SPY Update 12/3/07 (SPY @ 133.15): Intermediate Term Bottom May Have Formed

Click on image to enlarge.
12/3/07: The SPY has been in a clear downtrend since my last post ("The Ugliness Continues") but that maybe about to change. Stay tuned.
The market violated first its trendline on 10/17, then 3 days later its 4 week low. It attempted to rally, ran into resistance exactly at its July peak of 140 (funny how these support and resistance levels seem to really work out), then sold off straight down to 127, a point above the 126 spike low of the August sell-off.

Now the clear 4 week trendline has been broken with a strong, wide-ranging day and 2 days of trading after in a narrow range.
It look as though the bulls are setting up to run the market higher. Seasonality is on their side - December is a strong, safe month for the S&P 500 and we are entering the bullish stretch that usually lasts from mid-late November through January. Interest rates are lower on a year-over-year basis, and valuation is not outrageous (but questions about the E part of the PE remain given the subprime mess).
Fundamentally, things remain pretty awful, with the dollar having lost 40% of its value peak-to-trough, an economy that is 70% dependent on the consumer to leverage himself so he can buy some more thing she can't really afford, combined with profligate government deficit spending...
But then again, they say to buy when blood is on the streets.
They also say not to try to catch a falling knife or it'll be your blood in the streets.
I anticipate the market will retest its recent lows, maybe not get quite down the 127-128 range, form a double bottom, then surge back to the 140 level or beyond.
We shall see...

S&P 500 Best Days of Month

The stock market tends to make most of its gains during the following trading days of the month:
- last 3 days
- first 2 days
- days 9, 10, and 11 (probably because of 401k inflows).

- Yale Hirsch in the Stock Traders Almanac calls these days - the last 3 plus first 2 plus days 9-11 - the "Super 8 Days." They have gained 2.2% v. -.43% for all other days of the month for the last six years (2000-2006).
Note that these are trading days, not calendar days, sent to follow or incorporate the strategy, you must count the number of trading days from the beginning and until the end of the month.
- Stock Trader's Almanac, 2007

First Day of Month Very Bullish for Dow Jones Industrial Average

Over the past 10 years, the market has made all of its gains during the first trading day of each month. In fact, the average of all other days of the month was negative.

From September, 1997 - 6/1/06 1st d of month, there were 106 such first days.
- The Dow gained gained 41.7 points on average during the first day of each month vs. a 0.37 point loss the other days.
The first day of August performed worst falling 6 of the last eight times.

- Stock Trader's Almanac, 2007

First Months of the Quarter Tend to be Bullish

According to the Stock Trader's Almanac, 2007, the first month of each quarter tends to be bullish:
Quarter: 1st month average percentage change, S&P 500, 1950-June, 2006:
I (Jan) 1.4%
II (Apr) 1.3%
III (Jly) 0.9%
IV (Oct) 0.9%

Average: 1.2% v. 0.06% average for month #2 and 0 .17% month #3.
- Stock Trader's Almanac, 2007

SPY update: the ugliness resumes...

10/19/07 4:37 pm (SPY 149.6): Today's 300+ point plunge was no doubt heavily distorted by options expiration but may reflect some fundamental reality. The dollar's decline, oil's rise, and the great uncertainty over the extent of the credit damage were all cited as reasons for the decline, but of course those factors were present during the recent almost 20% trough-to-peak rise in the SPY.
Technically, one does not have to know the fundamental why of a market's movement to make money or protect profits. This market seems in a sloppy 20% range characterized by broken uptrend lines, false breakouts, then surges back to new highs. This is the third such break by my count, the first being in February, the last during the summer, and now.
Things to note: an uptrend line was broken at 155 last week. The market traded below this line all this week then broke decisively below its 20 day trailing low.
It will be interesting to see how the market behaves on Monday once the options overhang is gone and people have had time to digest CAT's pronouncements on the economy, etc. If the July-August script is followed, we are only halfway down if that, but the dollar has crumbled since then and the Turkey-Kurdish conflict has surged and represents a potential meltdown in the only part of Iraq that was starting to resemble something close to success.
Extremely bullish seasonality is just around the corner, during the November-January stretch in particular, so if there is a sell-off, the opportunity to make sharp gains in the next few months will increase.

On Innumeracy: A Minor Quibble about Percentage Calculations in an NPR Story

David Kestenbaum in "Japan Wrestles with Kyoto Accord Promises" stated that "instead of cutting emissions by 6 percent, they have grown about 8 percent. That adds up to a 14 percent problem."
Actually, this is incorrect and reflects a commonly-held misunderstanding of percentages. Percentage declines cannot be directly added to percentage gains to get a percentage difference. To illustrate this fallacy, lets say the base emission index had a value
of 100. A 6% decline would have brought the index to 94, Japans target. An 8% gain, however, would bring the index to 108. The percentage difference between 108 and 94 is 14.9%, not 14% as reported in your story.
When the percentage differences are small, as in this case, the error of simply adding the percentages is also small, but larger numbers illustrate the fallacy. For example, if Japan had wished to halve its emissions (a 50% reduction), the difference between this ideal and the actual 8% gain would be a whopping 116%, far greater than the 58% difference that Kestlebaums method would have computed.
Just a minor point in an otherwise solid story.

S&P 500 Breakout - Buy at 152.44

The S&P 500 - I will use its depository receipts or SPY from now on - broke out decisively yesterday to the upside. Around midday it was forming a high, tight flag, indicative of a potential explosion to the upside. A stop could have been placed just above 140 or so. It is unclear if it would have been filled near that range however as the market exploded up immediately after the fed announced its interest rate cuts.
Only a few weeks ago I was bearish and net short the market. Then I was flat; I spent the last 45" of trading (an excellent time to enter the market, by the way - the first 30" are the worst) getting long.
I bought SPY at prices from 151 to 152, as well as XLE, XLU, and the Dogs of the Dow, namely AT&T, Verizon, GM, GE, Citigroup (a phenomenal value right now, me thinks), JP Morgan. A full list is available at www.dogsofthedow.com. I also picked up some Target.
Any 3% surge on good volume across major trendlines and the 20 day high must be taken seriously. Historically, most followed through. The dangers remain, of course, but the willingness of the Fed to perpetuate the bubble if need be means investors should adhere to the old adage of not fighting the Fed. They seem serious about doing whatever it takes to restore liquidity and confidence.
I would expect some choppiness but an equal likelihood is a market that never comes down to "reasonable" levels, so starting to move in aggressively any cash on the side is critical to long-term performance.

SPY Sector Fund 36-Month Return Ranking System, 12/03-August, 2007

If you were to sort the SPY sector funds by trailing 36-month total return at the end of each year, then divide your money evenly among the top 3, you would have achieved excellent results over the past several years. Caveat: these ETFs were launched less than a decade ago, so the data is limited, but studies done with sector mutual funds going back to the 1980s have produced similar results. Also, I have a quibble with how Yahoo calculates its "adjusted prices" which are used here for total return calculation. These returns overstate slightly the returns per year when adjusted for dividends paid. For example, if a stock began the year at $100, paid a $3 dividend, then ended the year at $110, your total return would be 13% ($10 capital gain plus $3 dividend). Yahoo calculates it somewhat differently, in effect subtracting $3 from the initial price (making it $97 ($100 minus the $3 dividend paid)). This creates a return of 13.4% rather than $13% ($110/$97 - 1 = 13.4%). This may seem like a trivial difference, but compounded over a long data series, these differences add up. Nevertheless, for simplicity's sake, I used the Yahoo adjusted price for total return calculations, understanding that it overstates the returns of high dividend-yielding stocks.

I only had month-end data to August, 2007, so did not of course have the full year results available as of writing this.

Nevertheless, the results are impressive:

Model: Invest in the top 3 SPY sector funds by trailing 36 month return, December, 2003 - August, 2007*:
		Start Date:	Dec-03	Aug-07	<>
	Model:	SPY:	Diff:	3.67	years
10k:	18,180	14,048	4,132
rr:	17.7%	9.7%	8.0%
worst:	5.3%	4.8%	0.5%
% of years > SPY:	100.0%
10th %ile:	9.0%	4.9%	4.1%
rr + 10th %ile:	26.7%	14.6%	12.1%
$10,000 invested at year-end 2003 grows to $18,180 (17.7% per year) versus $14,048 in the SPY (9.7% per year), a difference of 8.0% percent per year. Every year of the study period led to an outperformance of the model versus the SPY. Here is a year-by-year breakdown with each of the winning funds: Date IYR XLB XLE XLF XLI XLK XLP XLU XLV XLY SPY avg rr: 10k-> SPY: Dec-99 Dec-00 Dec-01 Dec-02 10,000 10,000 Dec-03 30.2% 13.5% 12.9% 10.7% 18.8% 11,883 11,070 Dec-04 9.0% 4.0% 40.2% 4.8% 17.7% 13,992 11,605 Dec-05 33.2% 17.3% 19.8% 15.2% 23.4% 17,268 13,370 Dec-06 -10.4% 19.8% 6.4% 5.1% 5.3% 18,180 14,048 The number shown for each fund is the next 12 months return. If no number is shown, then the fund was not one of the top 3 by trailing 3 year return. Notice that real estate (IYR) and energy (XLE) dominated, appearing 4/4 and 3/4 the years in question. The model did even better, although it is a high risk strategy I would not recommend, if only the single top sector ETF were bought and held for the next 12 months: Model: Invest in the top 1 SPY sector funds by trailing 36 month return, December, 2003 - August, 2007*:
		Start Date:	Dec-03	Aug-07	<>
	Model:	SPY:	Diff:	3.67	years
10k:	19,941	14,048	5,893
rr:	20.7%	9.7%	11.0%
worst:	9.0%	4.8%	4.2%
% of years > SPY:	100.0%
10th %ile:	11.5%	4.9%	6.6%
rr + 10th %ile:	32.2%	14.6%	17.6%

S&P 500 Sector ETF List

For all it's worth, here is a list of the main S&P 500 Sectors (and an iShares real estate ETF as well):

SPDR BasicInd XLB
ConSvc XLV
ConStpl XLP
CycTrn XLY
Ene XLE
Financial XLF
Indus XLI
Tech XLK
Utility XLU
Real Estate IYR

S&P 500 update 9/13/07 spy

The S&P 500 appears to be setting up for a breakout. Although it has not yet breached its 20-day high, the market has been forming higher lows and challenging its recent highs, holding in the upper end of its recent trading range, after breaking a downtrend line from the false breakout high. A trend line, properly drawn, should only connect the highs preceding lower lows.
Volume on up days still appears weak, but support appears to have moved up to the 144 area.
It the market decisively breaks above 150, then another leg, perhaps to the 156 high, could follow. This would surprise everyone which is what the market tends to do most of the time.
Individual sectors that look interesting include energy (XLE) which has broken out again and is perhaps short-term over-extended, technology (XLK), and basic materials (XLB). Financials (XLF) remain in a downtrend, but the 31.5 climax low may hold for awhile at least. A close above the down-sloping 50 day, currently at 34.29 could be good for a nice trading swing.
The macro economic picture remains horrible and we are in a seasonally weak spot, but pre-election Novembers are usually strong. The market may breathe a sigh of relief if the President indicates he is finally changing course in Iraq, which would eliminate a major financial and credibility hemorrhage to the United States. We currently spend about 12 billion dollars a month there, so any sign that this might be the beginning of the end could be greeted positively, particularly by the credit markets.

S&P 500 Update 9/4/07 12:55 pm (SPY 148.78)

S&P 500 update midday: Well, things are starting to get interesting. This is why it is so important as a trader or investor to be agnostic (literally: not knowing) rather than wedded to one position all the time (long or short). The market has now broken its downtrend line after looking like it was going to make another leg down last week. This could represent a change in trend, but for confirmation, I would like to see the 20 day trailing high taken out (currently at 150.59 on the SPY).
Negatives remain. No pronouncements by the Feds or by Bush about bailing out some homeowners at the margin changes the macroeconomic uncertainty or the overall credit crunch. Nevertheless, the idea of not fighting the Fed remains in place.
Volume is low on up days, the market's action has been distorted by light pre- and post-Labor Day trading (when dramatic moves, usually to the upside, are not uncommon), and if you are bullish, this would be an over-extended entry point. A pullback to at least 144-145 if not the 138-139 range would be a normal and orderly development.
More later. Watch closely. September is a horrible month, but maybe we got some of the carnage out of the way in August... I'm still flat to short the S&P 500 (via puts).

S&P 500 Presidential Election Year Cycle

Seasonality

Presidential Election Year Cycle:

A fascinating relationship exists between the stock market and the 4 year Presidential election year cycle. If one combines this with month-to-month seasonality, you get the following table:

S&P 500 Average Monthly Returns by Presidential Election Cycle Year, January, 1949-June, 2007:
Month:	Post-election	Midterm	Pre-election	Election		Average:
Jan	0.9%	-0.6%	4.7%	0.7%		1.4%
Feb	-1.2%	1.1%	0.9%	0.1%		0.2%
Mar	-0.2%	0.4%	2.3%	0.8%		0.8%
Apr	1.0%	0.0%	3.7%	0.5%		1.3%
May	1.5%	-0.5%	0.1%	0.1%		0.3%
Jun	-0.7%	-1.9%	1.8%	1.9%		0.3%
Jul	1.6%	0.5%	1.3%	0.2%		0.9%
Aug	-1.8%	-0.4%	1.4%	0.9%		0.0%
Sep	-1.2%	-1.3%	-0.6%	0.3%		-0.7%
Oct	0.9%	3.5%	-1.3%	0.6%		0.9%
Nov	1.4%	2.9%	0.8%	2.1%		1.8%
Dec	0.3%	1.7%	3.5%	1.2%		1.7%
Avg:	0.2%	0.4%	1.5%	0.8%		0.7%

Annualized S&P 500 Average Monthly Returns by Presidential Election Cycle Year, January, 1949-June, 2007:
Month:	Post-election	Midterm	Pre-election	Election		Average:
Jan	11.4%	-6.8%	73.7%	8.4%		21.7%
Feb	-14.0%	14.1%	11.2%	0.6%		3.0%
Mar	-2.3%	4.5%	31.4%	9.7%		10.8%
Apr	12.8%	0.5%	53.8%	6.0%		18.3%
May	19.0%	-6.4%	0.8%	1.2%		3.7%
Jun	-8.2%	-20.7%	24.3%	25.1%		5.1%
Jul	21.0%	5.5%	16.8%	2.1%		11.4%
Aug	-19.2%	-4.8%	18.4%	11.7%		1.5%
Sep	-13.1%	-14.4%	-7.3%	3.5%		-7.8%
Oct	11.5%	50.3%	-14.9%	6.8%		13.4%
Nov	17.8%	41.6%	9.8%	28.3%		24.4%
Dec	3.7%	21.9%	51.9%	15.8%		23.3%
Avg:	3.4%	7.1%	22.5%	9.9%		10.7%

What I've done here is broken down the year of the Presidential election year cycle versus the month, showing each month of the 48-month presidential election year cycle. For example, the upper left column is the average January S&P 500 monthly percentage change (0.91%) during post-election years.

Average Monthly Change in the S&P 500 by Month and Presidential Year, 1945-July 07:
	Pres Yr:
	Post-election	Midterm	Pre-election	Election	All:
Month:	1	2	3	4
1	0.92%	-0.30%	4.24%	0.54%	1.37%
2	-1.12%	1.06%	0.88%	-0.28%	0.14%
3	-0.12%	0.14%	1.82%	0.82%	0.67%
4	1.03%	0.45%	3.24%	0.97%	1.43%
5	1.47%	-0.50%	0.15%	0.42%	0.38%
6	-0.79%	-1.83%	1.84%	2.04%	0.29%
7	1.57%	0.24%	1.19%	1.57%	0.77%
8	-1.29%	-0.50%	1.19%	-1.29%	-0.01%
9	-0.51%	-1.89%	-0.76%	-0.51%	-0.76%
10	1.22%	3.08%	-1.10%	1.22%	1.01%
11	1.49%	2.65%	0.68%	1.49%	1.62%
12	0.54%	1.72%	3.20%	0.54%	1.62%
All:	0.36%	0.36%	1.40%	0.73%

Shown here are the average monthly changes in the S&P 500 broken down by month and Presidential election year. Note that pre-election years have been the strongest over the past half-century. The S&P 500 gained 1.40% on average (18.2% annualized) during pre-election years, followed by .73% (9.1% annualized) for election years. Postelection and midterm years are equally bad, returning roughly half (.36%) of the average monthly gain of the entire period. Put another way, the vast majority of a stock investors returns would have occurred during the pre-election and election years with the market treading water (returning only 4.4 percent before dividends) during the other years.

A quick scanning of this table shows that the strongest months of the year fall in the pre-election year (January (4.24%), April (3.24%), and December (3.20%), which also has the fewest number of down months (2: September and October). Almost half of the postelection and midterm years were down on average. Notice also that September is uniformly bad on average, showing an average negative return in all years of the election cycle. What is interesting is that August is also becoming a very weak month with the exception of pre-election years. With the exception of the pre-election year, October seems to have become a stronger month.

Of course, there will be strong Septembers and sell-offs during pre-election years (the crash of 87 occurred during a pre-election year, for example, although the S&P 500 closed up for the year), but in investing, as in poker, returns are very much a function of average probabilities. There is no guarantee that historical seasonal patterns will hold, but in investing more aggressively during the October through April, particularly during pre-election and election years, while lightening up or avoiding the market altogether during the weak August-September period seems a wise strategy.