Man’s $21bn quant fund sees no need for high frequency trading

AHL, the managed futures single manager of Man Investments, says it sees no need to participate in the high-frequency area to secure alpha from its trend-following strategy.

Quantitative investment manager AHL says while it monitors high-frequency data and takes prices 4,000 times from the 36 countries that its global $21.2 billion managed futures portfolio trades every day, it sees no need to reduce execution speed down to a nanosecond to achieve its target return.

Keith Balmer, London-based portfolio manager for the managed futures unit of hedge fund Man Investments, says AHL's systematic trading approach, which predicates on algorithms and statistical models to cut out human discretionary intervention, often completes executions in a matter of seconds. Unlike high-frequency traders, whose alpha can degrade in nanoseconds, AHL's trend-following approach means its alpha degrades more slowly, with the result that it can sometimes afford to trade slower.

AHL's investment process is built around the belief that markets frequently demonstrate inefficiencies that can be predicted through careful statistical analysis.

High-frequency trading is a subset of algorithmic trading, which is estimated today to account for as much as 60% of the daily equity trading volume in the US. Its use aims to exploit arbitrage trading opportunities that stem from fragmentation of liquidity by generating large amounts of orders and liquidating these exposures as quickly as possible.

High-frequency trading has drawn increased regulatory scrutiny in the US and elsewhere after the 1,000-point crash in the Dow Jones Industrial Average on May 6. The so-called ‘flash crash' led to the Securities and Exchange Commission and the Commodity Futures Trading Commission considering the need for stricter oversight regarding the proliferation of trading algorithms widely considered to have exacerbated the crash.

Trend-following is historically a core strategy for managed futures funds such as AHL. While the proliferation of high-frequency trading in Europe, the US and increasingly in Asia means these traders could be picking up on trends observed by the minutes, Balmer says in AHL's case trend-following takes "one to two days on the fast end".

"One method of trading is on a fully automated basis, meaning we are looking at high-frequency data that gives us a lot of information and insights on how order books are evolving and how prices are moving," says Balmer. "Our algorithms would adjust how they put orders into the marketplace based on what they are seeing in that environment."

Research on how to maximise alpha generation while minimising market impact and execution costs done by Man Investments in collaboration with academics from the University of Oxford, where Man runs a research institute of quantitative finance for confidential AHL projects, found that "occasionally, it is better to slow down your execution than to speed it up", Balmer says.

"Sometimes our execution takes longer now than it used to, not because we do not have the capability to execute fast, but because now we have a more sophisticated way of placing orders into the marketplace to get a better price and therefore our algorithms adapt to different market conditions," he says.

Employing globally a total of 77 quantitative researchers, focusing chiefly on developing new trading models and methods of constructing portfolios, AHL in April last year opened a trading desk in Hong Kong to establish direct market access to Asian exchanges to facilitate faster connectivity with them and improve overall execution.

Balmer says, however, that most of the group's servers are located in London. This differs from some high-frequency trading specialist firms, which tend to rent space to place their own software and servers in closed proximity to the exchanges' trading engine to reduce latency.

For the three months to June this year, AHL Diversified, the single manager's key investment programme, returned a modest 0.9% against a 11.6% drop in the MSCI World Index price return hedged to US dollar.

Balmer's comments come amid rule changes in markets such as Australia and Japan that appear to be increasingly welcome to competition for equity trading. Automated shares trading platform Chi-X – Europe's second-largest exchange by turnover this year – has also started operations in Japan after the Tokyo Stock Exchange upgraded its trading system following the launch of its ¥13 billion ($153.6 million) Arrowhead technology systems in January. The arrival of alternative exchanges in Asia has in turn drawn more high-frequency players to look at opportunities to trade Asian markets within Asian hours.

Besides from displayed markets such as Chi-X, investment banks are also launching their own internal dark pools in Asia as part of their algorithmic trading offering to clients. For example, this week Deutsche Bank says it has launched its internal dark pool called Deutsche Bank automated trading system, together with its algorithm that routes clients' orders to other dark pools outside its own system.

A full article on the risks related with algorithmic trading and how dealers and investors are coping with them will be published in the September issue of the Asia Risk.