NYSE New Highs

Behind the advance/decline numbers, the most commonly-referenced breadth figures that analysts use are the number of new highs and new lows, usually over the past 52 weeks.

According to the textbooks, the greater the number of new highs, the more healthy the market, especially if that number is expanding over time.

The textbooks, then, should be cheering yesterday's performance.  The S&P 500 broke to a new high and was pushed along by over 18% of the issues trading on the New York Stock Exchange.  That's the greatest percentage since April.

But April wasn't exactly the best time to be sucked into buying stocks, and it turns out that's not uncommon when we see such a jump in the percentage of new highs.

The table below shows every time since 1962 that the number of issues trading at a 52-week high exceeded 18% of total issues.

A few times, such as 1982, 1986, and 2003, such an event was cause for celebration as the textbooks nailed the analysis.  Stocks chugged higher pretty much without interruption.

But more common was a gimpy market that struggled to maintain its momentum.  Outside of the three time periods mentioned above, the S&P sported a negative return 7 of remaining 10 times two weeks later.

Six months later, the S&P was up by an impressive amount after the three exceptions noted above.  But the other instances showed a median return of -0.9% six months later, with only one greater than +5%.

Overall, the returns were in line to slightly below random across all time frames except one month later when it was modestly above average.  There were still four losses greater than -2%, and two gains less than +1%, but the other 8 were solidly positive.

While not shown on the table, 12 of the 14 instances sported a positive return one year later (assuming positive returns for the last two instances in March and April of this year).  Overall, though, the median return was only +5.4%.  Five of the dates gave returns greater than +10% and three greater than +20%.

I would not consider this to be a sell signal, although it would have worked better that way in April (and January 11th when we saw a spike to 17.4% of total issues that just missed the cutoff for the table above).  But I sure as heck am not cheering this development as a major buy signal, at least for the short- to intermediate-term. 

S&P 500 Price Performance

The markets were firing on all cylinders yesterday (unless you look at anything dollar-denominated of course).  The pattern couldn't be more pretty for bulls - a large gap up, a new multi-month high, and enough buying pressure to close the S&P above its opening price.

With the S&P looking so good, let's see how it performed after similar setups.  The table below shows the S&P's performance over the next two weeks when the S&P 500 SPDR (SPY) gapped up on the open and never filled the gap (i.e. it never traded below the previous day's close), it closed higher than the open and at at least a six-month high.

Out of 16 occurrences since 1998, it managed to climb higher only twice.  And one of the positive occurrences (12/01/03) showed a maximum loss that was nearly three times greater than the maximum gain as the market fell back then jumped right before the two weeks was up.

Prior to 1998, it wasn't nearly as bearish.  There were another 12 occurrences, and it was higher two weeks later 9 of those times.  The median return was modest, +0.8%, but at least the average maximum gain (+1.9%) was well above the maximum loss (-0.6%).