S&P 500 Quarterly Return Model

[nodoodahs]

This is the projected return from today's close to 13 weeks from today.

AVERAGE MINIMUM MEDIAN MAXIMUM STDEV BEAT
5.83% -7.15% 6.13% 22.60% 6.64% 63.33%

This is a very strong buy signal for the S&P500. This score is in the top decile of the range.

I would put the odds of a new 4-year high for the index, to be set in the next 13 weeks, at 65%.
_________________
He was wearing my Harvard tie. Can you believe it? My Harvard tie. Like oh, sure, HE went to Harvard.

[HenryTo]

[nodoodahs]

From last Friday's close, predicted 13-week return of +3.6% which would have been 1239 from that point. The historic 13-week return for the base period is +2.88% so this is a reasonably strong buy signal.

Of course, the last really good buy signal was two hurricanes and -50 points ago ... LOL!
_________________
He was wearing my Harvard tie. Can you believe it? My Harvard tie. Like oh, sure, HE went to Harvard.

[nodoodahs]

+3.4% for end of year is this model's output. That would be 1270 and a strong 4-year high.
_________________
He was wearing my Harvard tie. Can you believe it? My Harvard tie. Like oh, sure, HE went to Harvard.

[nodoodahs]

For the close before Christmas, +2.2% from today. That would be 1258 and a four-year high.

Would be a nice Christmas present.

I don't think we'll dip below 1200 this year, but I also don't think we'll break 1300 either.
_________________
He was wearing my Harvard tie. Can you believe it? My Harvard tie. Like oh, sure, HE went to Harvard.

[nodoodahs]

For the 12/16 close, model is predicting +2.0%, which is below the historic +2.8%, so this is bearish from the average but predicting an absolute return.

+2.0% would be 1262 and a fresh 4-year high.

Still thinking a 4-year high has good odds but that 1300 is not in the cards for 2005.
_________________
He was wearing my Harvard tie. Can you believe it? My Harvard tie. Like oh, sure, HE went to Harvard.

[nodoodahs]

For the Dec 9th close. Model predicts +1.9% SPX return (below average which is +2.8%) and a close of 1266. At this point I still think another 4-year high is pending but I don't think we'll cross 1300 this year.
_________________
He was wearing my Harvard tie. Can you believe it? My Harvard tie. Like oh, sure, HE went to Harvard.

[nodoodahs]

Model is suggesting 1251 for the week after Thanksgiving. This is actually a slightly below median prediction for the data used in the experience period and one could interpret where the signal is (5th decile) as a neutral signal. However, "neutral" for the S&P 500 is a gain of 2.88% per quarter, non-seasonally adjusted.
_________________
He was wearing my Harvard tie. Can you believe it? My Harvard tie. Like oh, sure, HE went to Harvard.

[nodoodahs]

Well, now we're down 1.75% from that strong signal. However, the model isn't backing off much, suggesting that the close will be around 1243 before Thanksgiving.
_________________
He was wearing my Harvard tie. Can you believe it? My Harvard tie. Like oh, sure, HE went to Harvard.

[nodoodahs]

Well, we are two weeks in from that strong "buy" signal and the S&P is down 0.5% (not a typo, zero point five percent) which is a wash. Keep in mind, this model is trying to predict closing S&P 500 one quarter (13 weeks) from Friday.

This week's output, prediction for closing S&P 500 gain 13 weeks from today.

AVERAGE MEDIAN BEAT
4.27% 5.80% 61.67%

The "beat" refers to beating the historic +2.8% average quarterly return. Since we are less than 2% off what would be another 4-year high in the S&P 500, I must repeat the odds are good that we will make that high in the next quarter.
_________________
He was wearing my Harvard tie. Can you believe it? My Harvard tie. Like oh, sure, HE went to Harvard.

[nodoodahs]

This prediction is still in the top quintile of score range, very bullish. Strong buy signal. Prediction of another 4-year high for the S&P 500 in the next quarter is still on, probability > 65%.
_________________
He was wearing my Harvard tie. Can you believe it? My Harvard tie. Like oh, sure, HE went to Harvard.

[nodoodahs]

Hi Cary, welcome aboard!

I'm a value investor 1st, modeler 2nd, and starting to do some market timing lately. I started with a "bust" model and moved into doing some others.

The S&P 500 model is attempting to predict appreciation in the index over the next 13 weeks from data assembled after Friday's close. Inputs are derived from either current levels or recent changes in M3, 10 yr Treasury yields, Federal Funds Rate, and S&P 500 returns.

I've backtested the alpha on this model to 1982, with a move to cash 15% of the time and alpha of 3% annually over a "buy the index and hold" position. Since I'm working on bond and bust models as well, I may not move to cash, but may move to shorts or bonds.
_________________
He was wearing my Harvard tie. Can you believe it? My Harvard tie. Like oh, sure, HE went to Harvard.

[Cary]

Hi Bill. I'm new here. What model are you using, or is that confidential? If it is, what kind of indicators are you looking at in your model? Sounds interesting, as I'm seeing something totally different. Thanks.

Cary