Dec

12

The traditional explanation of a high put-call ratio was panic by the uninformed public. With the S&P 500 at a two-year high today, that scenario seems quite unlikely. Using another method to evaluate the little guys, the non-reportable positions in last week's COT report were net long 3.4% of the S&P 500 open interest (weighted average of bigs and e-minis). The non-reportable net long position has ranged from 0.9% to 5.2% in the past year.

Alston Mabry comments:

IMHO, the relationship between the ratio of leveraged-long instruments to leveraged-short instruments on the one hand, and the market on the other, went through a regime change this summer after the Flash Crash. To paraphrase Wally Shawn in The Princess Bride: "I don't think that means what I thought it meant."

William Weaver writes:

To pose another question to the list: What types of events can cause a regime change in a financial instrument?

To stick with equities, I use consumer spending data to define high vol and low vol regimes as I have found these fundamentals precede market action (not survey data).

Of my three business partners, two who have created models that best my own and use only price, volume, and open interest data.

Composition of OI? i.e. Commitment of Traders? Larry any thoughts on how a strategy may only work with one demographic instead of using the demographic as a signal itself? Maybe ICI data with fund holdings?

Sentiment data?

There was a company a few years back that suggested lookback periods consisted of all the data where a detrended version of price stayed within a standardized range, thus the last spike was the end of the prior regime. The flash crash could be a part of this.

Regulatory/mechanical changes: fractional to decimal, up-tick rule, naked short rules, no short rules, required reserves, etc.

So this gives us a few starting points: fundamentals, price/volume/OI related observations, demographic of a market, regulatory/mechanical, sentiment data and major dislocations in price (I believe this should be separated from the price category).

David Aronson replies: 

Will,

With regard to your question. We have recently added a type of modeling that searches simultaneously for an indictor to define distinct regimes (2 or 3) and then linear models that are optimal in each regime (distinct range of the regime indicator). I have not done much with it yet but we were motivated to add this tool because of the phenomena that you talk about in your post.


Comments

Name

Email

Website

Speak your mind

Archives

Resources & Links

Search