In light of Michael Lewis's new book, "Flash Boys: A Wall Street Revolt", and the overwhelming response, ranging from
regulatory and
critical, it may be important to review the sector and how information technology has impacted the space.
High-frequency trading ("HFT")
utilizes computer programs, algorithms, 'black box', and equivalent to execute trades at high speed. In this space, profits are positively correlated with time to trade execution. HFT may track the speed at which an asset's price is going up or down or the delta of an order book or may look at
technical trading indicators like MACD or Bollinger bands for example. Traders are not executing large block orders, but are entering/exiting smaller positions to capture pennies. In a sense, they are 'sweeping' nominal capital left on the table during a trade (or potentially many, many trades).
How a 'Flash Order' using HFT works
Source: Wikinvest
In terms of pros and cons to the market as a whole, advocates mainly point to
market liquidity and lower fees, while critics point to market manipulation as HFT engages in '
scalping' markets or may
even place orders it never intends to make or may artificially inflate asset prices (much like the FED. But that's another conversation all together).
Regardless of which side of the fence you are on (or even if you straddle), HFT comprises of
~70% of market volume. Yet is HFT doing anything that market participants haven't been doing before technology allowed HFT to flourish? Doesn't the market seek to react as fast as possible when new information is presented? Let's assume HFT has early access to info. How are HFTs getting information early and other participants not? That seems challenging to believe when position alpha is decided by milliseconds. Meaning, we could argue HFT and other participants have access to the same non-material and non-insider because HFT trade value vanishes in a fraction of a second. But is it hard to believe when market subscription services exist?
Pay for Play?
CME-NASDAQ has a subscription service to get data fast. The file seems to suggest that a subscriber is linked directly to market data centers allowing for one-way speeds of ~4.25 milliseconds. There are even works for a
laser network between the NYSE and NASDAQ. As long as the subscription service is cost efficient for the HFT, then anything that improves speed--even if by a time so nominal it would seem logically irrelevant--will have value. But what about to a pension fund manager? Mutual fund? Or an individual RIA? It seems hard to believe they would view such technology as valuable to their business. But is their lack of participation in this tech allowing HFT to capitalize on their trades?
Level Playing Field
Would be interesting to see how long HFT could last if all participants had access to the same data services. If anything, such a level information field could potentially justify HFT as they are just the first to the table and the most sophisticated when data arises. Yet, this seems too optimistic. As HFT seems reactionary to market activity. Also, given the large amount of volume, who would receive priority from an executing broker? A mutual fund that trades once a month or an HFT that trades 1000s of times a day? It seems if the broker is motivated by commissions, then the HFT is the preferred client.
This could potentially allow HFT firms's orders to be placed ahead of other clients. Such discrepancies make the idea of efficient markets outdated. Instead of fundamentals, participants seek to trade off the mechanisms of the market where no matter the movement, the trader first to the plate makes money in the trade. As regulators continue to navigate through these '
dark pools' they would probably be doing themselves a disservice if they overlooked the market's (from auction to ECN to OTC) themselves and just focused on firms that react to inefficiencies that the markets--not participants--create.
- Nile C.
4/28/14 2:39pmEST