that have happened in the West are the result of fragmentation, which is causing more damage than benefit at this stage,” he says.
In any case, most exchanges claim to have safety measures in place that will avoid the dangers posed by rogue algorithms and the like. SGX, for example, is set to introduce a 'circuit breaker' to its securities market, which should kick in if the market falls precipitously. Meanwhile, BM&FBovespa boasts a throttling mechanism that prevents the system from being overwhelmed by limiting the number of messages per second that a particular client can send to the exchange. Pre-trade credit controls and the option for broker dealers to set limits for each and every client are also popular. Micex, on the other hand, has no specific circuit breaker set-up relating to HFT, but steep falls in the Russian market already trigger suspensions in trading.
The spectre of protectionism is not so easy to exorcise. In recent months, a number of smaller exchanges have introduced market structure changes seemingly designed to cater for electronic trading. But this will not necessarily open up the bourses to international traders, says Citigroup’s Mr Nagella, warning that the welfare of domestic exchange members may take precedence in some regions. “When you start looking at these smaller countries there’s a desire to keep up with the neighbours in some sense, but there’s still a big issue of trying to protect the local brokers.”
The Middle East, for example, has seen significant technological investment, but international brokers and traders still complain that the regulatory environment makes markets difficult to trade in.
Tight capital controls in many countries are a further complication. Introducing high-speed electronic trading to international participants will probably lead to headaches for regulators concerned about the speed at which money flows in and out of its borders. Indeed, opening up a market to algorithmic trading requires a comprehensive effort from many parties, says Mr Najella. “It’s not necessarily the electronic technology that is a determinant of success; it’s more a symptom of the government and regulatory bodies deciding it is something they want to invest in.”
Lacking liquidity
Even if the political environment is more welcoming, an influx of trades is not guaranteed. The larger Latin American markets, as well as some in Asia, have been able to attract algorithmic trading flows with ease because they already boasted fairly sizeable levels of liquidity and volume in the market.
However, algorithmic trading does not work particularly well in markets with relatively small levels of notional value, flow and liquidity. And even when it is attempted, prospective traders cannot always import strategies they use elsewhere. Instead, algorithms must be adapted to account for local idiosyncrasies. Attempting to force-fit electronic operations onto a comparatively shallow market, concentrated on a handful of infrequently traded stocks, may represent a rather unappealing investment.
“There’s a first step where the market can’t be too thin to attract that liquidity. While the appeal for the exchanges of attracting algorithmic and HFT is to increase liquidity in the first place, you can’t start from zero,” says Société Générale’s Mr Loiseau. For some markets, he says, building up trading activity on the traditional side of the business and making structural improvements to the market is necessary before more advanced electronic trading can take place.
Appetite exists all the same in emerging markets, and many traders in search of fresh money-making opportunities will be willing to make that leap. The exchanges that have reached this critical mass – and that also benefit from a more enlightened regulatory regime – may be few in number but, for them and their home countries, the potential rewards are huge and present the opportunity of becoming a genuine regional hub and global financial centre.
