Yesterday was an impressive-looking rally for the major U.S. indexes, but more evidence is needed before we can say the tide has turned. Until then, caution is advised, particularly for the next few days, for reasons we will get to in a moment.
But first, the one big positive. The indexes produced some impressive-looking candlesticks yesterday, particularly on the Nasdaq, which gapped up 60 points at the open and managed to end the day 40 points higher than that. The white candlestick – meaning the close was higher than the open – produced as a result was evidence of real buying pressure. The bottom of those candlesticks – 2031 on the Nasdaq, 10,269 on the Dow, and 1187 on the S&P 500 – are now critical support, and could be tested on the first strong pullback.
As we noted in yesterday’s Market Close, market internals, such as up/down volume and new highs/new lows, were solid but not spectacular. Trading volume rose only modestly from Wednesday’s levels. Key indicators like MACD and PPO have yet to turn up. Sentiment went from bearish to bullish overnight, an indicator that a few too many investors think this rally is for real. And all the major indexes except the Dow ended yesterday right at resistance. So it could be just an oversold bounce until those indicators improve.
Today could turn out to be an important day for one big reason. For the last few weeks, the major U.S. stock indexes have been tracing out a familiar pattern – the same pattern the market traced out before the 1929 and 1987 crashes. A hard down day today would continue that pattern, which makes yesterday’s bullish action all the more important, because it could put a short-term floor under the market at just the right time.
The pattern developing in the Dow, S&P and Nasdaq last appeared in the stock market in March, according to George Slezak of FuturesFax.com, who follows the pattern as closely as anyone. It produced a sharp sell-off – but also one heck of a rally off the April 4 low. The pattern also appeared at the bottom of the 1929-1932 bear market, so it can produce some major upside reversals too. It does not always have an unhappy ending. In Slezak’s words, “many get this far and fizzle at any moment.” But this point of the pattern tends to produce strong, reliable signals on market direction, he said.
Through yesterday, the current patterns were a strong match for the 1987 and 1929 tops. The 1987 and 1929 tops had their final rallies on the quarter moon, which was yesterday. And both indexes plunged on the 55th day off the top, an important Fibonacci turn date. 55 days off the May 22 top will be either July 16 or 17 (Monday or Tuesday), depending on who’s doing the counting. And the 1929 and 1987 crashes both occurred within a Puetz crash window, which this market is in until Tuesday. And July 18 is an important cycle turn date (called a Bradley) similar in size and strength to the one that produced the April 4 bottom (July 18 is also the date of IBM’s much-anticipated earnings report). A lot of important turns clustering around the early part of next week could make whatever signal the market produces over the next week a reliable one.
One argument in favor of treating the 1929/1987 pattern as more than a passing curiosity is the high level of complacency in the market right now. The CBOE put-call ratio plunged from .95 at its peak on Wednesday to .34 at its low yesterday morning. That kind of one-day move – from extreme bearish sentiment to extreme bullish sentiment – normally takes at least several days. Yesterday’s dramatic drop in the PC ratio may have just been Microsoft-induced short-covering, but it took weeks after the April 4 bottom before the put-call ratio fell below .60. The PC ratio closed at .51 yesterday. Also, the Investors Intelligence survey of newsletter writers shows them bullish by 2-to-1, a level usually seen at tops. Bulls and bears were even at the April 4 bottom. Investors’ ability to ignore massive charges against earnings from Microsoft and others and a brewing crisis in Argentina that affects 20% of all emerging market bonds should be at least a little worrisome.
All of which means that the real test may still lie ahead.