By Irene Aldridge
Just over two weeks ago, NASDAQ stopped trading midday, fueling a new wave of speculation about the reliability and, ultimately, appropriateness of using computer technology in trading. Critics of trading algorithms readily jumped on the news bandwagon, happily denouncing technology as a source of all economic ills.
In reality, however, on Aug. 22, 2013, NASDAQ halted its servers to comply with instructions from the U.S Securities and Exchange Commission (SEC). Had NASDAQ continued trading as usual that day, it would undoubtedly face hefty fines on the order of $10 to $15 million, like the penalty assessed to the New York Stock Exchange by the SEC in September 2012 for failing to stop trading in similar circumstances.
How so? First of all, the halt was not driven by any disruptions in the trading engine of NASDAQ. In other words, all trading orders submitted to NASDAQ were processed in proper fashion without a hint of trouble, just as they always do. The actual disturbance triggering the trading halt occurred in the reporting functionality of NASDAQ. This reporting engine, known as Securities Information Processor (SIP), is an SEC-designed system common to all exchanges in the U.S. that is used to collect and archive the latest market information from all exchanges. Once the information is accumulated by the SIP, it is distributed back to the exchanges. This is done with the explicit goal of protecting individual investors by ensuring that every investor’s market order is executed at the best price available for a given security in the U.S. exchanges at any given time.
NASDAQ halted trading when it detected that the timestamp on NASDAQ messages sent to SIP differed from the timestamp of messages sent from SIP to NASDAQ. As I discussed in my new book, High-Frequency Trading: A Practical Guide to Algorithmic Strategies and Trading Systems, 2nd edition (Wiley), disruptions of this sort can be due to factors as benign as machine overload, or as malicious as hacking, the latter ruled out by NASDAQ. And according to the SEC guidance, NASDAQ has always had the authority and discretion to “(1) halt trading on Nasdaq of a Nasdaq-listed security to permit the dissemination of material news.” Since the latest trades and quotes are “material news” for many traders relying on technical and quantitative market analysis, as well as for investors that simply seek to obtain fair execution in the U.S. markets, NASDAQ acted appropriately under the circumstances by shutting down the systems until malicious causes of the incident were ruled out.
It’s worth reiterating that, just like in any other industry, technology is good for the financial sector. In addition to lowering costs of trading, making markets more transparent, and returning money to investors, financial technology has been making our markets efficient year after year. Ask any economics and finance professor about market efficiency, and they will define it as follows: efficient markets distribute and incorporate news instantaneously. Most academics rightly think that this is the primary function of the markets, and not casinos that some people have come to expect from exchanges instead.
Investors who call for a ban on market technology and high-frequency trading (the two are really the same discipline) should consider the following: Unlike 20 years ago, we are now bombarded by millions of data points every second — market data, economic indicators, news, Twitter, etc. Human eyes are biologically incapable of processing such volumes of information. (Think about this: movies display only 24 frames per second, and human brain thinks that it is watching a real-time show as people cannot absorb more data than 24 frames per second.) Without technology, financial markets simply cannot function any longer.
As a result, denouncing current technology does little to help the markets. To prevent future glitches in trading, markets need more technology, not less. Developing the new technological frontier is easier than it may seem to a casual observer since many models for scientific monitoring and prediction already exist. Implementation and deployment of newer automated surveillance and risk management methodologies may need to become a priority of organizations across the entire financial system, at NASDAQ and beyond.