|
WiserTrader Monthly |
|
|
__________________________________________________________________________
July 2007 Prescott, Arizona Contact Editor __________________________________________________________________________
|
|
|
Subscribe to the Weekly Newsletter
Published the 3rd Tuesday of the month
|
|
___________________________________________________________________________ Financial markets are an emotional place. Without a feel for market sentiment, anyone can fall victim to the stressful emotions that make 90% of all trades lose money at precisely the worst possible moment. For those in long positions, positive financial news is met with greed or complacency while negative news is met with caution or fear. For those in short positions, the opposite is true. Changes from one mode to the other are signaled by large swings in the VIX, VXN and put-call ratios. However, when market news is inconclusive, it is accompanied by uncertainty, indecisiveness and continuous volatility. This turbulent condition is signaled when stock and industry price movements are highly correlated with the major averages. ___________________________________________________________________________
Sample Correlation Coefficient
Some of the best sentiment indicators depend on polling investors, brokers, traders, money managers, hedge fund managers, technical and fundamental analysts, and very savvy market players, each with their own particular expertise in the markets. It is the equivalent to the perfect market diversification, only in terms of backgrounds, interests, and approaches. Financial news commentators who actually seek opinions from to floor traders are also worth listening to. A very curious oscillatory condition occurs when market participants are uncertain. The market will move sideways in a choppy fashion as stocks follow the major indices very closely. It is as if everyone is watching everyone else because maybe someone else knows what’s going on! Under these conditions stocks and industries follow the major averages very closely with a high degree of correlation.
When the correlation between industries, stocks and the major averages is perfect the correlation coefficient equals 1. When they are perfectly out of sync, the correlation coefficient is -1. This can be expressed as a percentage from -100% to +100%. One way to measure the correlation between stocks, industries and the major averages is to compare their William-%R values over a number of days. This avoids comparing very large price values say for the DOW, with a variety of large and small numbers for stocks and industry composite values. Such comparisons are thus restricted to values ranging from 0 to 100.
One must be consistent in the choice of the number of days, the period for which the indicator is measured and also the kind of comparison being made. In this work, the 10-period Williams-%R is compared over 11 days. That means 11 pairs of values are compared (11 for a stock or industry and 11 for the major averages). Ideally, it is best to use a William-%R period that is about half the periodic cycle. The 10-period Williams-%R is arbitrarily chosen in order to make a consistent comparison. Eleven days worth of data was chosen to fully encompass all the Williams-%R data for the pervious day. When calculating the correlation with the major averages, one can either (1) calculate it for each stock or industry and then combine the correlations into an average value or (2) average the Williams-%R values for the stocks or industries and then calculate the correlation based on their averages. The first method gives a wider range of possible values (-100% to +100%) that averages around 35% or less when the markets are calm and rising to 40% to 80% when the markets are highly volatile. The second method produces a consistently high value over a narrower range (0 to +100%). Therefore the first method is used in order to obtain a wider dynamic range of values.
Over the past month the 215 industry correlation with the major averages has ranged from 40% to 65% while the market has been extremely volatile, swinging back and forth every two weeks or so. As of this writing the industry correlation coefficient is about 36%, still a little on the volatile side.
Formally, the correlation coefficient is actually being approximated by taking a biased sample. We do not perform a test of significance or construct a confidence interval. We are making only a relative comparison. Hence it is called the sample correlation coefficient for two samples if data x and y. Let us suppose that the n = 11 Williams-%R values for stocks or industries are represented by sample x and n = 11 William percent-%R values for the major averages are represented by sample y. Then the sample correlation coefficient r is given by, *
Where,
This calculation is installed in the industries spreadsheet for the correlation between 215 industries and the major averages (DOW, NASDAQ Composite, S&P 500, Russell 2000 and NYSE Composite).
|
Table 1A
Leading Industries
July 16, 2007
|
Rank |
Relative To Avg. Position |
Rank of Position Change |
Industries |
Companies |
Last Gain |
1 wk Ago Gain |
1 mo Gain |
|
1 |
63 |
16 |
0.0% |
9.5% |
13.8% |
||
|
2 |
7 |
92 |
0.0% |
6.0% |
12.1% |
||
|
3 |
9 |
86 |
0.0% |
3.6% |
11.9% |
||
|
4 |
23 |
60 |
0.0% |
2.4% |
10.9% |
||
|
5 |
81 |
9 |
0.0% |
5.2% |
10.4% |
||
|
6 |
35 |
40 |
0.0% |
4.5% |
10.3% |
||
|
7 |
25 |
55 |
0.0% |
-1.0% |
10.0% |
||
|
8 |
-2 |
115 |
0.0% |
0.4% |
9.7% |
||
|
9 |
-6 |
124 |
0.0% |
0.7% |
9.7% |
||
|
10 |
4 |
97 |
0.0% |
1.7% |
9.5% |
||
|
11 |
52 |
26 |
0.0% |
2.9% |
9.2% |
||
|
12 |
23 |
63 |
0.0% |
1.3% |
8.9% |
||
|
13 |
39 |
37 |
0.0% |
1.5% |
8.9% |
||
|
14 |
28 |
51 |
0.0% |
0.7% |
8.8% |
||
|
15 |
3 |
102 |
0.0% |
3.0% |
8.8% |
||
|
16 |
-7 |
128 |
0.0% |
-0.8% |
8.7% |
||
|
17 |
105 |
6 |
0.0% |
2.4% |
8.4% |
||
|
18 |
41 |
34 |
0.0% |
1.1% |
8.1% |
||
|
19 |
39 |
36 |
0.0% |
3.3% |
7.7% |
||
|
20 |
10 |
84 |
0.0% |
1.0% |
7.4% |
||
|
21 |
72 |
12 |
0.0% |
0.3% |
7.3% |
||
|
22 |
80 |
10 |
0.0% |
4.6% |
7.1% |
||
|
23 |
7 |
90 |
0.0% |
1.8% |
6.9% |
||
|
24 |
70 |
14 |
0.0% |
0.4% |
6.9% |
||
|
25 |
-9 |
130 |
0.0% |
1.2% |
6.9% |
||
|
26 |
53 |
24 |
0.0% |
3.9% |
6.7% |
||
|
27 |
99 |
7 |
0.0% |
0.1% |
6.7% |
||
|
28 |
-1 |
110 |
0.0% |
1.0% |
6.6% |
||
|
29 |
43 |
33 |
0.0% |
-2.0% |
6.6% |
||
|
30 |
68 |
15 |
0.0% |
4.1% |
6.6% |
||
|
31 |
29 |
48 |
0.0% |
3.9% |
6.3% |
||
|
32 |
8 |
87 |
0.0% |
0.8% |
5.9% |
||
|
33 |
53 |
22 |
0.0% |
-2.6% |
5.7% |
||
|
34 |
45 |
32 |
0.0% |
3.6% |
5.7% |
||
|
35 |
114 |
5 |
0.0% |
1.8% |
5.7% |
||
|
36 |
27 |
52 |
0.0% |
1.6% |
5.6% |
||
|
|