PureBytes Links
Trading Reference Links
|
Gary Funck wrote:
> I thought it might be interesting to look at the S&P cash
> index in terms of some other factors: (1) implied
> volatility (VIX) and (2) long rates (TYX). The attached
> charts are "scatter charts" that show the values of SPY
> (S&P Depository Receipts which trade at 1/10 the S&P 500
> cash index) for various values of VIX and TYX.
>
> These charts have a bit of a Rorschach character to them,
> but offer some insights, gained from the removal of the
> time element. (in fact, the time element is not completely
> removed because it is embedded in the ever-increasing price
> of the S&P). The time period covered is 10/29/1993 to
> 01/24/2000.
>
> I'm partial to the S&P as a function of VIX chart -- it
> shows the ever-increasing trend in volatility as the S&P
> price has marched higher.
NW: Nice looking chart Gary. (I could have sworn I saw an Indian Chief riding
a horse <G>)
But, the period you picked just happens to be the period that is the up period
in the 18 year cycle of volatility for the stock market. For example, I think
if you start your study in 1987, you will see VIX
declining steadily, while the market went higher, into 1994. This cycle tells
me that VIX should peak
in 2006. There is also a 11.5 year cycle for volatility which peaked last
year. This may cause the appearance of the big peak having been last year and
then followed by a high plateu as this two
cycles move toward counterbalancing each other. The bottomline is that I am
sure that you just happened to pick just the right time segment to totally
skew your data. You need a study that is much
lnger in time. If you retest, please let us know how it goes and if I was
right about this.
This brings to mind what is probably the most famous fatally flawed
study of all time. It was 1932,
Literay Digest was one of the best selling magazines in the country. They did
a telephone poll to see
who would win the Presidenttial election, Franklin Roosevelt or Alf Landon.
Their results showed that that Alf Landon would win by a landslide.
Obviiosly they were very wrong, as it was FDR who won by a landslide. . What
happened? The problem was in the media through which they took their polls. In
1932, in the depths of the Depression, mostly wealthy people who tended to
also be Republicans had telephones. Therefore, when the poll was conducted by
telehphone they mostly reached Republicans who were inclined to vote for their
candicate, Alf Landon. Needless to say, this blunder was very embarassing for
the magazine and they receeded into obscurity. I began college as a Polictial
Science major. This is one of the prime examples shown of what can go wrong in
poll and testing.
Well, I hope this helped. As you can see, I am not just "polling your leg".
<G>
Unskewingly,
Norman
> It also shows the wide range in
> implied (VIX) during the 1998 LTM bailout/crisis. You can
> see the periods of time that the S&P marched higher as
> rates and/or volatility dropped, and the current phase
> where the market has been heading higher along with
> volatility and long rates. The charts also show a fair
> amount of dispersion in S&P in the last few months,
> reflecting perhaps the fact that relationship of the S&P
> price to long rates and/or volatility has both diverged and
> become more sporadic.
>
> (SPY/VIX chart attached in this message, and SPY/TYX in the
> next.)
>
> ------------------------------------------------------------------------
> [Image]
|