Saturday, September 12, 2009

More on Short-Term Stock Market Momentum

I recently posted an initial look at momentum in the stock market, drawing upon weekly data. This further look makes use of daily data.

Going back to 2000 in the S&P 500 Index (SPY), when the number of advancing stocks on the NYSE minus the number of declining stocks amounts to 20% or more of the issues traded, then we have taken out the current day's high price on the next trading day 69% of the time. Conversely, when we've seen such advance/decline strength on the day, we've taken out the current day's low on the next day only 26% of the time.

When the number of number of declining stocks minus the number of advancers amounts to 20% or more of the issues traded, then we have taken out the current day's high price on the next trading day only 29% of the time. But when we've seen such advance/decline weakness on the day, we've taken out the current day's low on the next day 66% of the time.

Also going back to 2000, if the current day closes above its pivot price level, the next day in SPY has taken out its prior day's high 68% of the time. It has taken out its prior day's low only 32% of the time.

If the current day closes below its pivot price, the next day in SPY has taken out its prior day's high only 31% of the time, but has taken out the previous day's low 67% of the time.

What we see is that strength and weakness tends to carry over to the next day's trade, though not necessarily to the next day's close. (See here for more on that dynamic).
.

5 comments:

JimRI said...

This pertains to the tweet you have on insider trading - selling.

Am I correct in seeing similar sentiment and market activity that we had before the 2000 collapse? I noticed then that the insiders of the tech companies whose stock was going to infinity, were selling like mad but the common word was that we were in a new era and there was no up limit.

Today, we see a market holding prices high when one would expect some fall back and people beginning to say that we won't have an fall back that it will continue to go up in a way defying historical behavior. We also see insiders selling.

Mike said...

In recent weeks, you have described several studies similar to this one. IMO, they are all misleading and useless. You have come up with several measures that are highly correlated with where the market closes in its intraday range. If the market closes toward the high end of its range, it is no surprise that the next day it is more likely to trade above the high than below the low (it would do this even if the next day's returns were random). Instead, you must look at the average return on the next day, or at a minimum calculate percentage of up days vs percentage of down days. This is bad stuff on an otherwise truly excellent blog.

Joslin Lolo said...

hi Dr.Brett, any variation in how I interpret your posts can leave me scratching my head for a few days...can you explain how you calulate "the number of declining stocks minus the number of advancers amounts to 20% or more of the issues traded"....thanks!

Brett Steenbarger, Ph.D. said...

Hi Mike,

I agree with you that, just by itself, knowing this momentum pattern doesn't constitute a valid trading signal or system. Where it is helpful, I find, is in situations where a market will move off its high or low overnight, but not reverse its prior trend. That leads to good trades where we can profit from a move that tests the prior day's high or low.

Brett

Brett Steenbarger, Ph.D. said...

Hi Joslin,

If you take the difference between the number of stocks declining and the number of those advancing and divide by the number of issues traded (and multiply x 100), you'll get the percentage that I refer to.

Brett