What to Expect From New SEC Rules on Executions

A ways back when I wrote about the best online brokers overall for 2001, a reader wrote: "I think you've left out a key factor in the value of online trading -- execution speed. All this talk of commissions is meaningless if you lose a half point" in execution.

How true. Many things that investors now consider in judging a broker -- email alerts, IPO access, even commissions -- would seem trivial if we knew that a broker got great or lousy executions for its customers.

Until now, however, there really has been no reliable way to evaluate executions. This year that's expected to change. Two new Securities and Exchange Commission rules aim to shine the proverbial bright light on trade executions. One requires brokers like Merrill Lynch and Schwab to disclose where they send customer trades for execution. The other requires disclosure from the folks who execute trades, otherwise known as "market centers." When viewed in tandem, the rules are supposed to give investors an idea of whether their broker is sending trades to market centers in their brokers' best interests, or their own.

The rules address a real need. But don't expect nirvana overnight. It will be some time before the results are available in any usable format, and even then, it's possible -- indeed likely -- that firms that crunch this data into some kind of helpful comparative format will charge for their findings. (For some light reading, check out the rules here.)

A Review of How Executions Get Done

First, a quick review on the execution process. Placing an order with a broker marks the beginning of a trade's odyssey. Once you place your order, the broker then has to route that order to a market center for execution. These market centers include specialists at exchanges like the New York Stock Exchange, "market makers" for Nasdaq stocks -- Knight Trading is an example -- and electronic communications networks (ECNs), which also primarily handle Nasdaq stocks. Sometimes, a broker has an affiliated market-making arm. Schwab, for example, has Schwab Capital Markets.

These different market centers create liquidity, ensuring that buyers and sellers will find someone to take the opposite side of their trades. ECNs are totally automated and are only set up to directly match buyers and sellers. Market makers, specialists and brokerages' own trading desks generally make money on the "bid/ask spread," which represents the difference between what sellers are asking in exchange for their stock and what buyers are bidding for shares of that stock.

Here's how it works. An investor sends his broker a market order to buy 100 shares of Cisco. The broker then routes that trade to, say, a market maker. The best prices being publicly quoted nationally -- the "spread" -- might be something like 18 1/8 bid (offer to buy) and 18 1/4 ask (offer to sell). If your order is filled at 18 1/4, that's right in line with the national best offer. If it's filled between the bid and ask, say 18 3/16, then you've done well.

The deeper within the spread your execution, the better for you and the less money for the market maker. If the bid is 18 1/8 and the ask 18 1/4, the market maker can buy at 18 1/8 and sell to you for 18 1/4. That's a gross profit of 12.5 cents per share. Gross, because the market maker will typically remit a portion of that profit to the broker who sent the trade his way as "payment for order flow." If the market maker gets you a better price than 18 1/4 -- what's known as "price improvement" -- that cuts into his profit.

What if you pay more than 18 1/4, say 18 5/16? That could mean any number of things. Perhaps someone, the broker or the market maker, dragged their feet and, in the meantime, the ask crept up to 18 5/16. Or it could mean that your broker -- acting in his own best interest -- sent that trade to someone who wasn't offering the best price. It's not supposed to happen that way, but part of the reason for these new disclosures is to find out exactly what is going on. (Fat spreads like the one in this example are increasingly rare for liquid stocks, as stock trading units long ago moved from eights to sixteenths, and more recently are transitioning into decimals. Still, the principle is the same.)

The SEC basically wants to put on display the market centers that do well by investors, those that don't, and which brokers are in bed with which of these players. The idea is that if investors have more information about who's executing trades well and who's not, pressure will build to provide better executions.

Market Center Disclosures

The first group that's subject to the new rules are the market centers. All I can say is, I wouldn't want to be the person in charge of this at these firms. They are being required to produce reams upon reams of information. It begins with every stock for which they receive a market or limit order. For each stock (not each trade, but each stock, e.g. GE or Microsoft), they must first divide the orders by type:

market orders (orders to be executed at the best price the market will bear);

marketable limit orders (limit orders that are executable given where the market is trading, e.g., a limit order to buy at $9.20 or greater if the spread is $9-$9.20);

inside-the-quote limit orders (e.g., a limit order to buy at $9.10 when the spread is $9-$9.20);

at-the-quote limit orders (e.g., a limit order to buy at $9 when the spread is $9-$9.20);

near-the-quote limit orders (e.g., a limit order to buy at $8.90-$8.99 when the spread is $9-$9.20).

Then, for each of those types of orders, they need to break it up by the number of shares per order -- either 100 to 499 shares, 500 to 1999 shares, 2000 to 4999 shares or 5000 shares or more.

Then, for each of these categories (e.g., market orders of Ariba, 100-499 shares, or near-the-quote limit orders of Yahoo!, 500-1999 shares), the market center needs to report on a number of variables, including:

Execution speed;

Fill rates, or the total number of shares they received orders for vs. the amount they actually executed;

Price improvement/disimprovement for market and marketable limit orders.

For this last item, the market centers need to say how many orders enjoyed "price improvement" -- a price better than the spread based on the nationwide best displayed bid and ask; and how many had price disimprovement -- a price worse than the spread. For each of these categories, the centers also need to quantify the amounts of improvement or disimprovement.

Each market center has to make all of this information publicly accessible from the Net and updated on a monthly basis. If you feel like you're sinking, so do many market centers. Which is why they recently secured a month delay in this rule's implementation. As it stands now, they need to start tracking this stuff in May and disclosing it by the end of June.

Broker Disclosures

Compared with the market centers, the brokers have it easy. All they need to do is tell investors where in general they're routing orders to and the relationships they have with those market centers. So broker ABC might disclose that it sent 70% of its Nasdaq market orders to market maker XYZ and received X cents a share in "payment for order flow" from that market maker. The idea is that you can take that information and then see how well that market center performs on Nasdaq market orders. If the performance is lame, you'll at least know if your broker is making a mint in the process.

Brokers are required to start reporting this information on a quarterly basis, starting at the end of October for the third quarter.

Getting Information on Your Own Trades

If you want to get personal, you'll also now be able to ask your broker where your specific trade was sent -- that is, which market center handled it -- and the time it was executed, going back for six months of trades. If you have records of when you sent the order, and what the bid/ask spread was at that time, you should be able to get an idea of whether your broker is gumming up the process. If, for example, your small Nasdaq trade took a minute or more to execute, and the market maker it was sent to has a stellar record of executing small Nasdaq trades within seconds, that delay is something you should probe your broker about.

What We'll Know and When

Don't expect to find all this good stuff in some easy-to-eat format, though. The rules only require each firm to disclose its own data, and the SEC won't do any comparative analysis on its own. The agency is expecting third parties to step in to sift through all the free data and create relative rankings. Yet, two firms that are likely candidates, broker-ranking firms Gomez and Keynote Systems, are both noncommittal so far.

"I'm not going to say yes until I actually see what the results are and whether [the data] fits our methodology," says Dan Burke, a senior analyst at Gomez. While execution quality is a major issue, Burke says, Gomez will be looking for some consistency to the results rather than seeing firms "flip-flop" from one reporting period to the next. "It's going to take us some time to sift through [the data] and build an understanding of it."

Dan Berkowitz, a Keynote spokesman, says that his company also hasn't made any decisions yet. "It's not the business we're in right now," says Berkowitz. "This is a totally foreign animal to us."

Until firms step up, we have no idea what firms that do crunch the data will charge for the fruits of their efforts.

The upshot? Comparisons of execution quality won't be anywhere near as simple or inexpensive as comparing commissions. But if they become just simple and cheap enough, they will be extremely valuable.