Algorithmic Trading: Making Money in Milliseconds

Front-Page Technology
Algorithmic Trading Making Money in Milliseconds

Today, the trading floor of the NYSE looks nothing like it once did. The most noticeable difference is the astonishing lack of people and activity. Instead of dozens of traders shouting orders over each trading desk, there are only three or four hovering around a bank of monitors. Software programs and touchscreen devices have become more efficient and cheaper to maintain than a workforce of traders — precisely why the day – trading profession is in decline. According to a Fortune article featuring Kenny Polcari, an NYSE trader, the Street had lost about 4,800 traders since he started working in 1980 [1]. Due to the rise of software-based trading, stock exchange floors have become tame and almost dull compared to the frenzied atmosphere that they once had about a decade ago.

Algorithmic trading has always been subject to backlash — much of which comes from traders who have been impacted by the rise of automated trading. However, part of this vitriol stems from a lack of understanding of this technology. Automated trading relies on algorithms programmed by specialized engineers, known as “quants”, who write programs that identify what a regular trader cannot.

Critics of algorithmic trading also point to “flash crashes” as a downside of the heavy reliance on digital trading. A flash crash occurs when an error in the program generates a downward spiral of stock prices. On May 6, 2010, a flash crash occurred, causing the Dow Jones Industrial Average to incur the largest intraday loss in history — a ghastly 998.5 points. To put that into perspective, the Dow usually swings a maximum of 100 points each day.

A lengthy investigation into the May flash crash concluded that a single event had prompted the swing. It was perpetuated by high frequency trades (HFTs), which are designed to exploit infinitesimal price swings generated by changing market demands. Though the per-trade gains may be small, over the course of millions of daily transactions, the monetary value of these gains multiply significantly.

The flash crash began when one hedge fund sold a large number of assets. The parameters of this algorithmic sell order accounted only for volume. It was carried out in 20 minutes, a mere fraction of the several hours required to prevent a wild swing [2]. Because HFTs are programmed to identify these changes indiscriminately, they followed suit and a massive sell-off ensued. Traders tried to respond with manual buys, but there was no way they could outmaneuver thousands of instantaneous, computerized orders. Eventually the market rallied, but the May flash crash was a wake-up call to investors, regulators, and Main Street. In an April poll conducted by the CFA Institute, 52 percent of respondents chose the option: “HFT does more harm than good to the market”, compared to the 14 percent that claimed HFT to be beneficial [3].

However, pointing fingers at algorithmic trading alone is an oversimplification of a very complex issue. It’s a technology that’s still difficult to understand, even for people whose livelihoods depend upon it. But at the end of the day, blame man, not machine. These algorithms don’t create themselves; they’re written and programmed by people who inevitably make mistakes. At the root of the May flash crash was someone who wrote a program that did not spread orders over time intervals. Mr. Polcari criticizes the heavy reliance on computer trading for lacking human intuition and judgment. Though this may be true, human error is always present as well. Ultimately, as our understanding of algorithmic trading evolves, so too will the mechanisms we have to guard against similar disasters.

Trepidation about technological advancement will always exist. Fears about electronic trading closely mimic the discontent of farmers and artisans displaced by cheaper machines during the Industrial Revolution. Not too long ago, manufacturing completely dominated the American economy, catapulting it to unprecedented levels. These manufacturing jobs are now a significantly smaller part of the economy, having been replaced by white-collar positions. Society evolves over time, and advancements will benefit one group as much as they hurt another. For every quant employed to write algorithms for Goldman, a desk trader could lose his job. It’s the unfortunate side effect of change, but it would be unfair to single out HFTs and algorithmic trading as an “enemy” of Wall Street traders and the industry.

1 Comment

  1. I don’t even understand how I stopped up right here,
    but I thought this submit was great. I do not know who you’re but
    certainly you’re going to a famous blogger in case you are not
    already. Cheers!

Leave a 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=""> <s> <strike> <strong>


Time limit is exhausted. Please reload CAPTCHA.

Lost Password