Will anonymous trading venues have ripple effects? BY SHERREE DECOVNY

Size: px

Start display at page:

Download "Will anonymous trading venues have ripple effects? BY SHERREE DECOVNY"

Erin Reed
6 years ago
Views:

1 Will anonymous trading venues have ripple effects? BY SHERREE DECOVNY 28 C F A M A G A Z I N E / J U L Y - A U G U S T

2 It seems to be a fundamental paradox of the modern investment management industry: While the amount of assets under management is growing, the average execution size is declining. On closer examination, however, the inverse relationship between these two trends makes perfect sense. The change is partly a function of scale. Large positions challenge asset managers to find ways to move large blocks of stock efficiently. The potential for information leakage exacerbates the problem. When you give an order to a broker and you hang up the phone, you ve lost all control over where that information goes, says Seth Merrin, CEO of Liquidnet in New York City, an electronic marketplace that facilitates institutional equities trading for asset management firms worldwide. The longer an order is outstanding in the market, the more leakage there is and the more the market moves against them. In short, he says, The problem is about capacity. But this is more than a numbers game. Further complicating matters is the fact that asset managers have a fiduciary duty ( and in many jurisdictions, a regulatory obligation) to deliver best execution for their customers. The need to avoid such undesirable effects has led to the development of dark pools, or anonymous matching venues in which quotes or other indicators of liquidity are not displayed. In the United States alone, there are now about 50 such venues, and 15 percent of buy-side firms trading volume is executed through them, according to estimates provided by Tabb Group, a research firm based in Westborough, MA. Although the term dark pool sounds mysterious, these venues actually reproduce a very human function. Historically, buyers and sellers on the floors of the exchanges had only a few people with whom they could trade. After years of being on the job, they learned how much information to give to whom and when. Brokers would differentiate themselves by specific skill sets or even by their personalities. In effect, dark pools simply mimic traditional trading practices and translate them to cyberspace. Indeed, despite the suggestion of some doubters that dark pools distort or destabilize markets, proponents of dark pools contend that these venues enhance overall market structure. M A N Y S H A D E S O F DA R K N E S S Dark pools come in three basic structures: scheduled crossing networks, continuous blind crossing networks, and indicated markets. In scheduled crossing networks (such as BIDS Trading, POSIT, NASDAQ, the NYSE s MatchPoint, and Instinet), the two sides of a trade cross during a set period. These networks typically do not display quotes but may have an order-imbalance indicator. Continuous blind crossing networks use limit-order books for which no quotes are given. (Internalization engines, which many large brokers have, are a subcategory of continuous blind crossing networks.) Indicated markets allow participants to post an indication of interest (IOI) and provide some level of transparency in order to attract liquidity. Liquidnet, Pipeline, ITG, Merrill Lynch, and NYFIX offer variations on this theme. Dark pools use three main pricing mechanisms: automatic, derived, and negotiated. With automatic pricing, the two parties agree to trade at the midpoint between the bid and ask prices. Derived pricing is based on the bid price, ask price, or another benchmark, such as the average of the trades for the past five minutes. In negotiated pricing, the price is determined by interaction between the two parties. Dark pools compete on the basis of their ability to attract liquidity and their distinctive ways of enabling users to interact with potential trading partners. They may also provide other services. BIDS TRADING Created as an open, flexible, low-cost, and efficient utility owned by investment banks and geared to meet the needs of buy-side traders, BIDS Trading also provides such tools as scorecards and filters that allow traders to manage information disclosure on a pretrade basis in order to obtain best execution. Buy- and sell-side firms can access BIDS Trading via an algorithmic- or program-trading platform. Algorithmic trading involves the electronic execution of traditional strategies driven by the buy side, such as implementation-shortfall and participation strategies. In program trading, an automated program is used with electronic market making on the electronic communication (or crossing) networks (ECNs) to trade equities. Orders can be automatched or sent in for negotiation and can even be submitted as firm or conditional, which allows traders to play on multiple venues simultaneously. For example, suppose a trader wants to buy a million shares of a certain stock. He is not willing to commit to posting an order on BIDS Trading alone, but he would like to be advised of an opportunity to do the trade. The trader can invite a conditional order to sell, say, 1 million shares at $0.10. To avoid leaking information to the wrong counterparties, BIDS Trading allows users to discriminate on past behavior through its system of scorecards and filters, which extends to the negotiation and conditional-order process. Each person in BIDS has a unique identifier, and we re going to track every time you come into a conditional order, how often you C F A M A G A Z I N E / J U L Y - A U G U S T

We re going to track every time you come into a conditional order, how often you convert from a conditional order into a firm order. That s an indication of what you might do in the future.

6 billion shares have been traded on BIDS Trading since it was launched in spring 2007. In the first quarter of 2008, BIDS Trading averaged 21 million shares a day, up from 18.

3 We re going to track every time you come into a conditional order, how often you convert from a conditional order into a firm order. That s an indication of what you might do in the future. TIM MAHONEY BIDS Trading convert from a conditional order into a firm order, explains Tim Mahoney, CEO of BIDS Trading. That s an indication of what you might do in the future. More than 3.6 billion shares have been traded on BIDS Trading since it was launched in spring In the first quarter of 2008, BIDS Trading averaged 21 million shares a day, up from 18.5 million shares a day in the fourth quarter of LIQUIDNET The eighth largest agency broker in the United States, the third largest in the United Kingdom, and one of the largest brokers in Europe, the Liquidnet block-trading venue differentiates itself by matching size with size and completely eliminating information leaks. When a match is found, only two people in the world know about that trade. Liquidnet s system will send an instant message or IOI to the trader if there s potential to negotiate and will perhaps execute the transaction at the midpoint between the bid and the ask prices. Liquidnet s traditional model caters to buy-side flow, but it has another model, called H2O, that attracts smaller orders through real-time streaming liquidity partners (SLPs), including aggregators of order flow en route to ECNs and exchanges. SLP order flow interacts with Liquidnet H2O when an order is created by a Liquidnet member. SLP orders that interact with Liquidnet H2O are agency orders that are electronically routed by a smart router to an execution venue with no direction or knowledge from human traders. Average execution size on Liquidnet has risen from year to year as traders have become more comfortable using the venue specifically to execute large block trades. When you sign onto ebay for the first time, chances are your first purchase is not going to be a car, says Seth Merrin. We have the same thing here. As people get more and more comfortable using us, they trade larger and larger sizes. In the United States, the average execution size through Liquidnet has increased from about 47,000 shares in 2006 to 56,000 shares in 2008 and transaction volumes have increased by an average of 44 percent each year for the past three years. Liquidnet has expanded to 32 markets, including Japan, Hong Kong, Singapore, South Korea, and Australia. It now has a global liquidity pool of 6 billion shares. NYFIX MILLENNIUM This venue differentiates itself by giving users the opportunity to achieve price or liquidity improvements while mitigating negative price impact for institutionalsize trades. Linked to the primary markets and regional exchanges, Millennium leverages the NYFIX network and its real-time matching algorithm to execute trades at or better than the National Best Bid or Offer, or NBBO, price. (Exchanges and market centers registered with the U.S. SEC send their trades and quotes to a central consolidator. Brokers are required to guarantee customers the best available ask price when they buy securities and the best available bid price when they sell securities as per the Consolidated Tape.) Brokers, buy-side firms, hedge funds, and retail customers can post both pass-through and conditional orders, and they can generate anonymous liquidity alerts to external dark pools and other passive liquidity sources. Conditionalorder functionality allows users to specify such parameters as minimum or maximum execution per tick, percentage of the displayed size, and minimum first-execution size. Users can enter both market and limit orders; peg orders to the bid, offer, midpoint, or last sale; and tie orders to the primary or consolidated market. NYFIX Millennium also offers a commission management program in connection with all NYFIX agency brokerage services, including direct market access (DMA) and algorithmic trading platforms. Clients can use NYFIX s neutral and independent trade-execution services without having to sacrifice existing relationships with trusted research providers. They can manage their commission allocations online by using a range of commission management tools including reporting, payments, and regulatory guidance which has the potential of significantly increasing clients work-flow efficiency. NYFIX Millennium set a record volume in January 2008 by matching more than 1.2 billion shares (or 58 million shares per day). Volume for 2007 was up 100 percent from In We are actively increasing our buyside interaction at the same time as continuing to grow our sell-side base. BRIAN CARR NYFIX 30 C F A M A G A Z I N E J U L Y - A U G U S T

Given that you can execute against our internal order flow or against IOIs in milliseconds, it s a viable destination to look for liquidity while you re on your way to the open market.

We are actively increasing our buy-side interaction at the same time as continuing to grow our sell-side base, says Brian Carr, CEO of NYFIX.

4 Given that you can execute against our internal order flow or against IOIs in milliseconds, it s a viable destination to look for liquidity while you re on your way to the open market. CRAIG VIANI VortEx addition, NYFIX launched its European platform, Euro Millennium, in March We are actively increasing our buy-side interaction at the same time as continuing to grow our sell-side base, says Brian Carr, CEO of NYFIX. VORTEX Both buy- and sell-side firms route their orders and trade through BNY ConvergEx Group s VortEx. VortEx combines conflict-free agency order flow (execution-only services offered by BNY ConvergEx) with qualified dark external liquidity (orders from external firms that are ready to trade). The two combined creates one of the deepest and most diverse high-speed dark pools on the Street, says Craig Viani, managing director and electronic trading product manager for VortEx. Buy-side firms may route orders to VortEx directly or indirectly via DMA products and ConvergEx s internal blocktrading, program-trading, and transition management desks. They also route orders through algorithms. A frequently used algorithm is intended to minimize slippage against volumeweighted average price by targeting the stock s expected volume profile within the user-specified time frame. As orders come into VortEx, they interact either with the internal liquidity or with other dark pools that send IOIs. VortEx aggregates the IOIs and trades against them. In this way, VortEx minimizes information leaks because it does not have to search for a match. The order size varies, but the average execution size is shares. Institutional customers tend to route larger block orders in and rest them in there for an extended period of time until a match for their order is found, says Viani. Wholesale customers tend to route smaller orders of shorter duration, given that they re coming off their algorithmic order flow or from their DMA flow. VortEx s volume grew significantly after it launched its resting order type, which provides access to hidden liquidity throughout the trading day, thereby increasing the probability of execution. It is designed to give clients additional options for executing trades in VortEx. In effect, the resting order stays in VortEx until it finds a match. VortEx s inflows grew percent from the first quarter to the second quarter of 2008 after the launch of this service, and executions are growing at about 15 percent a month. In total, million shares are executed in VortEx daily. Viani notes that VortEx benefits both institutions and broker/dealers who are seeking liquidity across the full range of venues, including exchanges, ECNs, and dark pools. Given that you can execute against our internal order flow or against IOIs in milliseconds, it s a viable destination to look for liquidity while you re on your way to the open market, he says. B R OA D E R I M P L I C AT I O N S The consensus view seems to be that the combination of dark pools and exchanges has improved the market structure. Although exchanges are excellent price-discovery mechanisms, they are not so good at quantity discovery. Thus, dark pools have increased competition and improved efficiency for institutional traders. Without dark pools, institutions would have to go outside the spread every time they wanted to trade a block. Moreover, dark pools have helped drive down total transaction costs from around 140 bps to about 40 bps, according to BIDS Trading s Tim Mahoney. He also notes, however, that the ability of traders to analyze post-trade data on how the order was executed has also contributed to the decline in transaction costs. As new analytical tools become available to traders and their skills improve, costs will go down even further. Still, some disadvantages come along with the benefits. The number one complaint is that the hit rate, or the matched liquidity that they re able to find, is sometimes very low, says Matt Simon, a research analyst at Tabb Group. Because of that, it can make their job a little more difficult. While they re doing all that, the price of the stock could be moving in a direction that s unfavorable to them. The number one complaint [from traders] is that the hit rate, or the matched liquidity that they re able to find, is sometimes very low. MATT SIMON Tabb Group C F A M A G A Z I N E J U L Y - A U G U S T

Orders are being partially filled in one venue and then have to be moved elsewhere: That s a lot of lost time, which does have an impact on total transaction costs.

5 Orders are being partially filled in one venue and then have to be moved elsewhere: That s a lot of lost time, which does have an impact on total transaction costs. ARY KHATCHIKIAN Portware Ary Khatchikian, president and chief technology officer of Portware, a New York City based provider of global, multiasset trading solutions, agrees. He says orders are being partially filled in one venue and then have to be moved elsewhere in what amounts to constant and conditional order routing. That s a lot of lost time, which does have an impact on total transaction costs, says Khatchikian. In addition, connecting to every dark pool simply is not feasible for most firms. Although connectivity has become less expensive over the years, Tabb Group estimates that it could cost about US$1 million for continuous access to every dark pool. Most buy-side firms connect to the dark pools via their order management systems, which aggregate all their orders and feed the data into the portfolio accounting systems to support back-end processes, such as compliance and reporting. The dark pools scrape the blotter that is, interact directly with these systems to gather IOIs. Dark pools also connect to execution management systems, which provide execution algorithms for order handling and DMA. Many market participants say the market has become too fragmented and believe that it s probably only a matter of time before dark pools consolidate, just as the ECNs did several years ago. Others don t see consolidation as inevitable because each dark pool has its own unique characteristics and flow that make it appealing to the buy side or sell side. For example, some specialize in very large sizes, some use algorithmic trading while others do not, and some do not admit sell-side flow. In fact, some large brokers have two dark pools and may find a business rationale for having even more dark pools, segmented along business lines. If anything, says Simon, I would almost expect that this kind of specialization is going to play out even more. Sherree DeCovny is a journalist specializing in finance and technology. Her work has appeared in leading U.S. and U.K. trade journals. She is also the author of several books and reports in her field. COPING WITH FRAGMENTATION Connecting to multiple trading venues is a significant challenge for most firms. Many send the order flow in their order management system to Portware, a multiasset trading system. In doing so, they can access up to 30 dark pools, builtin broker algorithms, and pretrade statistics from sell-side firms and third-party vendors. Large buy-side firms run their algorithms whether they are out of the box or custom made to hit low-cost DMA (direct market access) platforms on site in Portware s Strategy Server. Sell-side firms can use the Strategy Server to support their crossing networks, internalization engines, and dark pools. Some use it to offer Portware s custom algorithms to their buy-side customers. Nezu Asia, an equity specialist hedge fund manager based in Hong Kong that manages multiple funds and strategies, uses Portware to help traders and portfolio managers work together more efficiently. The Asian market is changing rapidly, with strong growth in electronic trading, new execution venues, and increased use of algorithms, says Nick Longcroft, a partner at Nezu Asia. Nezu needed an advanced, broker-neutral trading solution that was flexible enough to meet our emerging business requirements and scale as our firm grows. Another way for a firm to cope with multiple trading venues is by using smart order-routing technology and sophisticated algorithms to aggregate and tap into different liquidity pools both displayed and not displayed. For instance, EdgeTrade, a DMA agency-only algorithmic specialist based in New York City, offers a smart order-execution tool called FAN (Find and Nail). FAN simultaneously sends orders to aggregated dark pools and public markets and then continuously moves an order in line with shifting liquidity until best execution is fulfilled. Between August and October 2007, a daily average of 27.4 percent of clients executions through EdgeTrade s FAN were completed in dark pools. Since then, the percentage has climbed as high as percent. In addition to FAN, EdgeTrade offers Covert, which allows institutional traders to send electronic orders to dark pools only. Its Sumo product is an algorithm that enables traders to execute market orders without causing an adverse price movement. 32 C F A M A G A Z I N E / J U L Y - A U G U S T

DALTON STRATEGIC PARTNERSHIP LLP ORDER EXECUTION POLICY DECEMBER 2017

DALTON STRATEGIC PARTNERSHIP LLP ORDER EXECUTION POLICY DECEMBER 2017 General Policy Information Dalton Strategic Partnership (DSP) invests in various asset classes as part of the investment management