Friday, September 26, 2014

Have We Hit Bottom Yet? A Process View of Market Behavior

I've received more than a few messages lately wondering if we're "due for a bounce" or if we've reached a bottom.  

It's easy to throw ourselves into tizzies trying to catch precise price highs and lows.  I prefer to think of topping and bottoming as processes, connected by market cycles.  Rather than ask if we've hit the bottom, I'd rather identify when we are undergoing a bottoming process.  First I want to see that we've made a momentum low--a point of maximum negative breadth--and then I want to see how key measures (such as sentiment, correlation, etc.) behave on subsequent price weakness.  Even a seeming V bottom has a process to it, albeit one that is relatively compressed in time.

Let's take an example.  As of yesterday's close, fewer than 10% of SPX stocks were trading above their 3, 5, and 10-day moving averages.  Going back to 2007, there have been only five non-overlapping situations in which this has occurred in a sub-20 VIX market.  Those dates are 2/27/07; 6/6/11; 4/9/12; 6/5/13; and 1/27/14.  This is clearly too small a sample for statistical analysis, but will illustrate an important qualitative point. 

After three trading sessions, SPX was up four times and down only once.  All five occasions posted a higher close within three trading sessions.  However, after five trading sessions, only one occasion was trading higher and four were lower.  Four of the occasions posted a lower close within three trading sessions.  So are we due for a bounce?  Probably.  Have we made a bottom?  Not necessarily.

Indeed, if you look at the forward price paths of the five dates listed above, you'll see a pattern of bouncing higher, moving lower over a period of a number of days.  It is not uncommon for downside momentum to dissipate, even as we make subsequent price lows.  Eventually that downside momentum wanes to the point where price cannot move lower and bulls are empowered to take advantage of the situation.

Shorter price cycles lead to very short bottoming processes; longer cycles lead to more extended bottoming.  The bottoming process for the very large preceding cycle extended from October, 2008 through March of 2009.  The bottoming process for the recent low in early August was just a few days.  Identifying the kind of cycle we're in provides clues for how extended a bottoming process might be.  (The same is true of topping processes).

My chips remain off the table, but with bearish sentiment (equity put-call ratio above 1.10) and stretched downside breadth and rising volatility and correlation, I'm on the lookout for those indications of bottoming.

Further Reading:  Discipline and Devotion