Short-term Outlook (1-5 Days):  10% Long (from 9/23 at 112.65)

Go to:  Top  |  Short-term Outlook  |  Int-term Outlook  |  Equity Updates  |  Indicator Summary  |  Commodity Updates

Go to:  Top  |  Short-term Outlook  |  Int-term Outlook  |  Equity Updates  |  Indicator Summary  |  Commodity Updates

Equity Indicators - Updates and Extremes

Hindenburg Omen

Last month, we took a look at the Hindenburg Omen, an overly data-mined "market crash" signal.  It got a huge amount of attention at the time, and was looking prophetic for a couple of weeks, but has since petered out faster than King Jung il's assertion he's not a nepotist.

Arguments vary about when (or if) a Hindenburg Omen signal is rendered moot, but I'd say the market rallying is a pretty good sign that it's not working.  Actually, looking back at the Report from August, the S&P did decline about in line (or more) than historical averages, at least during the first few weeks.  It's the longer-term that's not looking as disastrous.

So let's go back to 1965 and look at every Hindenburg Omen, along with how the S&P 500 performed during the subsequent 30 days.  We'll separate out future performance based on if the market rallied well or fell hard during that first 30 days, to see if it had an impact longer-term.

By "rallied well", let's require that the S&P rise by at least +5% during that first 30 days, and also that the maximum gain is twice the maximum decline during that time.  For "fell hard", we'll require the opposite - a drop of at least -5%, with a maximum decline at least twice as large as the maximum gain.

From the table, we can see that there were 20 times the S&P fell hard during the first 30 days, versus only 9 times it rallied well.  That's right in line with historical expectations following the Omen.

What's interesting is how the S&P performed over the ensuing six months.  When the market bucked the Omen for the most part and did well during the first 30 days, then over the next six months it was positive every time, and the maximum gain averaged twice as large as the maximum loss during those six-month stretches.

When the market fell hard during the initial 30-day period, then it also showed a positive six-month return (which isn't too surprising given the market's long-term trend and tendency to rebound from shorter-term declines).  But the risk/reward wasn't nearly as positive as it was compared to those times when the market rallied after the Omen.

NYSE Unchanged Issues

Barry Ritholz pointed out a chart yesterday from Ron Griess showing the percentage of unchanged issues on the NYSE.  As the exchange went to decimal pricing, the number of unchanged issues dropped dramatically (it takes less to move a stock $0.01 than it did to move it 1/16 of a point).  The conclusion was that this is likely the cause of the numerous 90% up and down days we've been seeing lately.

Well, yes and no.

This is a concept we've discussed several times in the past couple of years, and it's important to remember.  We simply cannot rely on breadth extremes like we were able to a few short years ago.  But I don't think decimalization is the main culprit.  Here's why:

The chart shows the percentage of unchanged issues (the blue line) along with the number of extreme breadth days, this being defined as any day where more than 80% of stocks closed up or down on the day (the small gray lines at the top of the chart).

The correlation between the two is -0.14 (on a scale of -1 to +1), meaning that the two are very modestly inversely correlated.  So as the number of unchanged issues dropped, the number of extreme breadth days increased.  That's what we expect.

But there is a 0.22 correlation between those extreme breadth days and the VIX implied volatility index.  Which makes sense, too - as market volatility increases, so does the number of extreme breadth days.  In fact, this explains more of the move in breadth than decimalization does.

Even still, the cluster of extreme days since 2008 stands out, and neither decimalization nor volatility can explain it.  I know it's cliché by this point, but I do blame high-frequency trading for these distortions, and it's the main reason I rely on breadth as an indicator less than I used to...especially the vaunted 90% days.

Go to:  Top  |  Short-term Outlook  |  Int-term Outlook  |  Equity Updates  |  Indicator Summary  |  Commodity Updates