Wednesday, April 29, 2009

VIX Asset Allocation Models

ADDED: For those who land on this post from a google search, this topic is discussed in more detail in other posts which are linked in this Gateway Post:USING THE VIX MODEL AS A TIMING INDICATOR FOR LONGER TERM STOCK ALLOCATIONS

These recent posts may also be of some interest:
A practical application of the model to the S & P 500 in 2007 is discussed in this post:VIX Chart from 2007: Alerts and Triggers Major Disruption of Cyclical Stable Bull VIX Pattern
****************************Original Post:

I noticed from my FeedJit widget some interest from Germany about my discussions on using the movement in VIX as a factor in asset allocation decisions. VIX is the CBOE volatility index for the S & P 500. My discussions on this topic are scattered throughout this blog, and I may have time later to bring them into a one post summary.  The question is whether I want to do it or not. This would be a far more extensive exercise than  the recent posts on ING preferred stocks and the TC containing a Hertz bond.  

For now, I would just summarize a few key points.  The model was first developed as an outgrowth of my concern, which increased in intensity during 2007, about the impact of an improvident extension and expansion of credit in the U.S.  The first real time test of the model came in August 2007 when the first warning signal was flashed by the VIX moving decisively out of the stable bull market range between 10 and 20.  The model worked, though the model does not tell an investor how much to change the asset allocation which is up to each individual investor based on their unique characteristics, risk tolerance and needs . I divided the bull market pattern into two phases. Phase 1 is movement within a range of 15 to 20 in the VIX, and Phase 2 is a range bound 10 to 15.  The model gave two signals in 2007 to reduce stock exposure, and I followed it to a degree, probably saving myself a six figure sum.   

After developing the VIX model, I started working on one for the DJIA and Russell 2000.  The one for the DJIA showed a somewhat less volatile and more stable index than the S & P 500, but otherwise confirmed my observations about the VIX.  The symbol for the CBOE DJIA volatility index is VXD ( or ^VXD at Yahoo Finance).   I then backtested both models with all of the data that I had from the CBOE and found that it would have worked during the 1990s.  It does not work perfectly, but the past has shown so far that it provides a general direction for cyclical bull and bear markets.     The models on the Russell 2000 and Nasdaq Indexes had to be adjusted since the first observations of those volatility indexes revealed overall less stability than the VIX.  Based on my initial observations of the CBOE volatility indexes for both the Nasdaq and Russell 2000 indexes, I concluded that the model needs to be adjusted based on each index due to inherent differences in the stability of the indexes.  The new indexes for gold and oil will ultimately require even more extensive adjustments which I can not make due to the lack of sufficient historical data.GOLD OIL and EURO CBOE VOLATILITY INDEXES. The symbol for Nasdaq is VXN (^VXN at Yahoo Finance) and  RVX  for the Russell 2000 (^RVX).

Some of the links discussing why these models may work and a general discussion of how they would be applied are as follows:

There is volatility data on the S & P 100 that goes back to 1986 which would have caused under the model an asset allocation change in 1987 before the crash in October.  Trading and Asset Allocation in Stable and Unstable VIX Pattern 
A forced reduction in the stock allocation, with the amount up to each investor under the model, would have occurred at anytime during the period between  July 7, 1987 to August 10, 1987 caused by a break in the stable bull pattern in April.  

A certain amount of skepticism is needed with any trading model, and LB views them as a guide rather than a hard and fast rule.  They are also in need of constant revision based on observed current information which may require an adjustment or even an abandonment of the model altogether.   The models have so far been adjusted to permit certain kinds of trading activity during an unstable VIX pattern that the initial model prohibited.  RB calls all of the models "stinking models" and ignored all of them after its coup d'etat on 3/3/09.  LB is back in control now so the model is being followed after a three week buying binge by RB.  The model now permits buying common stocks only with cash flow from interest and dividends, reinvestment of dividends to buy more shares, and using the proceeds of one sale to buy another.  Further, as an exception to those exceptions, cash flow can not be used now until the S & P 500 closes a month above 815 which it will probably do on 4/30.  

I may come back to this topic at some point later.  While the VIX has been moving down, which is a good sign, it is still in an unstable pattern as characterized by the model. 

I suspect that Novartis (NVS), a recent buy, rose yesterday due to its likely involvement in developing a vaccine for the new swine flu strain.  BusinessWeek Personally, I would not buy NVS for that reason.  Novartis is one of the 3 drug manufacturers that make flu vaccines.  More confirmed swine flu cases were reported yesterday, and the experts have given up on what they refer to as containment.  As of this morning, there were no confirmed deaths in the U.S.  Yahoo! Finance

Yesterday, I talked about the junk rated bond from Hertz.  This article from a RealMoney contributor currently available on Yahoo Finance without a subscription discusses a few other bonds including FCY which I own.FCY: Odd lot limit order filled  FCY: Forest City Enterprises Senior Bond (FCY)/FCZIt is treated like the Hertz bond by LB.  I also own Prospect Energy mentioned by this columnist.  I was aware of the CBS bond maturing in 2056 but will not buy it due to the maturity date. I was not aware of the GAP bond maturing in 2039 so I will look into it.   Yahoo! Finance  Some people call bonds traded on the stock market as exchange traded bonds.  Buying them is easier than making a purchase for a similar security in the bond market, and an investor can buy just a few share if so desired.  I could never go into the bond market and buy $500 worth of a bond.  My orders for 5 $1000 bonds are mostly ignored. 

One research technique that I use is to have several portfolios at Yahoo Finance.  I subscribe to their real time quote service but that is not really necessary.  Each portfolio can have up to 200 securities in it, so I have two to cover my existing holdings.  This gives me real time information on all positions, and the Marketracker service which gives me flashing green and red lights as bid, ask, and transaction prices change .  More importantly for me, it aggregates the news for all of them in one place, which can be done for free by anyone.   So, as an example, I noticed today that a few of my holdings will report earnings so I am now alerted to do some reading later today.   I then have several watch portfolios, divided into types of securities (Bonds, ETFs, Closed End Funds, and Stocks).  Within the stock category, there is a further division into small, mid cap and large cap.  Then there are a few unique type of categories.  For those who know me, I take investing seriously, and I will generally start working on it at least two hours before the market opens.

No comments:

Post a Comment