While we have looked in the past at the incredible dominance of FOMC
days when it comes to stock market performance, recent intraday
performance of the major equity indices has had a somewhat repetitive
and rhythmic structure. We know volumes surge, pause, and surge;
Tradestation has dug one step deeper into the actual performance structure intraday and found some fascinating trends. From the
extremely
clear final-hour ramp to the oscillating bull-bear opening moves (and
the European close positive bias) across almost 30 years of price
behavior in bull and bear markets. The afternoons dominate market performance in bull markets and the morning session dominates the weakness in bear markets -
so
fade the opening rally, buy the dip, cover half into Europe, hope into
the close appears the 'empirical route of least resistance' - for now.
Active traders make their livelihood in the charts of the intraday
session, scanning the markets for recognizable patterns that are
persistent and profitable over time. However,
the intraday session is influenced by numerous factors.
For example, trading activity has been known to increase prior to and
after economic and earnings announcements. Developments in technical
analysis can also influence price momentum, market swings and trend
continuation. And then, of course, there’s always the completely
unforeseen event that throws the market completely out of whack. While a
certain degree of price movement will always be random, these and
countless other factors come together to create observable trading
biases. In this note below, the author will focus on trends and reversal
points in the intraday session, with the goal of identifying bullish
and bearish biases that active traders can put to use in their trading.
Intraday Bias Studies
In this section of the paper, intraday price trends of the
S&P 500 Index are spotlighted using data as far back as 1987. Some
of this information was conveyed in the March 8, 2011 Analysis Concepts
paper, “Mapping the Intraday Price Movement in the S&P 500 Index” (http://www.tradestation.com/education/labs/analysis-concepts/mapping-int...).
In this paper, a similar study is constructed from a finer interval
resolution (60 minute increments) with a variation in the construction
of return calculations. Another difference is that basic plus (+) and
minus (-) signs are used to depict whether the hour was positive or
negative in percentage terms. This creates a clearer visual
representation of the hourly trends that makes them easier to identify.
All results are created from average returns; these average returns are
calculated on an hourly interval but are generated from 30-minute bars
between 10 a.m. and 4 p.m., which includes pre- and post-market trading
(price changes from the 4 p.m. bar to the 10 a.m. bar).
At first glance in Table 2 (below), what stands out is the
number of positive periods at the 10 o’clock hour and in the 4 p.m.
hour, with the bulk of the returns from the 10 a.m. hour coming from the
pre-market session. The actual return from 9:30 a.m. to 10 a.m. is
positive, though Table 2 also shows a bullish bias in the 4 p.m. hour as
stocks make their way to the close. Going back to 1987, 21 of 25
occurrences had average returns that were positive for the 4 p.m.
interval. Also of interest is the weakness that typically occurs in the
11 a.m. hour (10 a.m. to 11 a.m.). Again, for data going back to 1987,
there were 18 occurrences where returns were negative for this interval.
The market seems, on average, to take a breather in the 11 a.m. hour
after its initial morning run-up.
Another
interesting statistic is that if stocks close higher on average into
the 3 p.m. hour, their probability of moving higher into the 4 p.m.
close is 70%.
Next, going back to September 11, 1984, trading biases in the S&P
500 Index intraday session are analyzed during longer-term bullish and
bearish market cycles. As mentioned earlier,
what really stands out in the data is a positive bias in the 4 p.m. hour of each bullish and bearish market cycle. Also, notice the
positive and negative biases in the 10 a.m. hour, correlated to each bull and bear market cycle.
Additionally, note that three of four bear market cycles had a negative
bias on average from the 10 a.m. hour into the 2 p.m. hour.
Bull and Bear Market Intraday Return Relationships
Depending on how one categorizes them, the
markets can experience cyclical periods of bull and bear runs for various lengths of time.
A more traditional approach is to classify these events in percentage
terms. Therefore, the rule applied here states that if the market
advances or declines by more than 20 percent, this will constitute a
bull or bear move. Price movement of this magnitude is recognized by
many financial market professionals as a change in market cycle.
Figure 7 (above) represents the
compounded total return of
the S&P 500 Index for the first, second, and third periods (9:30 to
11:40, 11:40 to 1:50, and 1:50 to 4:00) of the trading session within
each successive bull and bear market from 9/1/1983 to the present time.
In analyzing the data, the information is evident. First, the 9/1/1983
to 8/21/1987 and 12/4/1987 to 3/24/2000 bull markets, which occurred in
the first two decades of the data, had most of their returns formulated
from the last third of the trading session (1:50 to 4:00). At the same
time, the 10/4/2002 to 10/12/2007 bull market, along with the current
one, have had greater returns occur in the first third of the day's
session (9:30 to 11:40).
In Figure 8 (above), we can see that
in bull markets, the positive returns that the market experiences on average come from all three periods of the intraday session. However, the returns are
highest in the first (24.33 percent) and third (74.52 percent) periods, with the second period still being positive at 14.44 percent. We should point out the
return impact of the 268.84 percent in the third period of the 12/4/1987 to 3/24/2000 bull market. Even if we cut this number down by some factor, the returns are still significant for this period.
As we look at the sequence of returns in bear markets, they are also
very interesting. They typically start with painful selling in the first
third of trading, as Figure 9 (above) illustrates.
The average bear market return shows that from the 9:30 to 11:40 period, the return was -29.69 percent.
In bear market cycles, however, the market selling becomes less
pronounced as the day progresses. The second period of trading returned
-9.60 percent on average, while the third period returned -1.07 percent
on average.
So in bear market cycles, there seems to be some good
opportunity to either short early in the first third of the trading
session or buy on weakness somewhere in the last third of the session.