Brad Katsuyama may be best known as the antagonist in Michael Lewis’ controversial book “Flash Boys: A Wall Street Revolt,” but he is an innovator. As global head of electronic sales and trading at RBC Capital Markets he didn’t like the way his customer orders were being filled. He never seemed able to hit the entire bid or lift the entire offer that the market showed when executing orders. He realized that faster players were able to see what he was up to and get in front of his orders when the order moved from one trading venue to another.
Katsuyama and his team at RBC devised smart order routing technology, Thor, which allowed an RBC order to hit different trading venues at the same time.
Alpha Pages: You mentioned that exchanges have different levels of colocation. Explain.
Brad Katsuyama: Exchanges have three versions of cables that they will sell you with tiered pricing and speeds. You could poke around at the [NYSE] and NASDAQ. They have three versions of cables that they will sell you. First you pay to be in the room. Now all of a sudden you have three options to connect to me. You have a one gig cable at New York, it’s $10,000 a month for two cables; you have a 10-gig cable which is $25,000 a month for two cables and you also have a 40-gig cable for $40,000 a month for two cables. These cables are not that expensive alright? The difference in the gig is speed, microseconds. You can save two or three microseconds by getting a faster cable.
AP: So the economics of colocation provide the exchanges with revenue by creating a tiered system?
BK: Yeah. There are those participants who are colocated and those who are not. So if we let someone colocate right next to IEX, even though we’ve made every attempt at getting information as fast as we can they’re still going to be faster. So the only thing that we could do to actually ensure that they were not trading with information that we didn’t have, is put them farther away. We had to go anti-colocation. And the distance that we wanted them away was 43.5 miles. Unfortunately there was no 43.5-mile data center away from the data center where our matching engine was so what we did was put them five-and-a -half miles away and coil 38 miles of cable in a box to create distance.
AP: That’s how you created 350-microsecond latency?
BK: Yes, and this is the reason why. The market is 10 by 11 we have a 10.5 buyer, the market changes to 9 by 10, HFT firm gets the update, they place an order to sell stock at 10.5 on IEX. As it’s going around the cable we get the update that the market has changed to 9 by 10, we’ll slide the order to 9.5 and by the time the order gets there we will have moved this order to the correct price.
AP: Have exchanges, by enabling colocation, undermined fairness in the market?
BK: Yeah. The fact that there’s a variance in the speed of market data and the speed of participants, offering colocation undermines the exchange’s ability to ensure that no participant has information that they don’t have.
AP: Should the regulators restrict colocation?
BK: The problem with doing that is that they’ll just go across the street. It will slow them down slightly but they won’t force them to go 43.5 miles away.
AP: Are those colocation revenues meaningful?
BK: Trading is the fourth largest net revenue stream [for equity exchanges]. They bunch [colocation] under technology services [so it is meaningful].
FM: Can you compete without those fees?
BK: We charge a flat fee: 9¢ per 100 shares for everyone. It’s cheaper than the 30¢ per 100 shares to take liquidity off let’s say ARCA, but they will also pay a 29¢ rebate. So, the people who are used to collecting rebates we’re way more expensive. For the people who are used to paying the high fees, we’re cheaper.
FM: Is there something the regulators should do?
BK: It’s always better for the market to try to address market related problems before we look to the regulators. There are things that the regulators can do that aren’t directly related to touching the market, i.e., disclosure and certain transparency measures.