ganns take for mid week

S&P 500 (updated September 5)
Positions:
On the 100% short position in the September S&P 500 entered at approximately 1489.50, lower the protective buy stop to just above 1510.00.
Current Market Analysis

After hitting a near 4-week high as traders returned from the Labor Day holiday on Tuesday, September 4, the cash S&P 500 lost over 17 points, or more than 1% on Wednesday. Stocks were dragged down by weak economic data (the National Association of Realtors index of pending home sales plunged to a 6-year low) and the continuing spike higher in Libor (the London interbank offered rate), as European banks shy away from loaning each other money amid the credit crunch.

The previous rally had brought the stock market close to a moment of truth. After falling 12% in a month, the market abruptly reversed and had retraced approximately 65% of the decline. In the 4 major 7th-year runaway declines (1907, 1917, 1937 and 1987), which comprise 4 of the 10 greatest runaway declines in the stock market since 1886, the respective minor rallies measured only 6.7%, 7%, 3.7%, 5.9%, 8%, 15% and 7%. The 15% rally can probably be thrown out because it occurred near a bottom. Consequently, the limit of what one would expect during a rally in our market would be roughly 8%. So far, we have advanced as much as 9% in the September S&P 500 since the August 16 low.

All of the rallies during 7th-year runaway declines were short-lived, with most taking even less time than our recent 19-day uptrend, meaning a top should be imminent if we are to follow these earlier examples.

Looking at the other 6 of the 10 worst runaway declines (1932, 1929, 1931, 1903, 1920 and 2002), the minor rallies were 7% in 7 days, 6% in 6 days, 4% in 8 days, 14% in 9 days, 8% in 1 month and 19 days, 8 % in 1 month and 7 days, and 21% in 29 days. As with the 7th-year declines, it appears that our market should experience no more than about an 8% advance. In every non-7th year, except 1929, the declines occurred at or close to final bear market lows. Since ours would have to occur very close to a final top, the probability we are in a runaway decline is arguably less favorable than we would like.

Because the stock market rally was pushing up against the maximum limit expected under such a scenario, we lowered our protective buy stop to just above 1510.00 in the September S&P 500 (just 11 points above Tuesday's high), which was the August 8 rally high before stocks really unraveled.

Until the market tips its hand, the big question remains, was the monthlong slide in the stock market between July 16 and August 16 a mere correction or the start of a genuine bear market? And what would it take to confirm the existence of the latter?

Ever since the October 10, 2002 bear market low, we've looked for a final top in the 6th or 7th year of the decade. During the 20th century, all bull markets that launched from 2nd-year lows (1932-37, 1942-46, 1962-66 and 1982-87) peaked in either 6th or 7th years. The structure of the latest top in the Dow reminds us of the final high in November 1916, over 11 years before daily tabulation of the S&P indexes began, and prior to the availability of intraday prices for the Dow Industrials. In that year, the Dow reached all-time highs, exceeding the 100 mark for the first time, only to correct 5%, finding support at the last important all-time high (old highs often serve as support). After setting more new records and climbing as high as 110.15, the Dow came down again and briefly stabilized above the peak (104.15) that preceded its earlier 5% correction. But once the average broke through that old top, the slide turned into a rout.

Similarly, after establishing an all-time closing high of 12,786.64 on February 20, 2007, the Dow slid 5.76%, moving down toward the old January 2000 record. Since closing barely above 14,000 on July 19, the average broke beneath 13,000 and closed as low as 12,845.78, less than 0.5% above its February high. A Dow close significantly below 12,786.64 could provide the next key sign confirming a final top, and would likely lead to a rapid dive under the 12050.41 March closing low and retest the January 14, 2000 climactic bull market high of 11722.98 and the May 2006 top at 11642.65. Under that scenario, both the Dow and S&P 500 would easily overbalance (exceed) their biggest corrections of the 2002-07 bull market in absolute points and maybe even on a percentage basis.

As it stands, the 185.30-point drop in the cash S&P through August 16, though the index’s worst in years, is still not that much greater than a 165.38-point correction to a secondary low between December 2, 2002 and March 12, 2003, shortly after the start of the bull market. We’d like to see the magnitude of this decline exceeded by at least 20% to confirm a bear. In terms of time, an initial leg down in a bear market will normally eclipse the duration of the most recent corrections, but this downtrend has yet to even match the 1-month-and-6-day setback from May 8 to June 14, 2006.

Between 1989 and 1997, the Dow Industrials underwent a number of corrections in which the average sold off to over a 10% cumulative loss intraday, but, as has happened here, avoided crossing that dubious threshold on a close. However, in 1998, a mid-August rally following such a scenario faltered, resulting in a bloodbath at month’s end.

If we’ve witnessed no more than a garden-variety correction, we’ve probably seen the intraday low. Even so, additional turbulence could be in store. If a genuine bear market is under way, the consequences could be severe. Of bear markets from all-time highs in the Dow that unfolded on the heels of a prior 36-46% bear market loss, as seen in 2000-02, all but one eventually led to a minimum 45% decline and undercut the last bear market low.

During the uptrend through mid-July, the extreme length of the leg up starting from a June 14, 2006 low (we consider the subsequent advance a single leg, notwithstanding the sharp sell-off last February-March) represented a cause for concern in in the S&P 500, where 91% of all previous bull market legs higher lasted less than a year. The marginal new record intraday high in the S&P on July 16 extended its cyclical advance to 4 years, 9 months and 6 days, surpassing 1932-37 to become the 4th-longest bull market in the history of the index, which dates back to 1928.

Markets as a rule tend to marginally exceed important prior tops, as the S&P just did, before a genuine major reversal. However, no bear market has ever commenced from a marginal all-time high in the S&P 500. In February 1946, the S&P turned sharply lower as soon as it clipped its 1937 bull market high (which was well short of 1929's record), but the downturn resulted in only a 10% correction that was over before the end of the month. Also, the timespan between record highs before the latest ranks as the 3rd longest (behind only 1929-54 and 1973-80) in the history of the S&P. A market that finally overcomes a prolonged drought between all-time highs can sometimes enter a price vacuum upon clearing all prior overhead resistance. However, in this case, the bull market was more mature and decennial patterns more unfavorable than usually necessary to support a credible breakout.

The stock market is moving toward a seasonally dangerous timeframe, particularly in the 7th year of a decade, when many of the worst crashes materialize. But most bull markets making new highs in the second half of a 7th year manage to escape with no more than a steep correction. The 1987 crash (20-Year cycle), following a climactic late-August high on the heels of a 5-year bull market that. like ours, sprang from a 2nd-year low, stands as a notable exception. An August high in 1937 also set up that year's devastating crash, but it was merely a secondary top. The actual high was established in early March of 1937. Our market, though roughly as aged as 1932-37 and 1982-87, may not be eligible for as spectacular a decline because of its comparatively lesser magnitude (up "only" as much as 102% intraday from the October 2002 low in the S&P). Of markets hitting new highs late in the 7th year of a decade, both 1927 and 1997 sustained double-digit percentage drops limited to the month of October, while a similar-sized correction beginning in late September of 1967 stretched into March of the following year.

Also, declines of the magnitude we've already seen so far beginning from bull market highs in mid- to late July tend to lead to further selling and/or foreshadow weakness later in the year, especially into early fall. Past examples include 1919, 1933, 1943, 1975, 1990, 1998 and 1999.

The VIX (also known as the "fear gauge"), an indicator of expected future stock price volatility which late last year recorded its first daily closes below 10 since January 1994, recently spiked to its highest levels in well over 4 years.

We are unquestionably long overdue for at least a significant correction after a record interval of over 4 years and 4 months without a decline of at least 10% on a closing basis (rounded to the nearest percent) in the history of the major averages (Dow Jones and S&P).

Thanks to the recent sharp dive in stocks, credit markets are pricing in a cut in short-term rates by the Fed at its September meeting, if not sooner. But the interest-rate backdrop remains particularly worrisome even if a final stock market top is not at hand. Dating back to the American Civil War in 1861, important downturns in bond prices (rising long-term interest rates) have supplied valuable advance warning ahead of 23 of the 28 concluding bull market peaks in stocks, or 82%. Lead time ranged from 2 to 22 months. The 30-year T-Bond registered a significant high last December 1, then bounced back to successively lower tops in late February and May before falling off the table prior to its current strong rally. Also, the rate-sensitive Dow Jones Utilities Average topped out 8 weeks before the blue chips on May 21.

Despite the S&P's new closing record in July, the NASDAQ indices came nowhere near all-time highs. The NASDAQ Composite made it to little over half its record price from the year 2000, and the NASDAQ 100, which hit a 6-year high on July 19, came no closer than within 57% of its historic peak. Of course, this merely extended a pattern of massive divergences between the Dow and broader stock indices held back by a heavy technology weighting.
We believe that the severe problems recently plaguing subprime mortgage lenders, whether or not they continue to spill over into the broader economy, are symptomatic of a greatly increased (until recently) appetite for risk in the financial markets that could cause a major stock market decline to snowball. Record amounts of margin debt and mergers/corporate buyouts, often employing an unheard-of degree of leverage, combined with last year's significant fall in housing and auto sales, recall conditions that prevailed at the top in 1929. So you can see why we're so concerned about the next bear market in stocks.

The recent spate of leveraged buyouts financed by private equity didn't necessarily mean the acquisition targets were cheap, just that financing was easily obtained. Historically, takeovers tend to peak when stocks are overpriced, as in the 1920s. Late July's trading should dispell any notion that buyout activity can prop up the market. If anything, it's the other way around. With stocks tanking, financing for a multitude of possible deals suddenly came into question. In a similar vein, today's strong reported earnings don't guarantee future support for the stock market. Corporate profits relative to GDP peaked in 1929, on the eve of the Depression, and in 1966, right before the Dow fell 72%, adjusted for inflation, in 16 years.

Our gain of as much as 100% in the closing S&P since the 2002 low has somewhat exceeded the rise in 1962-66, but significantly trails the other 3 much more robust 2nd-year bull markets (1932-37, 1942-46 and 1982-87).
At 1 year, 1 month and 2 days (through July 16), if we have experienced a final leg, it would represent the 2nd longest final leg ever (of 17) in the S&P. But the 4 lengthiest prior advances into final tops resulted in moves of 74%, 46%, 24% and 51%, respectively. In other words, when final legs up persist beyond 5 months, the tendency is for the moves to be far more robust than we saw in our market. One could infer from this that our market was not demonstrating the type of aggressiveness you would ordinarily expect, which could somewhat diminish the prospect that a final top is in.

Nonetheless, the bull market -- and its most recent leg -- were already very mature by normal standards.

As for 7th years of decades in the last century, bear markets in 1907 and 1917 made final lows in mid-November or later. In both 1937 and 1987, the stock market crashed spectacularly in October. The 1956-57 bear market sustained virtually its entire loss between mid-July and late October of the 7th year. Both 1927 and 1997 featured harrowing double-digit percentage corrections in October. And in 1967 the bull market, though only in its 2nd year, began a 6-month 12.5% Dow correction in the last week of September.

Technicals

As of Tuesday (September 4), the September S&P 500 had advanced as much 9% off the August 16 low. If history is our guide, then this is about as far as the market should go. Our analysis of the 10 greatest runaway declines in the history of the stock market, including four major 7th-year runaways, indicates that we should experience not much more than about an 8% rally. Coincidentally, both the S&P 500 and the Dow Industrials rallied near Fibonacci .618 retracement levels on Tuesday and then backed off sharply on Wednesday. If the September S&P 500 can now break below 1435.00 (the low of each of the last two weeks), it could cause a swift decline to the August 16 low of 1375.00 or the weekly March nearest-futures low of 1364.00 (all-sessions).