Half of what you see on the screen isn’t real
Suppose you see the S&P 500 Index (ES) futures
rise a full point in a short period of time on solid volume. Seeing the market
rally, you jump aboard, only to have the entire move retrace.
What happened?
In a nutshell, that move you saw on the screen
wasn’t real.
Yes, the index rose on buying, but what you
didn’t see on the screen was that the very same buyers of the index were selling
the S&P 500 Index ETF (SPY). Once the futures moved a full point higher,
exceeding fair value, they then sold the futures against a basket of stocks in
the index to bring the futures and cash back into line. The moves were real, in
the sense that they genuinely cost traders money. They’re not real, in the
sense that buying or selling did not reveal bullish or bearish tendencies on
the parts of large market participants.
According to
H. L. Camp, about 45% of all
NYSE volume is now attributable to program trading. What that means is that
the buying or selling you see on the screen may or may not reflect genuine
demand or supply in the marketplace.
Most readers are familiar with the NYSE TICK, a
measure of how many New York stocks are trading on upticks (at their offer
price) vs. downticks (at their bid price). Let’s say that, instead of measuring
the number of NYSE stocks trading at their offer prices vs. those trading at
their bids, we simply focus on the Dow 30 Industrial stocks and investigate how
many of them are trading bid vs. offer. The resulting statistic is called TIKI
and, it too, can be viewed as a sentiment measure. When Dow buyers are
aggressive, they will be willing to transact at the stocks’ offer prices, and
you’ll see TIKI values skyrocket above +20. When Dow sellers are aggressive,
they’re willing to bail out at the stocks’ bid prices and TIKI will plunge below
-20.
Because the Dow stocks are quite liquid and
trade frequently, the TIKI moves much faster than the NYSE TICK. Its values are
also distributed very differently from the TICK; TICK and TIKI correlate
significantly (around .60), but hardly perfectly. The majority of variance
in TIKI cannot be explained by the general buy/sell sentiment captured by TICK.
Something else, apart from general market buying or selling impacts TIKI
values.
The reason for this is that TIKI is highly sensitive to program trading.
Whenever a program is executed that calls for the simultaneous buying or selling
of a basket of stocks (arbing stocks against index futures would be a common
example), TIKI values will shoot very high or very low. The Dow stocks, being
liquid, are frequent components of such stock baskets. When the Dow stocks move
in unison, it is often because programs are being set off.
One way we know this is by looking at the distribution of TIKI values on a 10
second basis. The odds of a very high number of Dow stocks upticking or
downticking at exactly the same time should be quite small if we assume that
there is an even probability of the next tick being an uptick or downtick in
each issue. What we see, however, is many more extreme values than would be
predicted by chance. These bulges at the extreme are the result of systematic
buying and selling by institutions, often as part of arb (non-directional)
trade.
If you get that idea, then it will make sense to you that absolute TIKI values
are not especially helpful in gauging the sentiment of the market. TIKI can soar
or plunge, simply because institutions are buying or selling stocks at the same
time that they sell or buy index futures. It is the correlation
between TIKI and price that is crucial. When TIKI hits extremes
and price is moving in a correlated fashion, we know this is part of directional
trade–not arb.
So let us take a
moving correlation between TIKI and price change in the S&P 500 Index
(
SPY |
Quote |
Chart |
News |
PowerRating).
I have cumulated each day’s TIKI values, adjusted them for a zero mean, and
correlated TIKI and daily price change over a moving 10-day window going back to
2004 (N = 682 trading days).
The average 10-day correlation between daily TIKI and daily price change in SPY
over this period has been .63. When we have a strong TIKI/price correlation
(above .80; N = 108), the next ten days in SPY average a gain of .73% (73 up, 35
down). That is significantly stronger than the average 10-day price change in
SPY of .26% (397 up, 285 down).
When the TIKI/price correlation is relatively low (below .50; N =133), the next
ten days in SPY average a loss of -.41% (57 up, 76 down). That is significantly
weaker than the average 10-day performance.
What this suggests is that, when TIKI is well correlated with price, the market
tends to outperform. When TIKI is poorly correlated with price, the market tends
to underperform. This pattern, I have found, is also present at intraday time
frames. A reasonable explanation for the findings is that low correlation
periods represent occasions of high program/arb trading, whereas high
correlation periods represent periods of high directional trade.
We last saw very high TIKI/price change correlations on September 21 and 22,
when the values were about .88. The recent price strength in the equity markets
have followed from that. We are now at relatively average levels of correlation
(.60). Much of May and June–a period of correction–featured very low
correlations.
The moral of the story is that markets today are different than they were a
decade or two ago. You can’t necessarily trust what is on your screen or what
your price-based indicators are telling you. Who
is in the markets ultimately impacts what markets do.
Brett N. Steenbarger, Ph.D. is Associate Clinical
Professor of Psychiatry and Behavioral Sciences at SUNY Upstate Medical
University in Syracuse, NY and author of
The Psychology of Trading (Wiley, 2003). As Director of Trader
Development for Kingstree Trading, LLC in Chicago, he has mentored numerous
professional traders and coordinated a training program for traders. An active
trader of the stock indexes, Brett utilizes statistically-based pattern
recognition for intraday trading. Brett does not offer commercial services to
traders, but maintains an archive of articles and a trading blog at
www.brettsteenbarger.com and a
blog of market analytics at
www.traderfeed.blogspot.com. His book, Enhancing Trader Performance,
is due for publication this fall (Wiley).