Capital markets firms are facing a data explosion that requires them to change the way they plan and implement their technology. Algorithmic trading, which accounts for just 10 percent of current order flow, has driven up the volume of pricing data faster than analysts can track it. The TABB Group estimates algorithmic trading will account for 60 to 70 percent of total flow within a few years, and cites estimates that firms may have to accommodate 130,000 messages per second. The speed of trading and the growth of cross-asset class trading are placing a stress on reference data

Robert Iati, a partner at the TABB Group, said these changes will pose a challenge for firms which have generously funded their trading desks while regarding reference data as a cost center that should be kept on a subsistence diet.

“As each marketplace steadily integrates electronic trading platforms into its core, trading strategies are advancing across asset classes in search of ways to exploit opportunities outside their traditional markets,” writes Iati. “This will tie together traditionally disparate products, facilitate cross-product hedging and accelerate the need for consolidated risk management and an integrated view of the customer, the trading desk and the organization.”

In a business that has become so dependent on programs, algorithms and other black box mechanisms to achieve best execution, explains Iati, “the pressure to gain the slightest edge in the marketplace has driven firms to adopt a laser-like focus on data.” (for the full report see http://www.vhayu.com/briefTABB.html)

The financial services mentality has been: “This is the technology infrastructure we have now, how do we work with it?” Instead, it should be: “Where do we want to go, and what do we need to get there.”

Leading firms are examining their architecture for ways to reduce latency. Increasingly firms are storing huge volumes of data in memory so analytical engines such as Vhayu can process it fast enough to exploit market opportunities. They are also reviewing their physical architecture – can they reduce latency by moving their trading desks closer to the exchanges?

As more trading moves to high speed electronic orders, the trader’s role is evolving. Algorithmic trading is too fast for human traders, who conduct their trades by writing algorithms that they test, fine-tune, and speed up or slow down during the trading day. Changing the algorithm during the day places huge demands on the trading infrastructure and on the reference data.

That said, I think the individual trader will continue to play a key role in certain areas. New tools are emerging, such as the visualization provided by Oculus, coupled with an extremely fast analytical engine like Vhayu, which can handle 350,000 updates a second now and projects capacity of 1 million updates per second with 64-bit technology. Combining high speed analytics and visualization and Tablet PCs, could offer the potential for an unorthodox trader to find new ways to view the markets and uncover opportunities that are hidden by other approaches.

Some trading will continue to rely on personal contacts, which can be enhanced with instant messaging technology that shows who is present on the trading desks of your own firm or on the desks of trusted counterparties. If I am doing a large proprietary trade, I might want to bounce my ideas off a colleague. Or, if I am a fixed income trader who needs an FX trade done, I might hand it off to a friend at an FX specialist, and he could reciprocate by sending me a bond transaction sometime in the future.

This is a race that will just continue to escalate because the more algorithmic trading that’s done, the more every other market participant will be forced into working at high speeds and with huge volumes of market data to find profitable opportunities.

The proposed merger of the New York Stock Exchange and Archipelago is the latest example of how electronic trading is impacting the industry. New York needed a new way to trade, and Archipelago had it.

Prime brokers may see their role expand into providing even more services to trading firms by buying software companies – Citigroup and Lava, Bank of New York and Netik, or building software that they then sell, such as Citi’s Best Execution Consulting Services (BECS).

As regulators and shareholders push for better risk management and compliance, banks will have to buy or build the tools to help their clients meet the new demands. Some of their most profitable clients, hedge funds, maintain their edge through a tight focus on trading while outsourcing the utilities that are vital to their business but don’t provide any competitive edge. For prime brokers, technology is shifting from a support function to a core competency required to maintain and grow their business.

The complex life of a trade (illustrated on facing page) shows how the business runs today, but not where it could be in five years.

Can peer-to-peer trading replace exchanges, as John Gapper recently suggested in the Financial Times? Will prime brokers cross orders internally and reduce the complexity of current settlement practices?

Over the last five years, firms have driven out the inefficiencies on the trading desk. When we are now looking at pennies and milliseconds it is clear that we can’t achieve much more there. Now it is time to turn attention to the back-office processes and the complexity of exchanges and custodians for the clearing and settlement processes. Much of the post-trade complexity adds weight rather than value. I expect that investment banks will also re-examine their structures. How many are still running different silos for retail, private banking, and institutional business? And how many have different architectures, and different applications, in London, New York, and Tokyo to support the same business functionality?

At Microsoft we have focused on experience in finance – the experience of the customer, the front-office professional and the operations staff. Microsoft-based applications are widely used to provide the tools for high-speed analytics and trading from the traders’ desktops to the high-performance Microsoft SQL Server 2005 database whose beta version supports Townsend Analytics. Just about every finance professional in the world uses Excel for a view of business, and often reports results either internally or to investors through a combination of Excel and Word that is easily consumed by business users who might not be familiar with all the technology that made the results possible. On the operations side of the business, advanced users such as Stephens, Inc. and Raymond James are finding that a modern back-office securities system from CSS provides powerful real-time reporting capabilities that delights their customers while allowing operations staff to get their work done faster and go home earlier after trading ends.

For more information about experience³Capital Markets, see www.microsoft.com/experiencecapmarkets

The Next Frontier: Fixed Income

The most interesting developments in equities trading are behind us, according to Till Guldimann, vice chairman at SunGard. Now the action in the United States is moving into the fixed income markets.

Here dealers with significant inventories and strong client relationships are fighting the electronification of fixed income trading.

“If you talk to the firms that are strong in fixed income, they say it will not happen, but if you talk to the strong equities firms, they will say it definitely is coming. Companies are defending their turf.”

The second major trend Guldimann sees in the US is improved simultaneous access for traders to the cash and listed derivatives markets.

“It is quite different from Europe where you had exchanges doing cash and listed derivatives on the same venue.” In the US, regulation separated the markets a long time ago and technology is starting to bridge the gap.

Third on his list of top trends is the strong growth in alternative trading networks for the buy-side, like LiquidNet whose recent round of funding valued it at $1.8 billion, only somewhat less than the New York Stock Exchange at the time.

“This is amazing – a buy-side venue that didn’t exist a few years ago and now has a value similar to the NYSE. But look at the market cap of eBay, which is $50 billion. The financial institutions missed the boat when it comes to inventing the new electronic markets,” Guldimann said.

Look to Hoboken

Great ideas in the financial markets are more apt to come from Hoboken or Jersey City than from Wall Street, he added. “You almost always see real innovation coming from the fringes rather than the center because the center gets so puffed up,” he said. Innovations often have unintended consequences.

One result of direct market access (DMA) was algorithmic trading because once you have direct electronic access, the response time was so fast you needed to make electronic decisions about how to route and how to trade, and that was algorithmic trading. Now everyone has these trading algorithms on top of DMA.

“The next game is total cost analysis. Every broker will now tell his institutional clients that his trading programs, his algorithms, are the best. And how will he prove that? He will send his client a detailed analysis of every single trade with time stamps and price drifts. An order for 100,000 shares came in at this time, when the market was this or that and we executed at these prices over the next three minutes,” Guldimann said. “This kind of total cost analysis is developed by each broker and each broker sends these reports to the clients in an effort to prove he is the best. The client gets multiples of these reports, and each is different so the performances are not easy to compare.”

“Will a third party jump in to provide an independent evaluation of brokers’ execution? That is where the movement is today, but I can’t tell you what the outcome will be,” he said.

An eBay-Style Trading Community Grows in Canada

While many predictions for trading simply extend existing practices with the addition of speed, complexity, and lots more market data, a group of seasoned electronic trading professionals in Toronto is creating a system on which institutions and brokers can trade large blocks of stock without moving the price. The system uses a community-based technology approach for counterparties to cooperate anonymously. A computer matching system then executes the trades when an agreed upon price is reached.

The group, led by Doug Steiner, CEO of Toronto-based Perimeter Financial Corp., has built and deployed several electronic trading systems. Steiner describes the company’s newest system, called BlockBook, as having unique characteristics such as “symmetrical price discovery” where buyers and sellers try to find each other in a tit-for-tat fashion – while a computer polices pricing and information delivery to counterparties on each order.

The system allows everyone to see when interest is developing in a particular stock, when prices between buyer and seller get within trading range, when negotiations commence and when a sales takes place. Each negotiation has public and private information. The more serious you are about the potential to trade, the more information and feedback you receive from your actions and others. When the trade is completed, the price and size of the trade is disclosed without identifying the participants.

Markets Inc, the company that built BlockBook, is backed financially by the largest institutional funds in Canada, a select group of forward thinking investment dealers.

“We felt that since the buy-side is so concentrated in Canada, they could be very influential in how they wanted this market to unfold,” Steiner said. The system was built with Microsoft .NET and uses the FIX standard for messaging.

If this sounds familiar, it could be in its similarity to LiquidNet, the system developed by Seth Merrin for institutional trading. The difference, said Steiner, is that BlockBook is multi-lateral while LiquidNet is bilateral, and BlockBook includes all capital market participants. “This market is too small to exclude pools of capital from a trade,” said Steiner’s whose group worked for three years on the infrastructure. For example, on a 200,000-share sale from a fund, the system could provide a single buyer, or 10 buyers of 20,000 shares each.

BlockBook is launching with 30 buy-side and sell-side participants, including several of the largest funds in Canada, such as the Ontario Teachers’ Pension Plan and BC Investment Management Corp.

Dealers are participating in the system because it offers the potential for more trades.

Steiner’s favorite model for trading is eBay. BlockBook won’t tolerate users who try to game the system and has included a method for expelling traders who misbehave.

“You have to respect the amount of trust between the buy-side and the sell-side. To do that, we think you have to have people involved in a community,” Steiner said.

If all this sounds very Canadian, and a bit naïve, the concept is backed up by rules and technology.

Participants have to place a real order into the network before they can participate in price discovery. And anyone trying to game the system will be ranked, and if the community agrees, thrown off, just like eBay. His community policing approach knocks out dishonesty and gaming, and allows firms to build cost-effective trading systems that aren’t using up a large part of their capacity processing prices that will never become orders.

“Instead of 50 tiny algorthmic orders, we want five huge orders that are put into a system that polices people and allows them to trust each other. I think the trend toward algorithmic trading and slicing is there only because of the trust factor, and in the US the central markets are liquid enough so people can find they are being gamed. We build systems where the price doesn’t move away when a large order is placed. If it does, there is a very severe cost to the person who is gaming.”