When Kent Rossiter joined Allianz Global Investors in 1999, he was the new guy. Despite the passage of time, that is a status he still holds among the company’s senior traders, due to the stability of the team which has not changed too much in the years since. The same cannot always be said for the structure of the modern trading desk, however.
Originally, the company operated six different trading desks in Asia; however, these have long since been merged. It’s a detail that hints at the increasing focus on efficiency, as trading desks become more proficient, more asset classes are added and more local markets are connected up. At the same time, the availability of liquidity has combined with pressure on costs to influence the way the trading desk operates. In Asia, the availability of liquidity is an important factor.
“You can’t always trade when you want,” Rossiter told the Buy-side Perspectives. “That means being patient at times and very aggressive at other times. If the market’s not right, you have to be willing to step out for a few days.”
Globally, liquidity has been an issue of debate for some time. While this might seem obvious, it’s not so much that liquidity was less important in the past, but more that it may be getting harder to access in the modern trading environment. Changes such as the pullback of the sell-side under the onslaught of global regulation have played a major part, but so too have technological changes and the fragmentation brought about by new platforms.
Numerous initiatives have sprung up over the years in an attempt to aggregate the liquidity electronically; however, the successes of these platforms have been limited, and many have fallen by the wayside. Even in equities, platforms such as Chi-East failed to achieve critical mass. Partly as a result, trading desks in Asia, like the rest of the world, are gradually becoming more multi-asset – leading to debate over the precise role of the individual trader and whether or not individual traders should be trading across multiple asset classes themselves.
“Ideally we would have enough traders to handle every niche product and asset class,” said Rossiter. “But in reality the liquidity is unpredictable and in certain asset types, there are periods when there is not much trading activity. We can’t necessarily afford to have specialized traders not doing anything for three days. You can’t run an efficient desk that way. We are a multi-asset operation and that includes CDS, fixed income, etc. One of the challenges is we have a lot of PMs that want us to trade different products. The bottleneck is not just trading, but legal and compliance, getting set up, making sure operationally that you can settle the trades, and of course monitoring the risk. There are many boxes that need to be checked.”
Allianz Global Investors’ trading desk in Hong Kong is mostly split by country. According to Rossiter, this is because the similarities and practices are “probably more important” than doing so by sectors as may be the case in the US for example. As for trading other asset classes, this is split between the traders because “practically speaking, we may be very busy in bond trading one day, and super quiet the next. And the same goes for equities, etc. We just need the bandwidth to be flexible enough to handle the uneven nature of how orders hit our desk.”
While the buy-side is gradually embracing a more multi-asset mode of doing business, there are signs that this shift has not always been matched by brokers. “Between asset classes some of the senior managers at the brokers know our firm holistically, but down on the desk the equity trader and fixed income trader at one of the banks may not even know each other and they really are more focused on their own asset classes, and not how Allianz Global Investors may use their firm’s offerings in other areas,” said Rossiter.
When it comes to selecting the trading venue for equities, there are rarely multiple venues to choose from in Asia, since most countries in the Asia Pacific region are still primarily focused around the national exchange. However, some markets such as Japan do feature multiple venues, and in such cases, Rossiter’s trading desk has to pay attention to which venue it is working orders on. To complicate matters, the presence or otherwise of high-frequency traders can be a significant factor in the decision of where and how to trade.
One method to mitigate the risk of being picked off is to use minimum fill sizes for dark fills or some of the PTS. But in reality, Rossiter warns that the trader could be interacting with HFT all day long even on the primary exchange.
“There’s a mentality in Asia to trade at one-third volume by default,” he said. “Trouble is, there are many savvy HFT programs out there that can take advantage of our signals and disadvantage us. It’s hard to pick a standard volume. The fact that HFTs are not present when I’m matching blocks is a great reason to do that, but realistically a lot of the order sizes are smaller. If I have to work in the market, I’d rather be slow. If you’re running at 30% volume you will have market impact but if you’re at 5% at least you won’t make that impact. Slow down where you can. In those cases where I must be trading in the market I try to at least be less predictable, taking breaks here and there, and bold moves between.”
That’s not to say there is no place whatsoever for alternative trading methods in the market, however. Under the right conditions, in limited circumstances, then there can be an opportunity to trade without too much risk. “With a smaller sized order where you really don’t have impact then there’s little chance an HFT client could game you, and in that case if a PTS has a better price then why not take it? We’re not against dark pool trading with our brokers but do have some settings in place which we hope help us avoid interacting with some of the more toxic flow found in such pools,” Rossiter added.
Broker lists are another factor that have experienced pressure in recent times in some parts of the world. At present, Rossiter estimates that nearly 70% of Allianz Global Investors’ trades in Asia can be executed with the firm’s top 10 brokers; however, while there is significant concentration of flows, the tail of the broker list is much longer. It includes brokers that do not have the scale of the bulge-bracket firms, as well as niche brokers that may only have operations in a few specific markets. Nevertheless, some of these brokers may have a dominant market share in a particular country or market, and thus the buy-side needs to have trading lines set up so that it can take advantage of liquidity and gain information on mid- and small-cap stocks.
“You could say fragmentation is a problem for Asia, which of course it is if you’re not patient,” added Rossiter. “But in most cases we can get our trades done, just that they may take more than a day or two. On real tricky trades we usually try to find brokers who know the company well, have done roadshows or presentations lately, or published research. We’ll also look at which corporate broking relationships they may have, or who did their IPO if not that long ago. We’ll look out for top shareholders and their publically disclosed stake changes to see if we could identify a natural contra, but if not we also consider facilitation at times. All this means our broker list has to be decently wide and provide flexibility to trade.”
From a trading technology perspective, the relationship between the buy-side and the sell-side is currently inextricably linked to the dilemma between high-touch and low-touch trading methods. Above all, the use of algorithms is arguably linked to the level of commission fees paid by the buy-side to the sell-side. For some buy-siders, this can lead to some surprising conclusions – including a will from the buy-side to resist further erosion of broker commission fees because of a feeling that further use of algos would undermine the other services provided by the sell-side.
“I am a member of a discussion group and last week we were speaking in detail about algorithmic trading,” said Rossiter. We found that there may be 100 algos out there but human nature means the top three get used, i.e. in any given year a random trader really only focuses on a few. They may tinker with the volume parameters, but they’re not really using a different algo. If anything, the proportion of algo trading we have may be too much. The commission pool that we pay our brokers has been shrinking year-on-year for years. As I recall, 2010 was the last time we had an uptick. A number of years have seen double-digit drops. Meanwhile, our AUM has not decreased so it’s basically a lower blended rate, and perhaps a little less active trading. I don’t want the service from the Street to fall as a result of lower commissions being paid, and I don’t think we’ll be using more algos.”
The Hong Kong Way
Kent Rossiter, Allianz Global Investors
"You could say fragmentation is a problem for Asia, which of course it is if you’re not patient..."
"If the market’s not right, you have to be willing to step out for a few days."
Head of Global Buy-side Research