Often in these comments, I've made no bones about the fact that I am not a fan of divergence analysis.  I think it's an over-hyped, over-blown, inconsistent strategy that has one looking for turning points all the time, mostly in frustration.

The one on most folks' radars at the moment is the massive positive divergence we're seeing between the stock indices and the number of new 52-week lows on the NYSE.  As we discussed last October, there was an absolutely incredible number of stocks hitting a new yearly low on the same day then, a truly historic event.

While the number of new lows has been increasing over the past week as the indices slump toward their 2008 lows, they are nowhere near the peak we saw last fall.  We could see more than 600 stocks on the NYSE hit a new low - more than twice what we saw yesterday - and it would still be 75% less than the peak reading from October.

That has technicians all atwitter because it would give us one of the most classic positive divergences in existence.  We'd have the indices hitting new lows, but a massively less number of underlying stocks doing so...so that's gotta be a good thing, right?

Not necessarily.

The table below shows all instances since 1962 when we saw the S&P 500 hit a new yearly low, on the same day that new lows on the NYSE hit a one-year high.  In other words, a capitulatory event like we saw in October.

Then, the market rallied at some point and came back down to hit another 52-week low within six months, but the number of new lows on the NYSE was less than 25% of the peak reading of the past year - a massive positive divergence like we'd probably see now if the S&P drops to 740 in the next couple of days.

For the most part, performance in the S&P 500 going forward was positive across all time frames.  But it was only weakly so, and from one month to six months later, its average return was less than a random comparable return during the study period.

Longer-term, all but one of the instances showed a positive return, but the failure in 1973 was a hefty -17%.  And while we don't have enough data from the instance in July 2008, that one will almost certainly show a large negative return as well.

 Disclaimer  |  Privacy Policy