Recent Herding Tendency Evident in the 20-day New High/Low Data
By
Rennie on Tuesday, May 5th, 2009 at 2:29 am
Stocks closed with solid gains Monday as institutional sellers remained conspicuously absent. TICKscore settled at +21, Cumulative TICK +103,000. Breadth closed 5:1 positive, sending the Cumulative Breadth line well above its January high. Note that it actually surpassed that level just over a week ago on April 24th. New 20-day highs swamped new 20-day lows by a huge 13:1 margin. Putting together a New High-Low Index (new highs/(new highs + new lows)) utilizing 20-day data as opposed to the standard 52-week data yields a high 93% reading Monday. That’s the third 90%+ reading in the past week and the 18th reading above 90% since the start of 2009. Eighteen is a remarkably high number of occurrences considering that there were only TWO DAYS in the four-year period from 2005 through 2008 when this statistic closed above 90%. This is a good illustration of the herding tendency that’s become especially prevalent on the upside over the past two months (15 of the 18 occurrences were in the last forty trading days).
In the table below I’ve listed each date when the raw 20-day High-Low Index closed over 90%, followed by the date when the High-Low Index closed back under 90% and the performance of the S&P500 in this time frame. Nine cases isn’t much to go on, but it does suggest the market’s a bit extended on a short-term basis…
20-Day High-Low Index (raw daily number) >90%
05/04/09 – ??/??/??… S&P500 ???
04/29/09 – 05/01/09… S&P500 +0.4%
04/16/09 – 04/20/09… S&P500 -3.8%
04/09/09 – 04/14/09… S&P500 -1.8%
04/02/09 – 04/07/09… S&P500 -2.3%
03/23/09 – 03/30/09… S&P500 -4.3%
03/19/09 – 03/20/09… S&P500 -2.0%
01/02/09 – 01/07/09… S&P500 -2.7%
01/09/06 – 01/10/06… S&P500 -0.0%
07/11/05 – 07/12/05… S&P500 +0.2%
Monday’s rally sent the OEX Volatility Index (VXO) down to new lows for the year and a close under its lower bollinger band. Unlike a stock, volatility generally won?t remain below its lower band for more than a couple of sessions before rallying. Given the inverse relationship between volatility and stocks, rebounding volatility tends to correspond with a flat-to-down market. Listed below are the last thirty instances when the VXO closed under it?s lower band…
VXO Closes Under its Lower Bollinger Band
05/04/09? OEX ??? three sessions later
04/17/09? OEX -3.4% three sessions later
04/09/09? OEX -0.6% three sessions later
12/19/08? OEX -1.7% three sessions later
05/19/08? OEX -2.5% three sessions later
02/25/08? OEX -0.1% three sessions later
12/21/07? OEX -0.5% three sessions later
09/21/07? OEX +0.1% three sessions later
11/17/06? OEX +0.2% three sessions later
08/16/06? OEX +0.5% three sessions later
11/03/05? OEX +0.1% three sessions later
09/09/05? OEX -1.1% three sessions later
06/21/05? OEX -1.9% three sessions later
04/12/05? OEX -3.2% three sessions later
02/04/05? OEX -0.5% three sessions later
02/01/05? OEX +1.2% three sessions later (*)
12/21/04? OEX +0.1% three sessions later
06/08/04? OEX -1.2% three sessions later
07/25/03? OEX -1.1% three sessions later
05/16/03? OEX -2.2% three sessions later
04/17/03? OEX +3.0% three sessions later (*)
04/14/03? OEX +0.9% three sessions later
11/20/02? OEX +2.1% three sessions later (*)
11/15/02? OEX +0.7% three sessions later
12/27/01? OEX -0.3% three sessions later
12/19/01? OEX -0.6% three sessions later
10/26/01? OEX -4.2% three sessions later
05/21/01? OEX -1.5% three sessions later
01/23/01? OEX -0.4% three sessions later
08/08/00? OEX -0.2% three sessions later
06/29/00? OEX +0.5% three sessions later
In 19 out of 30 cases, or 63% of the time, the S&P500 closed at a lower level three days later, well above the 47% at-any-time odds for a lower S&P three days later. While not quite accurate enough to make the cut for a sell signal, note that in the majority of cases this setup has translated into limited upside potential over the next few sessions. In only three cases out of thirty, or 10% of the time, was the OEX up 1% or more three days later, while it was down 1%+ a total of eleven times.
Technical oscillators are also suggesting the market is at or near a short-term top. The NYSE McClellan topped +200 Monday and will trigger a short-term sell at Tuesday’s close if it remains above +150. For that to occur, breadth would have to close anywhere north of -200. Also, the 2-day RSI for the S&P500 closed over 98 Monday, which in recent
years has been a good indication that a top is nearby. We’ve seen this signal triggered a total of 22 times since 1999, 19 of which led to a lower S&P close by the time the 2-day RSI closed back under 98…
S&P500 2-Day RSI Closes Over 98
05/04/09… S&P500 ??? when RSI <98
01/02/09… S&P500 -0.5% when RSI <98 (one day)
04/10/07… S&P500 -0.7% when RSI <98 (one day)
10/26/06… S&P500 -0.4% when RSI <98 (two days)
03/17/06… S&P500 -0.2% when RSI <98 (one day)
11/22/05… S&P500 -0.3% when RSI <98 (three days)
09/30/05… S&P500 -0.2% when RSI <98 (one day)
07/15/05… S&P500 -0.6% when RSI <98 (one day)
06/17/05… S&P500 -0.1% when RSI <98 (one day)
11/03/04… S&P500 +1.9% when RSI <98 (three days)
05/25/04… S&P500 +0.7% when RSI <98 (three days)
04/05/04… S&P500 -0.2% when RSI <98 (one day)
09/02/03… S&P500 -0.1% when RSI <98 (three days)
06/04/03… S&P500 +0.2% when RSI <98 (two days)
03/19/03… S&P500 -1.0% when RSI <98 (three days)
05/21/01… S&P500 -0.3% when RSI <98 (one day)
04/19/01… S&P500 -0.9% when RSI <98 (one day)
08/08/00… S&P500 -0.7% when RSI <98 (one day)
07/14/00… S&P500 -1.1% when RSI <98 (two days)
12/31/99… S&P500 -1.0% when RSI <98 (one day)
12/23/99… S&P500 -0.1% when RSI <98 (one day)
07/02/99… S&P500 -0.2% when RSI <98 (one day)
04/12/99… S&P500 -0.7% when RSI <98 (one day)
Prior to 1999 this signal wasn’t nearly as effective as it’s been in recent years.
Finally, it should be noted that Monday represented the fifth consecutive day of positive NYSE breadth. In recent years, such a long string of positive breadth days has been followed by generally toppy conditions looking out 3-6 trading days. In 28 out of the last 30 separate occurrences, the S&P posted a close below the setup day 3-6 trading days later, well ahead of the 66% at-any-time odds…
Breadth Positive Five in a Row
05/04/09? ???
03/16/09? No lower S&P close in 3-6 sessions
01/06/09? Lower S&P close three sessions later
11/28/08? Lower S&P close three sessions later
11/03/08? Lower S&P close three sessions later
07/22/08? Lower S&P close three sessions later
05/16/08? Lower S&P close three sessions later
04/04/08? Lower S&P close three sessions later
08/23/07? Lower S&P close three sessions later
07/03/07? Lower S&P close four sessions later
06/01/07? Lower S&P close three sessions later
05/07/07? Lower S&P close three sessions later
04/04/07? Lower S&P close four sessions later
03/23/07? Lower S&P close three sessions later
02/01/07? Lower S&P close six sessions later
01/12/07? Lower S&P close three sessions later
11/16/06? Lower S&P close six sessions later
08/30/06? Lower S&P close four sessions later
08/18/06? Lower S&P close three sessions later
03/16/06? Lower S&P close three sessions later
01/09/06? Lower S&P close three sessions later
12/14/05? Lower S&P close three sessions later
11/23/05? Lower S&P close three sessions later
06/15/05? Lower S&P close six sessions later
06/02/05? Lower S&P close three sessions later
02/16/05? Lower S&P close three sessions later
12/14/04? Lower S&P close three sessions later
11/15/04? Lower S&P close three sessions later
10/29/04? No lower S&P close in 3-6 sessions
10/04/04? Lower S&P close three sessions later
08/02/04? Lower S&P close three sessions later
Recent Herding Tendency Evident in the 20-day New High/Low Data
By Rennie on Tuesday, May 5th, 2009 at 2:29 amStocks closed with solid gains Monday as institutional sellers remained conspicuously absent. TICKscore settled at +21, Cumulative TICK +103,000. Breadth closed 5:1 positive, sending the Cumulative Breadth line well above its January high. Note that it actually surpassed that level just over a week ago on April 24th. New 20-day highs swamped new 20-day lows by a huge 13:1 margin. Putting together a New High-Low Index (new highs/(new highs + new lows)) utilizing 20-day data as opposed to the standard 52-week data yields a high 93% reading Monday. That’s the third 90%+ reading in the past week and the 18th reading above 90% since the start of 2009. Eighteen is a remarkably high number of occurrences considering that there were only TWO DAYS in the four-year period from 2005 through 2008 when this statistic closed above 90%. This is a good illustration of the herding tendency that’s become especially prevalent on the upside over the past two months (15 of the 18 occurrences were in the last forty trading days).
In the table below I’ve listed each date when the raw 20-day High-Low Index closed over 90%, followed by the date when the High-Low Index closed back under 90% and the performance of the S&P500 in this time frame. Nine cases isn’t much to go on, but it does suggest the market’s a bit extended on a short-term basis…
20-Day High-Low Index (raw daily number) >90%
05/04/09 – ??/??/??… S&P500 ???
04/29/09 – 05/01/09… S&P500 +0.4%
04/16/09 – 04/20/09… S&P500 -3.8%
04/09/09 – 04/14/09… S&P500 -1.8%
04/02/09 – 04/07/09… S&P500 -2.3%
03/23/09 – 03/30/09… S&P500 -4.3%
03/19/09 – 03/20/09… S&P500 -2.0%
01/02/09 – 01/07/09… S&P500 -2.7%
01/09/06 – 01/10/06… S&P500 -0.0%
07/11/05 – 07/12/05… S&P500 +0.2%
Monday’s rally sent the OEX Volatility Index (VXO) down to new lows for the year and a close under its lower bollinger band. Unlike a stock, volatility generally won?t remain below its lower band for more than a couple of sessions before rallying. Given the inverse relationship between volatility and stocks, rebounding volatility tends to correspond with a flat-to-down market. Listed below are the last thirty instances when the VXO closed under it?s lower band…
VXO Closes Under its Lower Bollinger Band
05/04/09? OEX ??? three sessions later
04/17/09? OEX -3.4% three sessions later
04/09/09? OEX -0.6% three sessions later
12/19/08? OEX -1.7% three sessions later
05/19/08? OEX -2.5% three sessions later
02/25/08? OEX -0.1% three sessions later
12/21/07? OEX -0.5% three sessions later
09/21/07? OEX +0.1% three sessions later
11/17/06? OEX +0.2% three sessions later
08/16/06? OEX +0.5% three sessions later
11/03/05? OEX +0.1% three sessions later
09/09/05? OEX -1.1% three sessions later
06/21/05? OEX -1.9% three sessions later
04/12/05? OEX -3.2% three sessions later
02/04/05? OEX -0.5% three sessions later
02/01/05? OEX +1.2% three sessions later (*)
12/21/04? OEX +0.1% three sessions later
06/08/04? OEX -1.2% three sessions later
07/25/03? OEX -1.1% three sessions later
05/16/03? OEX -2.2% three sessions later
04/17/03? OEX +3.0% three sessions later (*)
04/14/03? OEX +0.9% three sessions later
11/20/02? OEX +2.1% three sessions later (*)
11/15/02? OEX +0.7% three sessions later
12/27/01? OEX -0.3% three sessions later
12/19/01? OEX -0.6% three sessions later
10/26/01? OEX -4.2% three sessions later
05/21/01? OEX -1.5% three sessions later
01/23/01? OEX -0.4% three sessions later
08/08/00? OEX -0.2% three sessions later
06/29/00? OEX +0.5% three sessions later
In 19 out of 30 cases, or 63% of the time, the S&P500 closed at a lower level three days later, well above the 47% at-any-time odds for a lower S&P three days later. While not quite accurate enough to make the cut for a sell signal, note that in the majority of cases this setup has translated into limited upside potential over the next few sessions. In only three cases out of thirty, or 10% of the time, was the OEX up 1% or more three days later, while it was down 1%+ a total of eleven times.
Technical oscillators are also suggesting the market is at or near a short-term top. The NYSE McClellan topped +200 Monday and will trigger a short-term sell at Tuesday’s close if it remains above +150. For that to occur, breadth would have to close anywhere north of -200. Also, the 2-day RSI for the S&P500 closed over 98 Monday, which in recent
years has been a good indication that a top is nearby. We’ve seen this signal triggered a total of 22 times since 1999, 19 of which led to a lower S&P close by the time the 2-day RSI closed back under 98…
S&P500 2-Day RSI Closes Over 98
05/04/09… S&P500 ??? when RSI <98
01/02/09… S&P500 -0.5% when RSI <98 (one day)
04/10/07… S&P500 -0.7% when RSI <98 (one day)
10/26/06… S&P500 -0.4% when RSI <98 (two days)
03/17/06… S&P500 -0.2% when RSI <98 (one day)
11/22/05… S&P500 -0.3% when RSI <98 (three days)
09/30/05… S&P500 -0.2% when RSI <98 (one day)
07/15/05… S&P500 -0.6% when RSI <98 (one day)
06/17/05… S&P500 -0.1% when RSI <98 (one day)
11/03/04… S&P500 +1.9% when RSI <98 (three days)
05/25/04… S&P500 +0.7% when RSI <98 (three days)
04/05/04… S&P500 -0.2% when RSI <98 (one day)
09/02/03… S&P500 -0.1% when RSI <98 (three days)
06/04/03… S&P500 +0.2% when RSI <98 (two days)
03/19/03… S&P500 -1.0% when RSI <98 (three days)
05/21/01… S&P500 -0.3% when RSI <98 (one day)
04/19/01… S&P500 -0.9% when RSI <98 (one day)
08/08/00… S&P500 -0.7% when RSI <98 (one day)
07/14/00… S&P500 -1.1% when RSI <98 (two days)
12/31/99… S&P500 -1.0% when RSI <98 (one day)
12/23/99… S&P500 -0.1% when RSI <98 (one day)
07/02/99… S&P500 -0.2% when RSI <98 (one day)
04/12/99… S&P500 -0.7% when RSI <98 (one day)
Prior to 1999 this signal wasn’t nearly as effective as it’s been in recent years.
Finally, it should be noted that Monday represented the fifth consecutive day of positive NYSE breadth. In recent years, such a long string of positive breadth days has been followed by generally toppy conditions looking out 3-6 trading days. In 28 out of the last 30 separate occurrences, the S&P posted a close below the setup day 3-6 trading days later, well ahead of the 66% at-any-time odds…
Breadth Positive Five in a Row
05/04/09? ???
03/16/09? No lower S&P close in 3-6 sessions
01/06/09? Lower S&P close three sessions later
11/28/08? Lower S&P close three sessions later
11/03/08? Lower S&P close three sessions later
07/22/08? Lower S&P close three sessions later
05/16/08? Lower S&P close three sessions later
04/04/08? Lower S&P close three sessions later
08/23/07? Lower S&P close three sessions later
07/03/07? Lower S&P close four sessions later
06/01/07? Lower S&P close three sessions later
05/07/07? Lower S&P close three sessions later
04/04/07? Lower S&P close four sessions later
03/23/07? Lower S&P close three sessions later
02/01/07? Lower S&P close six sessions later
01/12/07? Lower S&P close three sessions later
11/16/06? Lower S&P close six sessions later
08/30/06? Lower S&P close four sessions later
08/18/06? Lower S&P close three sessions later
03/16/06? Lower S&P close three sessions later
01/09/06? Lower S&P close three sessions later
12/14/05? Lower S&P close three sessions later
11/23/05? Lower S&P close three sessions later
06/15/05? Lower S&P close six sessions later
06/02/05? Lower S&P close three sessions later
02/16/05? Lower S&P close three sessions later
12/14/04? Lower S&P close three sessions later
11/15/04? Lower S&P close three sessions later
10/29/04? No lower S&P close in 3-6 sessions
10/04/04? Lower S&P close three sessions later
08/02/04? Lower S&P close three sessions later