10
May
2007
Several years ago, when I was living in New Mexico, I had a girlfriend who used read very esoteric books about cutting edge theories on biology, evolution, and astronomy. We were talking about “what-if” scenarios one day and the conversation drifted to some called “punctuated equilibrium.” It was explained to me, at the time, that species evolve slowly in an environment that’s in equilibrium. A sudden leap in a species evolutionary development happens when a catastrophic internal or external event occurs. The example she used was the asteroid wiping out the dinosaurs theory. The dinosaurs lived and evolved in a relatively state of equilibrium until an asteroid killed them suddenly and allowed mammals to evolve rapidly.
Wikipedia defines it as:
Punctuated equilibrium (also called punctuated equilibria) is a theory in evolutionary biology, which states that most sexually reproducing species will show little change for most of their geological history. When phenotypic evolution occurs, it is localized in rare events of branching speciation (called cladogenesis), and occurs relatively quickly compared to the species’ full and stable duration on earth. [via wikipedia]
I understand that some of these theories have changed over the years but its premise stayed with me for years. Can the upset of punctuated equilibrium (PE) or something similar explain the sudden emergence or death of trends?
Although we like to believe in market equilibrium (PE?) and slow evolution of prices when new fundamentals occur, I’m a firm believer that the markets themselves are not always seeking equilibrium. Sentiment and fundamentals drive a trend and then the trend in turn drives the sentiment and fundamentals of that market. Trends become reinforcing and suck more and more capital into them until they crash. Then the crash becomes self reinforcing as the sentiment and fundamentals change and a new trend emerges downward. Where’s the equilibrium in that?
What truly interests me in trend following is the moment a financial asteroid hits the trend. What are the events or sudden changes in the financial environment that will allow some trends to die and cause others to evolve? Was it single event or several smaller events together that killed a trend or caused financial havoc? The first example that comes to mind was the Russian domestic debt default in the late 90’s that led to Long Term Capital Management (LTCM)’s sudden demise.
I know I don’t have all the answers, all I have is an interesting brain tease, and an interesting biological theory that I’m trying to superimpose on existing trends, hoping to uncover future financial asteroids. ![]()
10 Responses
Ashton
May 10th, 2007 at 10:16 am
1Interesting post…I think your “financial asteroid” theory is similar to Nassim Taleb’s “Black Swan”. The type of events you’re referring to are outliers, something MPT fails to recognize, which is party why these events have such a great impact. Taleb’s hedge fund has been very successful in trading these types of extreme events. Your quest to uncover financial asteroids may well turn out to be rewarding.
Tom
May 11th, 2007 at 12:54 pm
2Ashton: Thanks for stopping by. Yes its those pesky outliers cause a lot of havoc, how often do we just ignore them? Often in data mining we just clip them out of the data as they skew the output. Well the models work great in an equilibrium state but is that reality? The reality is that we will have these “financial asteroids” every so often.
I’ve read a few articles from Mr. Taleb and find him to be interesting and insightful. I plan on reading his new book, “Black Swan.” Have you checked out Mandelbrot’s, “The (Mis)Behavior of Markets?” Its along similar lines that these “asteroids” are part of the overall market organism.
Ashton
May 14th, 2007 at 10:27 am
3I’m a huge fan of Mandelbrot’s work & think that’s one of the best books on the markets. I prefer Taleb’s “Fooled by Randomness” to “The Black Swan” as its focus is more relevant to trading. Highly recommended if you haven’t read it yet. You can find some interesting articles/webcasts by Taleb (& others) on http://www.wilmott.com
Tom
May 14th, 2007 at 3:40 pm
4I was introduced to Mandelbrot when I read the Chaos Theory book by Glick and his work at the time focused on weather patterns. I devoured (Mis)Behavior when it came out, fantastic book and its the reason my focus shifted to quantitative and then neural net modeling.
Thanks for the Wilmott link, I’m sure I’ll disappear for a few weeks now!
muxControl
May 15th, 2007 at 3:58 pm
5The demise of LTCM was a disaster waiting to happen. It could have been (and was i think) predicted by some. Lets look at 9/11. Could that and its effects have been predicted? And the recent mass panic in China which rocketed the FTSE, blowing £80bn off its value. Its hard to imagine that the lead up to these type of events can be detected in a time-series. What information would a time-series have regarding the plans of some al-qua-eda guy wanting to blow something up, thereby causing astriod like aftermath?
Can all types of financial astriods be detected?
Tom
May 15th, 2007 at 4:51 pm
6MuxControl: If you would be so kind and point me in the direction of an article about the prediction of the LTCM demise, I’d really appreciate it. Anyone dealing with the amount of leverage LTCM had at its height was asking for trouble, if you ask me.
Some asteroids can be spotted ahead of time, like the insane activity in the Shenzhen. It’ll probably end in tears with some ripple effect worldwide. I don’t know what the extent of the damage would be worldwide but it could turn into a financial asteroid alright and kill a lot trends. Other financial asteroids probably can’t be spotted until its too late.
Time series allows you see the trends, the asteroids hitting them allows you see their demise and ripple effect.
muxControl
May 16th, 2007 at 8:30 am
7So then isn’t it easy to confuse early indication of a financial astriod with the time series changing from trend mode to “non-trend” mode? Because it seems that the onset of both appear to have similar characteristics i.e. correlations become less linear in the transition hence the resulting outlier points.
Tom
May 16th, 2007 at 9:06 am
8Mux: The trick is to how to quantify the event into a trend smashing or minor disruptive occurrence. I’m still working on that one!
Kerwin
May 19th, 2007 at 8:51 am
9Looking at trendfollowing in the forex market, i would say we get a financial asteroid at 2 times a week, when the market responds to data about inflation, production, nfp etc. As a trendfollower i have to close out my positions ahead of this data as the trend can either be turned or exacerbated on the data. This puts the time series non-stationary, which makes most backtest using TA null and void during this period of volatility.
Calculating Historical Volatility | Neural Market Trends
December 29th, 2007 at 12:43 pm
10[...] IVUnderstanding Fuzzy Trend Following in ExcelBuilding an AI financial market model - Lesson VFinancial Asteroids & Their Effects on TrendsMonte Carlo Simulations For S&P500 Volatility Timing [...]
RSS feed for comments on this post · TrackBack URI
Leave a reply
previous post: Algorithmic Trading
next post: $100 Forex Experiment - Still Trapped
to top of page...