Making Crosses Work

Price improvement doesn't always drive trading decisions

In the November issue, Traders Magazine reported on comments made at a recent industry conference regarding the effect of penny price increments on institutional trading. One options executive stated that large trades in penny option classes are increasingly executed over the counter and not on an exchange.

Although the same person later noted that the practice is not widespread, the thought that tradingexchange-listed options in penny price increments may drive trades to the OTC market is surprising, especially given the recent market turmoil created in large part by problems stemming from the OTC market. Such turmoil illustrates investors’ lack of confidence in a market without evident competition. That turmoil also vividly demonstrates the lack oftransparency within both the trading and position marking process in the OTC market.

Penny pricing is certainly a hot-button issue for options traders, and it is used by its detractors to conjecture all manner of unintended consequences. Many times these predictions provide valid insight into the speaker’s real agenda. An analysis of the comment that “Traders are having a very difficult time putting trades on the tape” may suggest that the real issue is upstairs dealers’ desire to keep the most lucrative crosses in-house, not any flaw in existing crossing mechanisms.

As option spreads began to tighten after multiple listing in 1999, it became more difficult for market makers to remain profitable. Some of these firms found that it made sense to pay brokers in order to get the first look at customer order flow and be named as contra on the crosses. Brokers profited because they could now bill both sides of the order.

Electronic Speed

By rule, a human floor broker with instructions to cross an option order must first request a verbal quote from the trading crowd, and then try to cross the order inside the best of the combined verbal and electronic quote. Since penny pricing has dramatically tightened options quotes, and these narrower markets update at electronic speed, a human broker may not be able to create the required audit trail and get the cross on the tape before the market moves away from his predetermined price.

Given the slow speed of manual crossing on exchange floors, one would think that these brokers would use the ISE’s electronic crossing mechanism instead. While the ISE certainly speeds up the process, its crossing mechanism presents a new challenge to cross-only order flow.

The problem for upstairs dealers is that they cannot attempt a cross electronically without exposing a customer’s order to the price-discovery process via a three-second auction. Some floor brokers may simply inquire whether a cross can occur at a given level without actually exposing a customer order to the price-discovery process. If they find that other (non-paying) traders want to participate, the broker shops the order to different floors until he can find one where he can do the largest portion of the cross.

While this practice has been condoned by trading floors for many years, the ISE system is designed to obtain the best price. Therefore, crosses may be broken up on the ISE when one side of the intended cross gets price improvement. For orders larger than 500 contracts, the cross will go up clean, unless the entire order gets the same amount of price improvement. My block-trading desk uses the ISE’s electronic auction mechanism to put up crosses regularly.

Once a cross is exposed electronically on ISE, it may trade at a better price for the customer, and the firm side cannot pull the cross back when it finds it will not participate. Many upstairs firms are not willing to risk being price-improved electronically, and therefore avoid using the ISE to cross. This is because the electronic auction process may prevent the most profitable trades from being internalized, and therefore make price improvement to the customer a secondary consideration.

The Scenario

The Traders Magazine article reported a trading scenario: “A dealer bringing a 1,000-lot trade down to the floor … hoping to print it at $1.18 per contract, could be thwarted by a subsequent bid of $1.19 for, say, 10 contracts.” That would be the case if the broker took the trade to a floor, rather than using an electronic system such as ISE’s. At ISE, the trader could cross the 1,000-lot at $1.18, but would have to wait three seconds while the exchange’s electronic auction sought a better price for the customer’s order. If no better price existed for the entire trade, it would print at $1.18. Since large-size trades are exempt from linkage rules, the $1.19 bid for a 10-lot would, in this case, be ignored. However, if another liquidity provider was willing to improve the entire order, the customer would get the better price.

In the interest of full disclosure, my firm, Interactive Brokers, supports penny pricing because price improvement is our primary concern. Sometimes customers evaluate the cost of trading only by comparing commissions, and fail to evaluate the quality of execution. Quality of execution, or the actual price where the trade is filled, is usually the most important price component. Consider that saving 10 cents per contract in commissions but paying even a penny more in execution price will cost an option trader $90 on a 100-lot.

If you are a customer looking for price improvement to the NBBO, you are more likely to see such price improvement if the other customers and liquidity providers can show improvements in small increments.

Imagine a limit order to sell 50 XYZ at $2.10 when the market is quoted as $2.10-$2.20, 1,000 up. Multiple market makers that each participate at the $2 bid may each only get a small piece of the 50-lot, but stepping up a penny and paying $2.01 will give a single market maker the whole trade. If the midpoint ($2.05) is really fair value, most market makers would rather make 4 cents 50 times than 5 cents on a 5-lot. In this case, the customer made an additional $50 on his option sale, because penny pricing encouraged other traders to compete for his order.

If that option class does not trade in penny prices, the minimum improvement would come at a cost of $5 per contract, and on a quote of $3.00-$3.20, the minimum improvement is $10 per contract. These large price increments are often too much in lost edge for a market maker to pay. Therefore, the customer may not see any price improvement in this class.

Price improvement can also occur when an order is routed to an exchange that supports an electronic price-improvement auction. These auction mechanisms currently exist not only on ISE, but also on Arca and BOX, and typically last under three seconds. Since even non-penny option classes are eligible for penny price improvement via these auctions, option customers may benefit if their brokers have the ability to expose their orders to these auctions.

Multiple listing, electronic trading and the reduction of minimum ticks from sixteenths to nickels to pennies have had a profound effect on the options industry. Market makers and specialists have seen their ranks decline, while volume has increased with faster and cheaper access to the markets. While everyone has strong feelings about the introduction of penny pricing for options trading, institutional dealers’ desire to avoid price discovery on the most profitable orders does not seem to be a valid reason to delay the rollout of penny pricing.

Kevin Fischer is the manager of the options block-trading desk at Interactive Brokers in Greenwich, Conn.

(c) 2008 Traders Magazine and SourceMedia, Inc. All Rights Reserved.