High Frequency Trading: The Unnecessary Middleman

High frequency trading has been a hot topic of discussion over the past few years in the investor community. In his book, Flash Boys, Michael Lewis claims that the “market is rigged” because of high frequency traders. As well, Mark Cuban, Owner of the Dallas Mavericks, has publicly spoken out against high frequency trading in his post here, and to quote Mark:

If you know that by getting to the front of the line  you are able to see or anticipate some material number of  the trades that are about to happen, you are GUARANTEED to make a profit.  What is the definition of a rigged market ? When you are guaranteed to make a profit.

If you look at high frequency trading firm Virtu Financial, set to go public in 2014, they have only had 1 day of trading losses since 2009. There are many other firms out there in similar situations. So what is going on at these high frequency trading firms and how are they consistently being able to profit on trades?

Brief Overview of HFT

High frequency trading covers any trading strategy that uses sophisticated technological tools and computer algorithms to execute trades at rapid speeds. The high frequency trading strategy that is often scrutinized and most controversial is latency arbitrage. When I mention high frequency traders from here on, I am referring specifically to latency arbitragers. To put it briefly, latency arbitrage is profiting off Person A placing an order to purchase stock and the latency arbitrageur purchasing and selling the stock before Person A’s order is complete. This is done by using algorithms to first predict/anticipate Person A’s order, and using ultra-fast computer technology (to achieve the lowest latency possible) to execute a trade before Person A’s order is complete. High frequency traders end up being an unnecessary middleman.

Brad Katsuyama had a great analogy with ticket scalpers and Stub Hub in an interview he did with CBS. Imagine you needed to buy 4 tickets to a concert for you and your family. You log onto Stub Hub and place an order for 4 adjacent seats to the concert for $20 per ticket. After the purchase is half-complete for the first 2 tickets at $20 each ticket, imagine a scalper somehow knew that your entire order wasn’t complete and you needed 2 more seats. With this information the scalper purchases the 2 adjacent for $20 each, and then sells these tickets immediately back on Stub Hub for $25 each. You end up paying $10 more, and the scalper profits $10. Now imagine all this happened within fractions of a second, automated by high-powered computers and occurring millions of times per day. That’s high frequency trading.

Improving Liquidity

The most common argument that high frequency traders use is that HFT improves liquidity. HFTs often argue that because of their large quantities of trades, they make stocks more liquid and marketable to investors. The definition of market liquidity, from Investopedia, is as follows:

The degree to which an asset or security can be bought or sold in the market without affecting the asset’s price. Liquidity is characterized by a high level of trading activity. Assets that can be easily bought or sold are known as liquid assets.

Are high frequency traders increasing (legitimate) trading activity or making stocks easier to purchase/sell? Not really. If you think about it, in an latency arbitrage, that trade between the 2 original parties was going to happen regardless of the existence of high frequency traders. The high frequency traders are simply squeezing themselves in between the trade for fractions of a second in order to scalp a bit of profit. This is just phantom volume, designed to go in an out of a position in a flash to secure a tiny profit, not really increasing legitimate trading activity at all. This does not improve liquidity.

Which Investors Are Affected?

High frequency trading may not affect individual retail investors directly as their trades are simply not large enough to span across multiple exchanges and trigger HFT algorithms to anticipate the retail investor’s order. However individual investors are affected indirectly through institutional investors (mutual funds, pension funds, university endowments etc.). Over past decade, the rise of mutual funds and ETFs has caused direct retail investing to decrease, this money instead being invested indirectly through institutions.

This means that high frequency traders affect the investor community as a whole, both individual investors and institutional investors. As I mentioned, a large amount of individuals invest through institutions and in turn institutions make large enough trades for high frequency traders to scalp profits off of, costing both the institution and the individual.

Free Rider Problem

The situation is similar to the free rider problem in economic theory. One person who always sneaks onto the public transit bus and doesn’t pay will consistently save money. Although this free rider is costing the city a tangible amount of money, the amount is extremely minuscule. Large profit for the individual, small cost to the system. The problem starts when other bus riders notice that this individual is consistently saving money, and may be motivated to do the same thing. As the amount of free riders start to grow, the system starts to take a toll in terms of costs, and eventually get to a point where either the system collapses or corrective action is taken.

Similar to the stock market, high frequency traders scalping profits off legitimate investors end up costing the investor community. If trading firms believe that they could make guaranteed profits by following an HFT strategy that uses latency arbitrage, the number of high frequency traders in the market will continue to increase to the point where either corrective action is taken or the system collapses.

Minimum Holding Period and Simultaneous Execution

From a regulatory standpoint, the most reasonable solution to the HFT problem is setting a minimum holding period of even just a few seconds. This shouldn’t have any material impact on liquidity and would stop most of the high frequency trading, all but the more larger traders across global exchanges. If the SEC mandated this minimum holding period, a lot of HFT firms that use latency arbitrage would either have to close shop or find some other way to profit off of their algorithms. In theory, this should solve the problem, however becomes a bit more complicated when exchanges compensate market takers/makers for the added “volume” they bring to the exchange.

From a buyside standpoint, using a trading platform that executes the trades across the different exchanges it needs to complete the order simultaneously, instead of one exchange after another, may fend of high frequency traders attempting to scalp profits off the order. This is what Brad Katsuyama’s THOR platform attempts to achieve.

Even though there are many legitimate HFT strategies out there that overall benefit the market and it’s liquidity, the ones that attempt to scalp a profit by using ultra-low latencies end up hurting legitimate investors. These high frequency traders divert from the original purpose/intention of the stock market and end up costing the investor community as a whole.

Jathu Vasantharajah

This entry was posted in Investment Philosophies. Bookmark the permalink.

3 Responses to High Frequency Trading: The Unnecessary Middleman

  1. Lloyd Marsh says:

    The scalper analogy really ties it all together. The algorithms used in HFT effectively create value only for the firms facilitating the trades, while creating market inefficiencies. Assuming a market with only HFT firms, no trade would happen at equilibrium creating a giant increase in DWL (dead weight loss). Furthermore, if the consumer bears the fees for each trade companies issuing public offerings will have to rework their valuations to take into account the increased costs of facilitating each trade. Then there’s further inefficiencies as companies completely undervalue or overvalue share prices as they attempt to gauge the true demand within the market.

    Haha regardless of the increased inefficiency, this is a really cool concept. Modelling consumer behavior is damn near impossible, and with that one firm having one day of losses since 2009…that’s huge. However, it makes me wonder what went wrong on the day they did experience some losses. Was it a problem with the algorithm? Were losses experienced across the market? Did their servers crash?

    Thanks for another awesome article bro, definitely learning a lot!

  2. Jathu says:

    Thanks for your insight Lloyd. Great point about the inefficiencies, and the increased costs of facilitating trades. I’m curious as well as to what happened on that one day Virtu experienced a loss!

    Hope all is well,


  3. And what caused the loss of liquidity? Well, it appears that one big factor was the flight of high-frequency traders from the market. The algos of the quants just didn’t work well when the HFTs refused to provide liquidity.

Leave a Reply to Jathu Cancel reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>