Heavy Volume at a Two-Month Low Not a Short-term Positive

Declining issues outnumbered advancers by a wide margin throughout Thursday’s session, a clue that the early rebound effort was on shaky ground. When the S&P poked its head into positive territory around 12:30 ET, breadth remained 2:1 negative on the big board. Nasdaq breadth also remained decidedly bearish even as the NDX briefly sported a 1.5% gain. By the close, those negatives proved too much to overcome, and stock indices gave up the morning rebound and then some. TICKscore settled at -117, its second-lowest close of the year (behind Wednesday’s session), following a monstrous wave of institutional selling in the last two hours. Cumulative TICK settled at a similarly extreme -133,750. Over 1,800 issues hit new 52-week lows on the NYSE, only the second time we’ve ever seen over 1,500 new lows (the first time being in early October).

Notice that we’ve finally begun to see some readings below 1.50 on the NASDAQ/NYSE Volume Ratio, although they’re still not low enough. The 20-day moving average has a long way to go before it even pulls back to average territory around the 1.25 level (see longer-term chart). Prices are likely to be considerably lower when that time comes.

NYSE volume settled over its upper bollinger band Thursday, a potentially dangerous sign given that it coincided with a new 52-week low for the S&P500 and a new 52-week high for the VXO. Big board volume has settled over its upper bollinger band prior to some of the largest one-day declines over the past forty years, including the ’87 crash, the 1998 ‘mini-crash’ and the 9/11 terrorist attacks. The table below lists the last thirty instances in which NYSE volume closed more than two standard deviations above its 20-day average and the S&P settled at its lowest level of the past two months. Note that in 27 out of 30 cases, or 90% of the time, the S&P posted a subsequently lower close at some point over the next five trading days. That’s significantly better than the 72% at-any-time odds of a lower S&P close within five days, indicating that when the S&P closes at its lowest level in at least two months and there’s a noticeable surge in volume, it rarely marks the ultimate bottom…

S&P Closes at Two-Month Low, NYSE Volume Over Upper Band
11/20/08… ???
10/10/08… No lower close within five sessions
06/27/08… Lower S&P close three sessions later
06/20/08… Lower S&P close two sessions later
11/08/07… Lower S&P close one session later
07/26/07… Lower S&P close one session later
02/27/07… Lower S&P close three sessions later
05/17/06… Lower S&P close one session later
04/15/05… Lower S&P close three sessions later
07/21/04… Lower S&P close two sessions later
03/11/04… Lower S&P close two sessions later
07/19/02… Lower S&P close one session later
06/26/02… Lower S&P close three sessions later
06/07/02… Lower S&P close two sessions later
06/06/02… Lower S&P close one session later
01/29/02… Lower S&P Close four sessions later
09/19/01… Lower S&P close one session later
09/17/01… Lower S&P close one session later
09/07/01… Lower S&P close two sessions later
09/06/01… Lower S&P close one session later
03/22/01… No lower close within five sessions
03/16/01… Lower S&P close two sessions later
02/22/01… Lower S&P close one session later
11/30/00… No lower close within five sessions
10/11/00… Lower S&P close one session later
08/27/98… Lower S&P close one session later
08/04/98… Lower S&P close five sessions later
07/16/96… Lower S&P close five sessions later
11/23/94… Lower S&P close five sessions later
10/05/92… Lower S&P close one session later
06/17/92… Lower S&P close one session later

SPY volume also closed over its upper bollinger band Thursday. If it closes over its upper band again on Friday, it would have negative short-term implications. For this to occur, SPY volume would need to exceed today’s volume of 813 million shares. That’s not particularly likely based solely on the fact that it’s options expiration – usually it’s the Thursday before expiration (today) that sees a spike in volume – but it would be noteworthy if it does occur because it would suggest an unusually large amount of institutional participation. This has typically translated into choppy conditions over the short-term, with a close below the setup day a common occurrence. Since 1996, the first year when average SPY volume topped a million shares, there have been 26 separate instances in which daily SPY volume topped its upper bollinger band two consecutive sessions, all of which are listed in the table below…

SPY Volume Over Upper Bollinger Band Two Days in a Row
09/16/08… Lower SPY close one session later
09/09/08… Lower SPY close four sessions later
07/11/08… Lower SPY close one session later
03/14/08… Lower SPY close one session later
07/27/07… Lower SPY close five sessions later
07/19/06… Lower SPY close one session later
01/19/06… Lower SPY close one session later
10/06/05… Lower SPY close two sessions later
07/01/04… Lower SPY close one session later
03/11/04… Lower SPY close six sessions later
10/01/03… No lower SPY close within six sessions
01/24/03… Lower SPY close one session later
07/11/02… Lower SPY close one session later
01/30/02… Lower SPY close three sessions later
02/23/01… Lower SPY close three sessions later
09/25/00… Lower SPY close one session later
04/14/99… Lower SPY close one session later
08/28/98… Lower SPY close one session later
07/23/98… Lower SPY close one session later
04/27/98… Lower SPY close one session later
03/06/98… Lower SPY close one session later
01/12/98… No lower SPY close within six sessions
10/28/97… Lower SPY close one session later
06/24/97… Lower SPY close one session later
09/03/96… Lower SPY close two sessions later
07/16/96… Lower SPY close five sessions later

Thursday marked the second consecutive 80%+ (actually 90%+) down volume session on the NYSE, triggering the 1-2 day buy setup recently discussed in my November 17th column. However, that setup is neutralized by the short-term negative implications of Thursday’s high-volume decline discussed above. They’re both up on the board, making the near-term outlook somewhat mixed, but in this environment I would tend to favor those setups that are going with the trend rather than contra-trend. I’d also add that running the same study from my 11/17 commentary utilizing a pair of 90% (rather than 80%) down volume days yields far fewer signals (six since 1990) but also a much lower winning percentage (50%), suggesting a cluster of extreme down volume sessions is not as effective at signaling a short-term bottom.

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

 Print This Post