Machine trading
Algorithmic trading is a hot topic in financial circles these days. Despite it already being a very important and visible phenomenon on Wall Street and the world’s financial markets, in my opinion we are only at the very beginning of what will be a transformational shift in how trading (in all types of markets) is conducted. The market research firm Celent last year predicted a five year compound growth rate of 13% (between 2003 and 2008):
Celent predicts that algorithmic-based equity trading will increase from approximately 14 percent of overall trade volumes to almost 25 percent by 2008, representing a compound growth rate of 13 percent. Traditional buy-side firms, who have thus far been slow to embrace algorithmic trading, represent the industry’s largest growth segment, with a five year compound growth rate of 30 percent.
I think they are low. I think growth in machine or algorithmic trading has entered a period of geometric growth and will see sustained and significant innovation for the next 5 to 10 years. Advances and ideas from areas such as econophysics and artificial intelligence will lead the way:
(From MoneyScience:) Econophysics was started in the late 1990’s by several physicists working in the subfield of statistical mechanics. They spontaneously decided to tackle the complex problems posed by economics, especially by financial markets. Unsatisfied with the traditional explanations of economists (many of which lacked empirical justification) they applied tools and methods from physics - first to try to match financial data sets, and then to explain more general economic phenomena.
(other sites on the subject include Econophysics Forum, Focus Session on Econophysics, American Physical Society, and Didier Sornette, UCLA)
(Thanks to the very interesting Vanilla Put blog for the post on econophysics.)
This rise of algorithmic trading will go hand in hand with the increasing sophistication and scale of electronic exchanges, ECNs and other computerized trading venues.
This is one area however where I would have to admit that the street is not standing still. For instance, just today Bank of America announced its purchase of (selected assets of) Financial Labs:
Financial Labs, LLC was founded in 2003 by a team of Harvard University-educated physicists and astrophysicists with extensive training in computational mathematics, and the managing partners of FX Solutions, one of the major firms in the global retail foreign exchange trading industry. The skills acquired by a scientist in analyzing large data sets and building quantitative models are similar to those needed to identify profitable trading opportunities and fleeting inefficiencies in financial markets; thus, a symbiotic combination of trained scientists with seasoned, professional Wall Street foreign currency traders paved the way for the formation of Financial Labs, LLC.
In a simlar vein, Citigroup acquired Lava Trading in the summer of 2004:
Lava Trading Inc. is an innovative technology firm that develops unique, high-performance trading solutions for the financial services industry. Our products are created as true ASP solutions built to withstand maximum volumes in the most volatile conditions. As a cost effective and neutral provider, our OTC, Listed and Foreign Exchange solutions are used by leading broker/dealers, including most of the top 20 U.S. investment banks, market makers, hedge funds and institutional investors.
The transition to more and more computerised electronic trading will fundamentally alter the financial ecosystem. Like with any significant environmental change, some species will not survive while others - including entirely new ones - will thrive. It creates a great opportunity for both firms and people, but for those unwilling or unable to adapt, the future is bleak.
January 31st, 2006 at 8:06 pm
Is Econophysics a scientific discipline mashup
There are three areas that I think are worth looking into or including in this type of discussion. The first you have mentioned is Game Theory.
The second is Neuroeconomics which applies economics, neuroscience and psychology to attempt to understand how we make choices. There is a website you can find by googling and a now dead blog. I also read a white paper whcih I found on the blog which I must admit didn’t seem to lead anywhere. Probably cos I didnt get it. Although it is another inter discipline Mashup
The last is chaos stuff particularly the work that Mandelbrot is doing published in his book The (Mis)behaviour of Markets: A Fractal View of Risk, Ruin and Reward which I read, understood a deal of and thought there was a lot of interesting stuff there to find and understand.
February 1st, 2006 at 10:32 pm
[…] In the descriptions above replace: betting with trading, sportsbook with orderbook, betting exchange with financial exchange, sporting odds with prices or quotes, Betfair with the name of any stock exchange or ECN…and you wouldn’t be able to tell the difference from a typical discussion of electronic and algorithmic trading tools and order management systems in the financial markets. Industry experts predict that across Europe betting exchanges by 2008 will match bets totaling USD 159 billion. Based on their current (2004) volume of USD 16 billion, this would mean a CAGR of 79%. […]
April 12th, 2006 at 7:26 am
Your blog is interesting and informative. Thank you
September 2nd, 2006 at 11:08 am
[…] Celent predicts that algorithmic-based equity trading will increase from approximately 14 percent of overall trade volumes to almost 25 percent by 2008. There are a few areas of immediate growth for the current technology - insurance, buy side. […]
October 24th, 2006 at 8:52 pm
Our firm just got our first international customer in the UK using our artificial intelligence and genetic algorithsm based trading product. We have many many years of research behind us and have used the technology for other markets and now moved into trading strategies. The results so far are extremely encouraging. It will be interesting to see how the market develops. www.aitellu.com
January 2nd, 2007 at 4:57 pm
there is at moment a big interest on trading algorithm, but so far I couln t find a single example (I mean with numbers and formulas)… maybe someone can help
January 8th, 2007 at 9:00 am
Amedeo, there is some information on wikipedia that you might find helpful; with for example the formula for one of the most common algorithms, VWAP. And lots of books available on amazon.com -hope this helps.