When QQQQ Volume Hits a Two-Month Low, Plus an Introduction to the VIX:VXV Ratio

Bill Luby over at VIX & More was one of the first to discuss a new product
launched by the CBOE at the end of last year - VXV, a variation of the popular
VIX that measures implied volatility for S&P options three months out instead
of one month. He first discussed this new volatility index in his December
10th column, where he provides a good overview of what the VXV is intended to
accomplish and how it could be utilized as an indicator, particularly in
relation to the VIX. As the CBOE itself noted in its introductory paper on the
VXV... ?The behavior of VXV relative to VIX illustrates the mean-reverting
properties of volatility and suggests that the slope of the 1- to 3-month SPX
implied volatility term structure could be used to predict future levels of
near-term (1-month) implied volatility..." The CBOE goes on to note that "When
the VXV was higher than 1-month VIX by more than 10%, the average closing VIX
level over the following 20 trading days was, on average, higher by at least
5%. " And the converse was also true - that is, when the VIX was significantly
higher than the VXV, the VIX typically traded lower over the next month. Bill
neatly summarized this in a follow-up column on December 11th.... "I expect
the VIX to have a tendency to fall and the markets to rise when the VIX:VXV
ratio is above 1.10; and will look for the VIX to rise and the markets fall
when the ratio is below 0.90, in classic mean reversion fashion."

At the time, there hadn't been any spikes above 1.10 or below .90, but since
that December column there have been two occurrences (see VIX-VXV Ratio
chart). And both played out as expected. The ratio fell below .90 on December
20th, leading to higher volatility and a lower stock market twenty trading
days later. And on January 22nd, the ratio spiked over 1.10, indicating the
VIX was 10% above the VXV. That led to falling volatility and rising stock
prices over the next twenty days. We almost saw another spike over 1.10 on
March 17th when the ratio hit 1.09, which as it turned out also led to lower
volatility/higher stock prices one month later. The sample size is not
statistically significant, but based on the CBOE's statement that essentially
backs up Bill's conclusions (and which was based on a study stretching back to
2002), it definitely seems this is a new indicator worth watching. We'll be
adding this to the charting & data sections of the site for daily monitoring.

Not much new to report following Tuesday's low-volume selloff. Institutions
have only been sporadically involved over the last few weeks, and only on the
big up days (3/18, 3/20, 3/24, 4/1). When they're not around, as is the case
currently, we get the kind of drift lower that started Tuesday. Most likely
this will continue over the near-term. QQQQ volume hit its lowest level in
over two months Tuesday, typically a negative sign for the short-term. The
absence of institutional firepower usually translates into a subsequently
lower QQQQ close within the next three sessions. Since 1999, this setup has
maintained an 85% accuracy level (39/46), well above the 65% at-any-time odds.
The last thirty occurrences are noted below...

QQQQ volume hits its lowest level in Forty Trading Days
04/08/08... QQQQ ???
12/24/07... QQQQ lower close two days later
11/23/07... QQQQ lower close one day later
10/08/07... QQQQ lower close three days later
10/02/07... QQQQ lower close one day later
07/03/07... QQQQ no lower close within three days
05/07/07... QQQQ lower close three days later
04/05/07... QQQQ lower close one day later
03/23/07... QQQQ lower close two days later
12/26/06... QQQQ no lower close within three days
11/24/06... QQQQ lower close one day later
08/31/06... QQQQ lower close three days later
07/03/06... QQQQ lower close one day later
05/08/06... QQQQ lower close one day later
03/27/06... QQQQ lower close one day later
11/25/05... QQQQ lower close one day later
11/14/05... QQQQ lower close one day later
09/12/05... QQQQ lower close one day later
08/19/05... QQQQ lower close two days later
05/27/05... QQQQ lower close one day later
11/26/04... QQQQ lower close one day later
10/11/04... QQQQ lower close one day later
08/26/04... QQQQ lower close two days later
04/12/04... QQQQ lower close one day later
11/28/03... QQQQ lower close three days later
08/15/03... QQQQ no lower close within three days
05/23/03... QQQQ no lower close within three days
03/28/03... QQQQ lower close one day later
12/24/02... QQQQ lower close one day later
11/29/02... QQQQ lower close two days later
11/15/02... QQQQ lower close one day later

Categories : Archived EOD Commentary, End-of-Day Commentary

 Print This Post