Thursday, October 12, 2006

When Selling Becomes Extreme: What Does It Bring?

Readers of this blog know that I rely upon the NYSE TICK as a short-term measure of market sentiment. Most of my studies have focused on the Cumulative TICK and momentum effects in the market, as well as patterns in the distribution of TICK values intraday. In this post, however, we'll examine the importance of single, extreme readings in the TICK.

On Wednesday, in the wake of selling on the release of the Fed minutes, we hit an NYSE TICK reading below -1200. This suggests very broad selling, with more than 1200 issues trading at their bid price vs. their offer. When selling becomes that extreme, what tends to come next?

Since 2004 (N = 695 trading days), we've only had 50 days in which the NYSE TICK has gone below -1100. Two days later, the S&P 500 Index (SPY) is up on average .37% (31 up, 19 down). That is quite a bit stronger than the average two-day gain of .06% (367 up, 328 down) for the entire sample.

Conversely, when the daily NYSE TICK never goes below -500 (N = 45), the next two days in SPY average a loss of -.14% (21 up, 29 down), notably weaker than average.

In other words, extreme selling seems to be bullish; the absence of extreme selling brings subnormal returns. Panic selling, in particular, tends to produce short-term trading opportunities.

Interestingly, on the panic selling days, if there is not a significant *buying* episode that day--an NYSE TICK reading stronger than +1100--the next day in SPY averages a *loss* of -.12% (11 up, 13 down). If there is a strong buying period on a day that features panic selling, the next day in SPY averages a next day gain of .54% (19 up, 7 down).

Extreme selling by itself cannot create a bull move. It's when panic selling draws in buyers--something we did on Wednesday--that short-term returns tend to be favorable. Average NYSE TICK readings are summarized daily on the Trading Psychology Weblog; going forward, I will also flag daily extremes.