Trade Execution

Once the appropriate trade strategy is determined by the portfolio manager and the trader, the trade must be executed in a market and in a manner consistent with the trade strategy chosen.

A variety of implementation choices are available based on the specific order, market, and trade strategy involved. Trade implementation choices range from higher-touch approaches, which involve greater degrees of human interaction for order completion, to fully automated trade execution through electronic trading venues with varying levels of trade transparency.

In general, trading in large blocks of securities requires a higher-touch approach involving greater human engagement and the need for a dealer or market maker to act as counterparty and principal to trade transactions. For these transactions, also called principal trades or broker risk trades, market makers and dealers become a disclosed counterparty to their clients’ orders and buy securities into or sell securities from their own inventory or book, assuming risk for the trade and absorbing temporary supply–demand imbalances. In the case of a less active security, the expected time to offset the trade for the dealer is longer. For taking on this additional risk, the dealer will demand greater compensation, generally by quoting a wider bid–ask spread.

Summary: High touch principal trades are used in executing urgent and large trades. Dealers buy/sell from own inventory, and can be more costly based on supply demand.

Markets characterized by dealer-provided quotes may be referred to as quote-driven, over-the-counter, or off-exchange markets. In such bilateral dealer markets, customers trade at prices quoted by dealers. Depending on the instrument traded, dealers may work for commercial banks, investment banks, broker/dealers, or proprietary trading firms.

In some cases, dealers may be unable or unwilling to hold the securities in their inventories and take on position (principal) risk. In agency trades, dealers try to arrange trades by acting as agents, or brokers, on behalf of the client.

Summary: High touch agency trades are used in executing large block trades that are non-urgent very illiquid. Dealers attempt to arrange trades as a broker.

High-touch approaches involve human sell-side traders as intermediaries. These traders, employed by sell-side brokerage firms, may first attempt to fill a customer order by matching it with offsetting orders from other customers before trying to fill it from their own position book. Crossing an order with a broker’s own book is known as a broker risk trade or principal trade. If this does not occur, the broker would then route the order to the open market and “slice,” or divide, the order into smaller pieces to trade in the market. This approach involves human judgment unique to each trade and is suited to trading illiquid securities in which the execution process is difficult to automate.

For relatively liquid, standardized securities where continuous two-way trading may exist, buyers and sellers display prices and quantities at which they are willing to transact (limit orders) on an exchange or other multilateral trading venue. In order-driven markets, order-matching systems run by exchanges, brokers, and other alternative trading systems use rules to arrange trades. Trading is done electronically with multiple venues, often through a consolidated limit order book that presents a view of the limit buy (bid)/sell (ask) prices and order sizes for all venues with orders for a security. Centralized clearing for trades exists on those venues. Equities, futures, and exchange-traded options are generally traded using this approach.

Electronic trading involves trading via computer and is used in more liquid markets. Trading here is typically order driven in that electronic systems allow buyers and sellers to advertise their limit orders in a central limit order book. A limit order is an order to trade at a certain (limit) price or better.

Electronic trading generally involves direct market access (DMA) and/or algorithmic trading. DMA allows buy-side portfolio managers/traders to access the order book of the exchange directly through a broker’s technology infrastructure.

Algorithmic Trading

Algorithmic trading is the use of programmed rules to electronically trade orders, primarily used for two purposes: profit seeking and trade execution.

  • Profit-seeking algorithms use real-time market data to determine which securities to buy and sell, and are employed by electronic market makers, quantitative funds, and high-frequency traders.
  • Execution algorithms trade according to the rules specified by the manager to meet their objectives. Types of execution algorithms include the following:
  • Scheduled algorithms—percent-of-volume (POV), VWAP, and TWAP algorithms. These execute trades using rules driven by historical volumes or specified time periods.
    • Scheduled algorithms are appropriate for relatively small orders in liquid markets for managers with less urgency and/or who are concerned with minimizing the market impact.

POV algorithms (a.k.a. participation algorithms) send orders according to a volume participation schedule.

  • Advantage: They automatically exploit increased liquidity when available.
  • Disadvantage: They continue to trade at any (potentially adverse) price, and may not fill the order in a specified time if there is a lack of trading.

VWAP and TWAP algorithms are time-slicing algorithms. VWAP algorithms attempt to match the VWAP price for the period by carving up the trade and sending orders based on historical intraday volumes. TWAP algorithms perform a similar task; however, they ensure an equal number of shares is traded in each time period.

  • Advantage: They ensure that a specified number of shares are executed in a specified time period.
  • Disadvantage: They may force trades in times of low liquidity or trade too little in times of high liquidity.

Liquidity-seeking algorithms (a.k.a. opportunistic algorithms) aim to take advantage of favorable liquidity conditions when offered by the market. These orders use both lit and dark venues.

  • Liquidity-seeking algorithms are appropriate for larger orders in less liquid markets with higher urgency while trying to mitigate the market impact. They are also appropriate when a manager is concerned that displaying limit orders may lead to information leakage, or when liquidity is typically thin with sporadic episodes of high liquidity.

Arrival price algorithms seek to trade close to market prices prevailing at the time the order is entered. These algorithms will trade more aggressively than other algorithms to trade more shares at close to the arrival price.

  • Arrival price algorithms are appropriate for relatively small orders in liquid markets for managers who believe prices are likely to move against them during the trade horizon, and therefore wish to trade more aggressively (e.g., a profit-seeking manager). They are also appropriate for more risk-averse managers who want to minimize execution risk.

Dark strategies/liquidity aggregators execute trades in dark pools, with aggregator algorithms attempting to optimize trading across multiple dark venues.

  • Dark strategies/liquidity aggregators are appropriate for large orders in illiquid markets, and arrival price or scheduled algorithms would likely lead to high market impact. Since there is a lower chance of execution in dark pools, these strategies are for managers that do not need to execute the full order immediately.

Smart order routers (SORs) are algorithms that determine the best destination (either lit or dark) to route an electronic order to get the best result. SORs focus on getting the best price for market orders, or the highest probability of execution for limit orders.

  • SORs are appropriate for small market orders with low market impact where the market can move quickly, or for small limit orders with low information leakage where there are multiple potential execution venues.

Recent developments in algorithmic trading include clustering and high-frequency market forecasting.

Clustering is a machine learning technique whereby a computer learns to identify which algorithm is optimal for different types of trades based on the key features of trades.

High-frequency market forecasting attempts to model short-term market direction.

Comparison of Markets

Equities are usually traded on stock exchanges (lit markets) and dark pools. Equity markets are the most technologically advanced, with most of the trading executed electronically, and the use of algorithms is commonplace.

Fixed-income markets tend to trade in a large number of heterogeneous securities, as many issuers have multiple issues outstanding. While the fixed-income markets tend to have low liquidity, typical order size is large. Due to these characteristics, trading is mostly conducted in dealer-based, quote-driven markets. Electronic RFQ systems are becoming more common; however, algorithmic trading is largely limited to only the most liquid on-the-run (most recently issued) U.S. Treasuries and futures contracts. Electronic trading, while growing for corporate bonds, remains relevant for only a small fraction of the universe. Other fixed-income securities markets generally use high-touch execution methods—urgent trades would require a principal broker risk trade, while less urgent trades would use a broker-agent approach.

Electronic trading is common for exchange-traded derivatives. Algorithmic trading is not as common as for equities markets, and is used more for futures than for options. Buy-side traders generally use DMA.

OTC derivatives trading takes place in a dealer quote-driven market, usually implemented through high-touch approaches. Since the credit crisis of 2008, there has been a move by regulators to increase transparency and central clearing of basic OTC derivatives such as interest rate swaps.

Spot foreign exchange trading takes place in OTC markets that use both electronic trading and high-touch broker approaches. The market works in three tiers: interbank, interdealer, and bank-to-client, with decreasing trade sizes and increasing spreads, respectively. For large urgent trades, RFQs are used with brokers. For large non-urgent trades, scheduled algorithms or high-touch agency approaches are used. Small trades are usually implemented using DMA.

Table of Contents

Leave a Comment