Dark Secrets: Where Does Your Retail Order Go?

Back when I started trading in 1999, the majority of trades took place on the exchanges themselves. But now, believe it or not, many retail orders never see the public exchanges. If these orders aren't sent to the public exchanges, where do they go? Many of these retail orders are internalized. In simple terms, that means the opposite party to your executed order was your broker/dealer themselves. These retail orders never leave the brokerage house. Other times your broker will ship your order to an over-the-counter (OTC) market-maker. The OTC market maker will pay your broker a fee, for the privilege of trading against your order. This is known in the industry as "payment for order flow". Why is there so much interest in trading against retail orders? Because many retail orders are "uninformed", meaning they don't have much knowledge of where the stock is going in the short-term. The most uninformed order is the "market" order, which is willing to pay any price to buy or sell the stock. Market orders are very attractive to trade against because it allows the broker/dealer, or OTC market maker the best chance of capturing the spread. Consider a stock like Bank of America BAC that can trade in a 1 cent spread for hours on end. Assume the NBBO (National Best Bid and Offer - the public quote) for BAC is $5.25 bid for 100,000 shares and $5.26 ask for 100,000 shares. A trader placing an order to buy this stock at $5.25 would have to stand in line behind the 100,000 shares that were bid before them. Similarly, a trader placing an order to sell the stock at $5.26 would have to stand in line behind the 100,000 shares that were offered ahead of them. Now assume, a retail trader sends a market order to sell BAC. In the old days, the market order would typically be executed against the participant bidding $5.25 on a first-come, first-served basis. But now the mechanics can be quite different. The OTC market maker (paying for the retail order flow) intercepts the retail market order, and buys the stock for their own account at $5.25, leaving the participant bidding the stock on the public exchange at $5.25 unfilled. Now assume a retail trader sends a market order to buy BAC. This order should be executed against the participant offering the stock on the exchange at $5.26 on a first-come, first-served basis. But again, the OTC market maker intercepts the retail market order, and sells the stock for their own account at $5.26, leaving the participant offering the stock on the public exchange at $5.26 unfilled. The OTC market maker pockets the one cent spread. The OTC market maker is in essence, jumping the order queue, allowing them to transact again and again at the front of the line, as retail market orders are sent from various retail brokerages. Do this thousands of times per day and the profit from capturing that one cent spread can add up to a substantial sum. If you like math, here is the calculated estimates: Considering the average daily volume over the past three months on BAC is 264 million shares per day. The SEC estimated in 2010 that 17.5% of trades are internalized(1), meaning 264 million shares x 0.175 = 46.2 million shares internalized. Considering the average spread in BAC is 1 cent, then OTC market makers have the potential to make: 46.2 M / 2 x $0.01 = $231,000 per day (note you divide by 2, because it takes 2 transactions to capture the spread; the buy, and the sell). $231,000/day x 250 trading days per year = $57.8 million per year just in BAC. As you can see, this queue jumping can add up to serious money over time. Where does that money come from? It comes from the participant who was left unfilled on the public exchange. Because if the participant was left unfilled, they would have to do one of two things:

1. pay the spread - which in the case of BAC is 1 cent or
2. remain unfilled - in which case the loss is unquantifiable as the participant may never be filled.

Why is this allowed? To justify the practice, OTC market makers will often offer price improvement to the retail market orders. Usually this is a few sub-pennies better than the NBBO. Instead of printing the market sell order at $5.25, they print the order at $5.2501. This saves the retail trader a whopping 1 cent on their $525 order. But let's give the benefit of the doubt here, and say that the OTC market maker offers $.001 price improvement per order, instead of $.0001 per order. 46.2M x $.001 = $46,200 per day or $11.6 million per year in price improvement in BAC. So the public is still out $46.2 million ($57.8M - $11.6M) in BAC. And that's assuming that the participant paid the spread. If they refused to pay the spread, the losses are unquantifiable, as the participant may never be filled. The bigger problem is the role these internalization practices play on displayed liquidity. What point is there to bidding a stock if some privileged participant is able to step in front of your bid at the moment the stock is about to be sold to you? This practice discourages participants from placing passive limit orders, and hence, the market is left with less liquidity. With less liquidity we become prone to future flash crashes (yes, remember that ugly day back on May 6th, 2010 when the market fell 6% in five minutes only to rally it all back). And we keep wondering, what caused that flash crash. Maybe we should take a deeper look into broker-dealer internalization as a possible cause. So next time your broker tells you that you received price improvement on your order, let them know that you'd rather they keep the sub-pennies, and execute your order against the quote on the public exchange. You'd be much better off in the long-run. (1)SEC Concept Release on Equity Market Structure, http://www.sec.gov/rules/concept/2010/34-61358.pdf, page 15